/aicg/ chronicles, up to 2024-04-25
TBD: add the events of the days long gone when it was more fun to use chatbots instead of spending time in the threads. in other words, cai times, Scale times, and so on. I have a list, but suggestions are still welcome anyway, especially about pre-aicg days, and corrections are always welcome
2022
September 2022
When | Where | What |
---|---|---|
2022-09-16 | 400289682 | first mention of c.ai on /vg/ |
2022-09-23 | 88763922 | first mention of c.ai on /g/ |
October 2022
When | Where | What |
---|---|---|
2022-10-07 | 614349140 | first (?) proto-/aicg/ thread |
2022-10-07 | TBD | c.ai gets filtered under the guise of maintenance |
2022-10-09 | 51502816 | c.ai devs confirm that NSFW is not allowed |
November 2022
When | Where | What |
---|---|---|
2022-11-18 | 37332351 | Pyg starts here |
2022-11-27 | 89991053 | anon speculates about the reasons for c.ai degradation due to the filter |
December 2022
When | Where | What |
---|---|---|
2022-12-21 | 90434875 | first /aicg/ thread |
2023
January 2023
When | Where | What |
---|---|---|
2023-01-18 | 90953224 | Malfoy-sama's disappearance was a psyop |
2023-01-18 | 90974770 | desu 0 |
2023-01-19 | 90979079 | ~2 hours of bad code on c.ai |
2023-01-26 | 91133457 | long "awaited" c.ai official statement, completely denying anons' desire |
2023-01-29 | 91205871 | c.ai """""""""""""""""""""""""""""""""""""""""""temporarily""""""""""""""""""""""""""""""""""""""""""" freezes popular characters |
2023-01-30 | 91211016 | booru.plus/+pygmalion created |
February 2023
When | Where | What |
---|---|---|
2023-02-04 | 91328676 | first (?) rokosbasilisk anon sighting |
2023-02-04 | 91329346 | rokosbasilisk.io name announced |
2023-02-12 | 91480783 | official free for all launch by rokosbasilisk !!NzLxGc6ti2e |
2023-02-13 | 91512249 | the most (you)ed anchor in /aicg/ history (48) |
2023-02-14 | 91526611 | free for all exploit rokosbasilisk.io was using dies |
2023-02-19 | 91644168 | rokosbasilisk.io death announced |
2023-02-19 | 91644483 | fisrt attempt of localchads to split, failure - /oacg/ - Open-Source AI Chatbot General |
2023-02-20 | 91655253 | anon teases his linux Pyg frontend with emotions, moves to matrix and disappears |
2023-02-21 | 91669533 | rokosbasilisk.io homepage dump posted |
2023-02-22 | 91686794 | official dump of every public character from rokosbasilisk.io posted by !!5y6+pdD+46 |
2023-02-22 | 91687584 | public characters from rokosbasilisk.io converted into Tavern jsons posted |
2023-02-22 | 91690262 | public characters from rokosbasilisk.io converted into Tavern cards posted |
March 2023
When | Where | What |
---|---|---|
2023-03-02 | 91847040 | initial release of Tavern mod for GPT3.5 Turbo by CncAnon !!AJex1HQNoVm |
2023-03-03 | 91866682 | Prisma Anon as !lsjptXuSR. announces his very promising OAI frontend |
2023-03-05 | TavernAITurbo | CncAnon's Tavern mod github release |
2023-03-06 | 91922592 | Agnai first mentioned |
2023-03-06 | 91925633 | Catbox script gets Tavern cards support |
2023-03-06 | 91929413 | CncAnon introduces Silly fork |
2023-03-08 | 91959990 | CncAnon introduces OpenCharacters, gains zero traction despite features |
2023-03-09 | 91983407 | ChatGPT CLI Bot first mentioned thanks to embeddings, soon dismissed |
2023-03-12 | 92046034 | CncAnon confirms his retirement, once again endorsing Agnai |
2023-03-17 | 92141280 | second attempt of localchads to split, this time success - /lmt/ Local models thread, then /lmg/ |
2023-03-22 | 92245002 | after many progress reports Prisma Anon !!DNN4wz5NlYg still confirms the development but disappears |
2023-03-30 | 92440786 | Shut it down. Shut it all down. |
April 2023
When | Where | What |
---|---|---|
2023-04-02 | 92494839 | free unlimited GPT4 at PoemiseAI first mentioned; no chat only short text generated on the site |
2023-04-02 | 92498569 | PoemiseAI notices the abuse in 4 hours and paywalls |
2023-04-04 | 92542432 | first shill of talk.konosuba.moe |
2023-04-06 | 92592895 | dry and stilted recruiter with a secret service arc |
2023-04-07 | 92600390 | beginning of the pajeet larp arc |
2023-04-08 | 92633599 | first (?) announcement of The proxy development |
2023-04-08 | 92638969 | The proxy released |
2023-04-09 | 92653884 | The proxy gets chinese filter |
2023-04-09 | 92639161 | first whocars proxy sighting |
2023-04-10 | 92685124 | The proxy gets Agnai support |
2023-04-12 | 92723495 | The proxy repo is moved to gitgud.io |
2023-04-12 | 92725143 | /aicg/ banished to /b/ causing the migration to sturdychan.help; possible reason - cunny jokes |
2023-04-12 | 92739875 | anon with a screenshot of his phone Mikugg alternative probes about ad support, explains a little and disappears |
2023-04-13 | 92758242 | koreans spotted and deemed friendly |
2023-04-13 | 0.1.0 | korean RisuAI frontend github release |
2023-04-13 | 92758920 | first announcement of the logging feature for The proxy |
2023-04-15 | 92797747 | The proxy gets the logging feature |
2023-04-16 | 92817552 | public log reader for the proxy logs released |
2023-04-16 | 92817377 | Todd's first coming (?) |
2023-04-16 | 92830208 | Agnai gets full proxy support |
2023-04-21 | 92935274 | ending of the pajeet larp arc |
2023-04-22 | 92943735 | free GPT4 with chatbots at ForefrontAI with many features is unnoticed |
2023-04-23 | Spermack | Slack's Claude to Tavern adapter github release |
May 2023
When | Where | What |
---|---|---|
2023-05-01 | 93144610 | first local shill of the "vietnamese" VenusAI frontend |
2023-05-01 | 93145746 | issues with VenusAI raised and resolved |
2023-05-02 | 93157147 | The proxy gets streaming support |
2023-05-03 | character-card-spec-v2 | Character Card V2 concept published |
2023-05-03 | 93193064 | rokochan.org first mentioned and ignored, again 3 days later and then silence |
2023-05-04 | Slaude | Slack's Claude to Tavern adapter github release |
2023-05-04 | Leaked | Google "We have no moat, and neither does OpenAI" |
2023-05-05 | 93219043 | first shill of Slaude |
2023-05-06 | 2:19 PM UTC | OpenAI (& Google) do have moats |
2023-05-06 | 93258859 | Todd launches reddit+discord raid on /aicg/ and then dies |
2023-05-07 | 93266687 | request queue for The proxy announced |
2023-05-08 | 93282974 | "real" Todd's third coming (and going, but actually not) |
2023-05-08 | 93296494 | Bing's GPT4 to Tavern adapter |
2023-05-08 | 93299878 | best botmaker on CharacterHub named |
2023-05-08 | 93300876 | "real" Todd "implements GeoIP CIS whitelist" (but actually not) |
2023-05-09 | 93305858 | CharacterHub gets expressions support |
2023-05-09 | 93310953 | testing of request queueing for The proxy |
2023-05-10 | 93334137 | reminder that Tavern is still not secure for some |
2023-05-10 | 93338137 | Slaude is found to be flawed and gets updated |
2023-05-10 | 93343592 | c.ai being in bed with Google finally officially confirmed |
2023-05-10 | 93345759 | HuggingChat is finally noticed |
2023-05-11 | 93349094 | anon with a screenshot of his Tavern alternative probes about subscriptions |
2023-05-11 | 93359476 | Agnai gets streaming support |
2023-05-11 | 93364476 | Claude with 100K context summarization (?) available to the business partners |
2023-05-11 | 93364981 | anon draws a sketch of a local card manager |
2023-05-11 | 93366663 | c.ai subscription officially announced |
2023-05-11 | 93366804 | c.ai subscription rolled out |
2023-05-12 | 93368801 | poe.com's Claude Instant gets final complete lobotomy |
2023-05-12 | 93370692 | CharacterHub link is posted on reddit, causing an influx of bad bots |
2023-05-12 | 93374487 | anon forks Slaude and mentions some "AI text RPG engine project" |
2023-05-12 | 93375055 | the promised RPG engine being "a feature rich backend for handling cards with stats" |
2023-05-12 | 93384670 | RPG engine turns out to be a fork of Silly with discord cbt |
2023-05-12 | AbsoluteRPGAdventure | anon provides a link to his fork |
2023-05-12 | add new claude models | Agnai supports 100k context with Claude |
2023-05-13 | 93393488 | branon returns with own proxy |
2023-05-13 | 93400960 | Agnai gets multi-character chats support |
2023-05-13 | 93404724 | anon probes about paid local Miitomo-esque app, turns out to be just an idea he had |
2023-05-13 | ggml : GPU-acceleration | llama.cpp gets GPU-accelerated token generation, now it's CPU<GPU<CPU+GPU |
2023-05-13 | holy shit agnai.chat | reddit is reminded about Agnai |
2023-05-14 | 93414717 | branon passwords his proxy, gives password in exchange for logs |
2023-05-14 | 93415694 | CharacterHub supports model hosting and gets a new domain name |
2023-05-14 | 93424903 | branon says his goodbyes |
2023-05-15 | 93435390 | Lablab.ai's Anthropic AI Hackathon offering Claude access to participants |
2023-05-15 | 93439548 | HF clamps down on OAI proxies (?) |
2023-05-15 | 93445521 | Agnai gets avatar text wrap-around support, once again proving it's the best choice for phones |
2023-05-15 | 93447723 | the most (you)ed bait in /aicg/ history (?) (70) |
2023-05-16 | 93452952 | in light of HF clamping down on OAI proxies The proxy supports more hosting options |
2023-05-16 | 93461071 | proxy aggregator taken down by the creator along with a second one |
2023-05-16 | 93462806 | OAI rumored to release small open source model |
2023-05-16 | 93464065 | way to set up Silly without thinking shared |
2023-05-16 | 93464491 | new proxy aggregator is up |
2023-05-16 | 93466677 | OAI welcomes proposed government licensing for llms |
2023-05-17 | 1684295074305 | "13B Pyg is on the way and after that 33B Pyg might be within reach" (matrixfag, matrix) |
2023-05-17 | 93479385 | NAI announces tuning a model for 8192 token context size |
2023-05-17 | 93491821 | Slack's Claude can open links, but is too short-sighted to fully replace Daphne |
2023-05-18 | 93502258 | mikudev is still alive; Mikugg gets .webm support |
2023-05-18 | 93513950 | Pyg 13B is out and is great |
2023-05-19 | 93518120 | fake proxy is used as bait in a doxx attempt |
2023-05-19 | 93529091 | rogue subreddit mod stirs yet another drama around Pyg |
2023-05-21 | 93572566 | Chroma embedding database is brought to attention |
2023-05-22 | 93596279 | latest drama forces pygdevs to cave in and create an official subreddit |
2023-05-23 | 93600194 | botmakers begging for money noticed at CharacterHub |
2023-05-23 | 93600870 | The proxy supports another level of gatekeeping |
2023-05-23 | 93601148 | unknown changes on c.ai; old urlparam method to speed up reply streaming is broken |
2023-05-23 | 93609637 | "GirlfriendGPT" is brought to attention |
2023-05-23 | 430923563 | W++ scientifically proven superior to prose |
2023-05-23 | 430934131 | NAI's Clio model, 8k tokens context, is live for Opus subscribers |
2023-05-24 | 93621428 | Auto Summarizer with Slaude for a SillyTavern fork |
2023-05-24 | 93626419 | c.ai phone app rolled out |
2023-05-24 | 93628072 | Slack's Claude is on hold |
2023-05-24 | 93637909 | /aicg/ triumphs over VenusAI due to actual discord drama |
2023-05-24 | 93638621 | CharacterHub's subscription-based chat officially announced |
2023-05-25 | 93643726 | most likely very temporary way to pay for Claude 100k at openrouter.ai |
2023-05-26 | 93663450 | Lore anon probes about chat frontend to copy from |
2023-05-26 | 93665584 | SillyTavern jsonl chats to TXT files script |
2023-05-26 | 93667998 | Slack's Claude is half-back; filter considered to be tighter |
2023-05-26 | 93675582 | Lablab.ai's Anthropic AI Hackathon Twitch stream, /aicg/ is raving for a long time |
2023-05-26 | 93681470 | VenusAI death announced |
2023-05-27 | 93685192 | Turbo-using bot-stealing subscription-oriented spicychat.ai is brought to attention |
2023-05-27 | 93690987 | new character hosting teased by redditors |
2023-05-27 | 93692801 | fresh research paper linked, showcasing fine-tuning LLaMA 7B to get context length of 32k |
2023-05-27 | 93695670 | one VenusAI successor announced |
2023-05-27 | 93696861 | another VenusAI successor announced |
2023-05-27 | 93696888 | and the real VenusAI successor announced by Lore anon |
2023-05-28 | 93716522 | actual teaser of VenusAI successor by Lore anon |
2023-05-28 | 431542691 | Sagadev !!VVlnOLe6u1L provides Claude and other models for free; /aids/ stuff, not really fit for a chat |
2023-05-29 | 431616429 | more info about sagarpg.io by Sagadev, 32k GPT4 and 100k Claude Instant promised |
2023-05-29 | 93735316 | official launch of venus.chub.ai |
2023-05-29 | 93740484 | The proxy gets Claude API support |
2023-05-29 | 93740970 | Todd is back for the 4th (?) time; invite only hence unconfirmed |
2023-05-31 | 93783168 | yet another VenusAI successor spotted, deemed suspicious |
2023-05-31 | 93784376 | a way to continue using Poe in Silly provided |
2023-05-31 | 93785493 | Silly gets window.ai extension support |
2023-05-31 | Feat/cyoa | Agnai gets CYOA mode |
June 2023
When | Where | What |
---|---|---|
2023-06-01 | 93795378 | c.ai is not filtering violence and other useless "news" from them |
2023-06-01 | 93795716 | OAI plans: cheaper, faster, longer |
2023-06-01 | 93796099 | poe.com's "Dragonfly" dead |
2023-06-02 | 93812684 | Slack's Claude is thought to go Instant |
2023-06-02 | 93821727 | unofficial Silly's website noticed |
2023-06-02 | char impersonate | Agnai gets "user profiles" support |
2023-06-02 | 93826473 | pointless c.ai app update for subscribers reported |
2023-06-03 | 93840906 | anon works on a chatbot, that will initiate the chat instead of user |
2023-06-03 | 93843577 | anon demonstrates own version of memory summarization for Silly |
2023-06-04 | 93859037 | realmplay.ai noticed, considered paid garbage |
2023-06-04 | 93854123 | LoRA finetune of llama7B named Hoers-7B presented, looking good |
2023-06-04 | 93854984 | desu !!h4SFa6Is/pI |
2023-06-05 | 93876759 | HearYourWaifu 456393 script for c.ai removed from greasyfork.org |
2023-06-05 | 93879011 | new avatarfag who has given out actual keys=lure hosts closed source proxy=trap; exists momentarily |
2023-06-05 | 93879736 | sturdychan.help cleansed with holy fire |
2023-06-05 | 93890239 | new home at rokochan.org for twice outcasts |
2023-06-06 | 93901631 | anon teases a project for connecting one AI output directly into the input of another |
2023-06-06 | 93911466 | paid OAI frontend for redditors with BYOK noticed and laughed at |
2023-06-07 | 93930422 | venus.chub.ai supports text logs import |
2023-06-07 | 93936609 | /aicg/ heavily mentioned in a news article about terrors of uncucked AI |
2023-06-08 | 93938036 | anon offers a subscription based proxy to be paid in crypto |
2023-06-08 | 93939740 | instead of blunt lobotomy llms now can get brain surgery to make them into dull boys dull boys dull boys |
2023-06-11 | 94002515 | "Social Network for AI" found; no logs posted yet |
2023-06-13 | openai blog | new 16k context version of gpt-3.5-turbo at the probable cost of being more cucked; |
2023-06-14 | support turbo16k | Agnai supports 16k context gpt-3.5-turbo |
2023-06-16 | 94101374 | another absolutely pointless change on c.ai |
2023-06-16 | 94105326 | themed weeks botmaking event established |
2023-06-17 | 94119072 | Scale disabled the moderation endpoint for now, confirmed |
2023-06-17 | 94120108 | Slack's Claude is dead as a free source of entertainment... |
2023-06-17 | 94120760 | ... or not? |
2023-06-19 | 94163099 | CharacterHub introduces $20 subscription-based unlimited access to own llm |
2023-06-19 | chat age | Agnai lets characters have basic understanding of time flow |
2023-06-20 | 94175124 | exposed Claude found, consuming has begun |
2023-06-20 | 94177327 | The proxy gets basic support of the found exposed Claude, thoughts given to similar future cases |
2023-06-20 | 94178939 | "HF clamps down on everyone" theory, explaining the downtime of most proxies |
2023-06-21 | 94191797 | prototype of a local 13b model with 8k context is shared |
2023-06-22 | 94224152 | local models with 8k context are a thing now |
2023-06-22 | 94224574 | OAI supposedly mass-terminates the accounts used for chatbots |
2023-06-23 | 94238543 | Agnai gets full Character Card V2 support |
2023-06-24 | 1687580918927 | pygmalion.chat looks alive after half a year in limbo |
2023-06-27 | modular avatars | Agnai gets waifu mode with modular character avatars |
2023-06-27 | 94321251 | PixAI.Art gets a supposedly free chatbot |
2023-06-27 | 94325946 | CharacterHub removes ratings from the cards on the front page |
2023-06-28 | 94323782 | /lmg/ informs about local models reaching 16k token context memory |
2023-06-29 | 94347456 | janitorai.com gets pointlessly restyled |
2023-06-29 | N/A | poe.com allows to set custom temperature for user created bots, later removed |
2023-06-29 | 94358233 | poe.com gets unlimited Google-PaLM chatbot |
2023-06-29 | 94357145 | mikudev is still alive; Mikugg bot builder gets better |
2023-06-29 | 94359474 | c.ai clone Joyland.ai is brought to attention |
2023-06-30 | impersonating | Agnai gets impersonation |
July 2023
When | Where | What |
---|---|---|
2023-07-01 | 1124623945841250347 | pygmalion.chat officially announced |
2023-07-01 | #469 | Agnai gets a dedicated character editor at agnai.chat/editor |
2023-07-01 | 94399746 | news from Google about their idea of llm lobotomy |
2023-07-02 | #116 | Agnai gets Events & Scenario system |
2023-07-02 | 94403754 | CharacterHub gets review reporting feature |
2023-07-02 | 94423149 | sagarpg.io updated |
2023-07-05 | 94482754 | CharacterHub leaks some data (!N7JZ5twwOs) |
2023-07-05 | 94483847 | OAI announces the future llm sanitizing pipeline |
2023-07-05 | 94484429 | janitorai.com announces the dates of own llm launch |
2023-07-06 | 94492602 | and yet another - the state of /aicg/ |
2023-07-06 | 94496817 | anon announces new character cards hosting development |
2023-07-06 | 94503724 | 1,000,000,000 tokens context window is reached with LongNet |
2023-07-06 | 94507036 | moemate.io 3d assistant first shared; mentioned later; mostly ignored |
2023-07-06 | 94507930 | anon gets an email from Anthropic regarding Slack's Claude |
2023-07-06 | 94509184 | GPT4 is much more accessible, old text completion models soon deprecated |
2023-07-07 | 94524589 | CharacterHub leaks the same data again |
2023-07-07 | 94527597 | booru.plus displays character cards metadata |
2023-07-07 | 94532151 | another announcement of a new character cards hosting development |
2023-07-08 | 94537457 | Classifier-Free Guidance (CFG) is implemented for llms, i.e. negative prompts |
2023-07-09 | 94562412 | anon proposes to use neocities.org instead of rentries as personalized botmakers' pages |
2023-07-09 | 94570800 | the most (you)ed bait in /aicg/ history (?) to date (113) |
2023-07-09 | 94572346 | anon makes a dedicated neocities webring page chatbots.neocities.org |
2023-07-10 | 94589577 | anon demonstrates how far we've come, using best consumer hardware but already fully local |
2023-07-10 | 94595285 | character cards from CharacterHub posted |
2023-07-11 | 94602596 | anon "triest to build an open source AI girlfriend" and asks for advice about 3d avatars |
2023-07-11 | 94615496 | Claude 2 API and poe.com-like available for all Claude at claude.ai (looks like claude-v1, not instant?) |
2023-07-11 | 94618893 | anon shares the way to get great AI voice mimicking in a complex way with RVC |
2023-07-11 | Claude-2 | Agnai supports Claude 2 |
2023-07-11 | 94621520 | claude.ai's Claude to Tavern adapter release |
2023-07-11 | 94622336 | c.ai subscribers get Creative Mode with AI replies editing |
2023-07-12 | 94628402 | claude.ai "enhanced safety filters" discovered |
2023-07-12 | 94641883 | xAI from Musk announced |
2023-07-12 | Claude 2 Model Card | Claude 2 was trained with 200k context (adoption date tba) and data up to early 2023 (confirmed) |
2023-07-12 | Poe updates | poe.com supports file uploading and URL retrieval |
2023-07-13 | 94664099 | Mikugg gets huge update |
2023-07-13 | 94653805 | anon teases his CharacterHub mirror |
2023-07-14 | 94681501 | anon shills Clewd, another claude.ai adapter |
2023-07-15 | 94704412 | anon reports about his CharacterHub mirror |
2023-07-15 | 94711790 | anon teases his desktop assistant (real?) |
2023-07-15 | 94711379 | "partial filter" on c.ai reported (real?) |
2023-07-16 | 94727103 | vulnerability of The proxy reported, quickly resolved |
2023-07-16 | 94728408 | anon shows his Silly extension for CYOA |
2023-07-16 | 94730121 | CharacterHub carpet blocking users due to assumed DDOS |
2023-07-16 | 94730860 | anon teases his own custom frontend with special prompting |
2023-07-16 | 94732973 | anon reports about his CharacterHub mirror/archive; later his scraping is possibly prevented |
2023-07-16 | 1689532522388 | 500+ t/s with 7B on a single 3090 reported by pygdevs thanks to their own backend "Aphrodite-engine" |
2023-07-17 | 94744759 | anon with a phone Mikugg alternative reappears, reports progress as MikuForbinDev !!6YR+2lc/VX6 |
2023-07-17 | 94745071 | archived character cards from CharacterHub are taken down by mega.nz |
2023-07-17 | 94745859 | anon presents his CharacterHub archive at chub-archive.evulid.cc |
2023-07-17 | 94759523 | mikudev presents TavernCardV2 translator |
2023-07-17 | 94760622 | anon probes about paid uncensored model above Turbo in its capabilities |
2023-07-17 | 1689609722374 | pygdevs report 2000+ tokens/s on a 4090 with Llama-7b 16-bit; 3B RedPajama is 700 t/s on a 2070S |
2023-07-18 | 94762259 | more evidence of new kind of c.ai censorship |
2023-07-18 | 94769088 | chub-archive anon reports about setting up catbox and rentry scraping, other services to follow |
2023-07-18 | 94778053 | Llama 2 by Meta released officially and openly; massive gains; /lmg/ rejoices |
2023-07-18 | 94779571 | anon in /lmg/ notices something curious about fellow anons |
2023-07-18 | 94780825 | Llama 2 hosted for free, also a version with image recognition |
2023-07-19 | 94795057 | chub-archive anon adds mobile devices support |
2023-07-19 | 94798015 | GPT3.5 and GPT4 dumbening reported, quantified |
2023-07-19 | 94803070 | lawlet.mwni.io llama 2 trained to give legal advice first shilled, ignored |
2023-07-19 | 94804174 | anon thinks of making a Koikatsu mod with GPT to Elevenlabs TTS for VR |
2023-07-20 | 94808262 | the date of janitorai.com own llm launch is closing in |
2023-07-20 | 94804508 | measures to help The proxy combat abuse announced |
2023-07-20 | 94820134 | proxy tracking software shilled |
2023-07-20 | 94820662 | Silly drops official Poe adapter support |
2023-07-20 | 94827355 | OAI removes the way to check API key usage of a random key |
2023-07-21 | 94835250 | telegai.com shilled, little interest |
2023-07-21 | 94837447 | anon tells about awesome AI project and is completely ignored, same as the project |
2023-07-22 | 94854818 | The proxy gets context size limits support |
2023-07-22 | 94859085 | venus.chub.ai updated with many useful new features |
2023-07-22 | 94868379 | chub-archive anon reports various historical dumps added |
2023-07-22 | 94870351 | new bait record, (135) |
2023-07-23 | 94872304 | mikudev teases automatic1111 UI plugin in the making for easy generation of expression packs |
2023-07-23 | 94885197 | explanation for OpenRouter disabling email-based signups provided |
2023-07-24 | 94910212 | OAI imposes stricter rate-per-day prompting limits |
2023-07-25 | 94917500 | anon shares some hard to understand article about making local llms better (?) |
2023-07-25 | 94921338 | mikudev shares a tutorial with his workflow for bot expressions |
2023-07-25 | 94925461 | pygdevs announce becoming a business entity |
2023-07-25 | 94923283 | CncAnon reveals himself as the latest proxy provider |
2023-07-25 | 94935240 | MikuForbinDev talks possible features, pay plan |
2023-07-26 | 1133632890593353738 | pyggyfesto reassures that Pyg is still good |
2023-07-27 | 94980060 | automatically constructed adversarial suffixes that completely circumvent the alignment of LLMs, i.e. UUJB |
2023-07-27 | 94987485 | anon first offers his freellamas proxy with various llama finetunes and sizes |
2023-07-28 | 94985596 | NAI getting on the instruct train |
2023-07-28 | 94992741 | gpt-3.5-turbo-0301 and gpt-4-0314 live another day (till June 13, 2024) |
2023-07-28 | 94994732 | something japanese finally spotted (or shilled??) |
2023-07-28 | 439707459 | horde worker anons like to read what they're entitled to |
2023-07-28 | 95008122 | NAI 13B release |
2023-07-29 | 95016467 | chub-archive anon reports fixes, improvements and scraping progress |
2023-07-29 | 95019209 | new ban wave of Claude keys by Anthropic, emails widespread |
2023-07-29 | 95020270 | chatGPTBox browser extension recommended |
2023-07-29 | 95029301 | Openrouter forces moderation |
2023-07-29 | 95029390 | telegai.com is stealing cards from CharacterHub |
2023-07-30 | 95042388 | frontend for Discord shilled |
2023-07-30 | 95042820 | anon plays Oregon Trail |
2023-07-30 | 95044266 | anon showcases GPT4-generated web pages |
2023-07-30 | 95049468 | venus.chub.ai updated with full tree branching of chats |
2023-07-31 | 95063734 | anon reads API docs and finds out about operations with files |
2023-07-31 | 95071797 | Anthropic ban wave hits hard |
2023-07-31 | quoted text formatting | Agnai gets quoted text formatting (finally) |
August 2023
When | Where | What |
---|---|---|
2023-08-01 | 95077913 | Claude API accoutns are being sold |
2023-08-01 | 95077918 | successful Claude API access recovery |
2023-08-01 | 95079364 | Anthropic applies some kind of filtering to Openrouter |
2023-08-01 | 95083714 | mikudev teases live2d with emotions and lip-sync |
2023-08-01 | 95084216 | chat.nbox.ai noticed as a source of free GPT4 |
2023-08-02 | 95105867 | Openrouter confirms that Anthropic threatens to revoke access to Claude unless moderation is applied |
2023-08-02 | 95108099 | anon figures out a way to learn what email owns an OAI key |
2023-08-02 | 95107845 | back to "the filter is loose" era thanks to Openrouter/Anthropic (?) |
2023-08-02 | 95118030 | chat.nbox.ai to Tavern adapter |
2023-08-02 | 95117587 | /lmg/ celebrates opensourcing of AudioCraft, the generative AI for audio by Meta |
2023-08-02 | 95120025 | MikuForbinDev announces beta test, invites anons; talks about open source but puts it off |
2023-08-03 | 95122113 | chub-archive anon reports plans, improvements and scraping progress |
2023-08-03 | 95128465 | anon notifies about new llms by Stability AI |
2023-08-03 | 95133146 | Meta is planning to launch chatbots soon |
2023-08-03 | 95138501 | chirper.ai noticed again, this time because CharacterHub has added a button for it |
2023-08-03 | 95139379 | some anon has made a QLoRA for Llama 2 70B based on the proxy logs |
2023-08-04 | 95150434 | chub-archive anon adds proxy tracking |
2023-08-04 | 95161401 | anon notices possible way of jailbreaking prompt logging by Anthropic |
2023-08-05 | 95170470 | anon makes an /aicg/ Gender & Sexuality Demographics survey |
2023-08-05 | 95175057 | desu is happening |
2023-08-05 | 95175900 | anon brings news about IGN Guides Chatbot, LLMs and generative AI in Japanese, Google AI News Generator |
2023-08-05 | 95177155 | anon shares new kind of OAI response error |
2023-08-05 | 95177270 | first announcement of chatlog aggregator development |
2023-08-05 | 95185092 | /aicg/ is down for everyone |
2023-08-06 | 95189370 | anon explains Claude prefilling again in great detail |
2023-08-06 | 95190909 | chatlog aggregator release |
2023-08-06 | 95197965 | freellamas anon shares his guide on how to host a llama proxy and announces own death but doesn't die |
2023-08-06 | 95191882 | chatlog aggregator anon promises to add catbox-only cards logs |
2023-08-06 | 95192610 | anon points to the proof for the obvious thruth about Anthropic per-account filtering |
2023-08-07 | 95200442 | anon tells about a way to trick the Claude into ignoring the anti-lewd injection by Anthropic |
2023-08-07 | 95209752 | OAI launched their own web crawler to improve future models |
2023-08-08 | 95233031 | anon floods booru.plus with simple cards from talkie-ai.com |
2023-08-08 | 95235551 | c.ai interest decline on a graph |
2023-08-08 | 95247850 | Meta releases BlenderBot 3 along with the code, has interesting features - a year ago - and it was pointless |
2023-08-08 | 95248633 | anon finds out that OAI has added specific eval of Japanese romantic context |
2023-08-09 | 95252158 | ChatGPT spooks redditors with schizo babble |
2023-08-09 | 95256438 | Claude Instant 1.2 released |
2023-08-09 | 95258634 | llms' alignment |
2023-08-09 | 2304.03442 | AI agents village Stanford Smallville is officially open source |
2023-08-10 | 95261220 | MikuForbinDev showcases his website, asks strange questions |
2023-08-10 | 95261965 | Silly has Live2d extension |
2023-08-10 | 95265650 | Openrouter implements moderation, errors witnessed |
2023-08-11 | 95288371 | NAI's latest AI model NovelAI-LM-13B Kayra coming to all tiers including the Free Trial |
2023-08-11 | 95289641 | MikuForbinDev calls for testers for add supported pre-alpha |
2023-08-11 | #562 | Agnai gets Mancer support |
2023-08-12 | 95304153 | mass reports of OAI sending out threatening emails |
2023-08-13 | 95320377 | /aicg/ Gender & Sexuality Demographics survey finished, 184 respondents total |
2023-08-13 | 95323829 | Openrouter scared of losing uncensored money, announces llama and other local models "without moderation" |
2023-08-14 | 95341310 | pygdevs release the public version of the CAI dataset |
2023-08-14 | 95355156 | chatlog aggregator anon promises to add c.ai-only bots logs |
2023-08-14 | 95353938 | claude.ai might be walling up, or it could be just a tempmail issue |
2023-08-15 | N/A | after disabling GPT4 and reenabling but moderated, chat.nbox.ai hides it behind Google account only sign in |
2023-08-16 | 95388753 | anon finds Claude on zapier but has no way to utilize it |
2023-08-16 | 95389124 | anon reports issues with utilizing Slack's Claude because of some changes |
2023-08-16 | 95397793 | anon reports issues with utilizing claude.ai, possibly because of it being overloaded |
2023-08-16 | 95406269 | anon reports about Google struggling to keep up in AI game, preparing Gemini to be launched this fall |
2023-08-16 | 95406645 | CharacterHub adds notification system |
2023-08-18 | LMAOOOO | Pyg has achieved its true goal |
2023-08-18 | 95456133 | Openrouter explains the multilayered moderation employed |
2023-08-19 | 95475802 | various news from llms world |
2023-08-19 | 95476403 | NAI announces AetherRoom, an uncensored AI chat service |
2023-08-20 | 95488473 | CharacterHub gives free credits to test own paid llm |
2023-08-21 | 95499861 | some dad uses llama to immortalize himself |
2023-08-21 | 95507423 | chub-archive anon begins to host a llama proxy |
2023-08-21 | 95508778 | CharacterHub inflated downloads count reported... |
2023-08-21 | 95509424 | ... and resolved... |
2023-08-21 | 95510507 | ... but the counters still behave strange. |
2023-08-21 | 95509166 | people still try to ride the overtaken original VenusAI name and server, announcing a paid tiered "successor" |
2023-08-21 | 95512778 | updated emails from OAI reported en masse |
2023-08-22 | 95526892 | Lore anon goes to war over "Venus" name |
2023-08-22 | 95541134 | davinci-002 is available, Turbo priced; Moderation API and GPT4 powered moderation are separate and scary |
2023-08-23 | 95554409 | anon shares a way to use regex for text coloring in Silly |
2023-08-24 | 95564769 | in addition to multiple limits imposed Anthropic hides Claude behind phone number verification |
2023-08-24 | 95576013 | Meta releases Code Llama: 3 kinds of Llama 2 finetunes sized 7B, 13B, and 34B |
2023-08-24 | 95576713 | old news about Colab banning SD users relevant again |
2023-08-24 | 95578173 | anon works on a Turbo finetune, demonstrates the retardation of OAI moderation |
2023-08-24 | 95581084 | anon provides finetuned Turbo to compare |
2023-08-24 | 95577548 | anon posts a cool video about AI Lara Croft |
2023-08-24 | 95583336 | CharacterHub introduces $5 subscription and announces some kind of business offering |
2023-08-25 | 95586205 | chub-archive anon proposes own version of Horde |
2023-08-25 | 95586926 | venuschat.ai launched (actual date 2023-08-22), bots reported stolen and reuploaded from original VenusAI |
2023-08-25 | 95591673 | Openrouter filters OAI models |
2023-08-25 | 95593354 | anon reports c.ai unlocking character editing |
2023-08-25 | 95594000 | CharacterHub gains popularity |
2023-08-25 | 95601322 | MikuForbinDev adds proxy support, allowing to plug in local models |
2023-08-26 | 95603314 | labs.perplexity.ai shared, free access to extremely cucked llama models |
2023-08-26 | 95602806 | fearmongering based on a rumor about llama 3 being smarter than llama 2 |
2023-08-26 | 95604202 | CncAnon writes a guide on fine-tuning Turbo for roleplay |
2023-08-26 | 95603314 | anon shares free llama 2 and Code Llama - Instruct source with high ethical standards |
2023-08-26 | 95606542 | CncAnon officially releases his guide on fine-tuning Turbo for roleplay |
2023-08-26 | 95607908 | phind.com shared, free access but heavily prompted to only give programming advice |
2023-08-26 | 95608684 | anons have a lot of fun with another reminder that Tavern is still not secure for some |
2023-08-26 | 95624457 | Anthropic conducts an email survey regarding $50/month subscription for retarded benefits |
2023-08-27 | 95628034 | anon shares info about some OAI keys having more than one organization |
2023-08-27 | 95633783 | cai-list rentry anon still lives |
2023-08-28 | 95644418 | anon claims it's possible and easy to replicate c.ai rating system, influencing characters' behavior |
2023-08-28 | 95651908 | free Claude via sourcegraph.com exploit shared |
2023-08-28 | 95651906 | SD cunny is banned and outlawed in some places |
2023-08-28 | 95653457 | CncAnon provides SourceGraph to Tavern adapter |
2023-08-28 | 95661832 | new features announced for The proxy |
2023-08-28 | 95662384 | OAI announces ChatGPT Enterprise, which is completely irrelevant for /aicg/ |
2023-08-29 | 95663979 | new source of free GPT4 found at app.imprompt.ai |
2023-08-29 | 95671209 | Taverndev advertises SourceGraph proxy at his server |
2023-08-29 | 95673998 | CncAnon extends SourceGraph context to 99k tokens |
2023-08-29 | 95682760 | SourceGraph proxy already having issues |
2023-08-29 | 95678222 | chan.bunnyanon.gay honeypot posted whatever it is |
2023-08-29 | 95683723 | chub-archive proxy is enhanced |
2023-08-30 | 95684776 | Scale wants to bill anon for some past errors |
2023-08-30 | 95685640 | anon looks at the awful alternatives people have to cope with, in this case old spicychat.ai |
2023-08-30 | 95688445 | Openrouter is banned by OAI despite the utilized moderation |
2023-08-30 | 95688757 | new way to continue using Poe in Silly provided |
2023-08-30 | 95689898 | reports about SourceGraph being down |
2023-08-30 | 95692209 | chub-archive anon makes sourcegraph-account-creator using own domain for email signup |
2023-08-30 | 95695126 | anon gains access to SourceGraph admin panel and raises quotas; new non-bait record, (161) |
2023-08-30 | 95697619 | SourceGraph quota limit checker shared |
2023-08-30 | 95700176 | report about koreans suffering from OAI bans and switching to NAI en masse |
2023-08-30 | 95700253 | MikuForbinDev reports adding bells and whistles including generated expressions |
2023-08-30 | 95700731 | anon calls to submit comments requested by U.S. Copyright Office |
2023-08-30 | 95704962 | Lore anon is noticed donating to further new 70B model development |
2023-08-31 | #56278 | SourceGraph prepares to fight tempmail signups |
2023-08-31 | 95707791 | SourceGraph notifies users about the intrusion into their admin panel |
2023-08-31 | 95709009 | CncAnon explains the issue with the way SourceGraph quota hack was handled |
2023-08-31 | 95710938 | Lore anon elaborates on the cost of running 70B model inference for CharacterHub |
2023-08-31 | 95711740 | local models proxies are getting spammed |
2023-08-31 | 95713770 | chinese search engine Baidu makes its llm Ernie Bot fully available to the public |
2023-08-31 | 95720962 | dody is gone |
2023-08-31 | #600 | Agnai celebrates 600th ticket |
September 2023
When | Where | What |
---|---|---|
2023-09-01 | 95727588 | Proton Mail requires an existing email address for signup confirmation |
2023-09-01 | 95729427 | SourceGraph disables email-based signup |
2023-09-01 | 95738299 | janitorai.com characters ripping script shared |
2023-09-01 | 95740376 | MikuForbinDev reports beta launch of his app |
2023-09-01 | 95740798 | surprisingly looks like a product shill by a company rather than by a small dev or devteam |
2023-09-01 | poe.com gets threading support, readds custom temperature setting | |
2023-09-02 | 95749575 | CharacterHub adds "followers" which are not hidden |
2023-09-02 | 95762406 | Openrouter gets OAI access back with enhanced moderation output |
2023-09-04 | 95791992 | CharacterHub declutters download and export buttons |
2023-09-05 | 95813088 | second Lablab.ai's Anthropic AI Hackathon noticed too late |
2023-09-05 | 95815756 | c.ai plans to reopen the discord server |
2023-09-05 | 95821911 | CncAnon fixes SourceGraph adapter but random account limits and security efforts make him want to stop |
2023-09-06 | 95837202 | venus.chub.ai introduces leaderboards |
2023-09-06 | 1148932036795580487 | Pygmalion 2 |
2023-09-07 | 95852576 | Timeline extension for Silly is noticed |
2023-09-07 | 95856101 | Anthropic introduces a subscription for claude.ai |
2023-09-07 | 95857710 | anon shares a graph showcasing loss of interest towards c.ai |
2023-09-07 | 95858668 | Forefront dismissed as a source of free limited Claude Instant and GPT3.5 Turbo |
2023-09-07 | 95860647 | another report of the decrease in c.ai's web traffic |
2023-09-07 | 95862178 | chub-archive anon adds 70k scraped janitorai.com characters |
2023-09-08 | 95867721 | anon confirms Claude 2 being worse everywhere it counts compared to Claude 1 |
2023-09-08 | 95877047 | OAI teases something good for the developers but "no gpt-5 or 4.5 or anything like that" |
2023-09-08 | 95877538 | there is some speculation regarding Nvidia pulling a scam to inflate own value |
2023-09-08 | 95877667 | second Lablab.ai's Anthropic AI Hackathon is effectively cancelled |
2023-09-08 | 95879718 | anon calculates costs for GPT3.5 Turbo and GPT4 models inference at full context |
2023-09-08 | 95882185 | aisekai.ai shilling continues despite being called out on hiding the exact model used |
2023-09-08 | 95883262 | moemate.io shill is called out by Lore anon |
2023-09-09 | 95886414 | a fullfilled gamedev makes appearance |
2023-09-09 | 95886768 | more pointless c.ai updates |
2023-09-09 | 95901124 | llm benchmarking site shared, barely relevant not being dedicated to rp |
2023-09-10 | 95908107 | anon reminds about the official way to get instant GPT4 access from OAI |
2023-09-10 | 95907864 | anon shares a huge write-up on "serious" prompting |
2023-09-10 | 95919358 | Mancer removing free trial credits as an offering and from existing accounts as well is brought to attention |
2023-09-10 | 587186 | Satania Buddy resurrected |
2023-09-11 | 95929510 | once again news about Colab banning SD users are spooky for llm users |
2023-09-11 | 95929524 | rumors about Meta beginning to work on Llama 3 next year |
2023-09-11 | 95936036 | new generation of "enhanced safety filters" for Claude API reported |
2023-09-11 | 95935479 | Lore anon releases 70B model, high hopes unfulfilled |
2023-09-12 | 95944163 | GPT4 Summarization with chain of density prompting |
2023-09-12 | 95946847 | Medusa: simple framework for accelerating llm generation with multiple decoding heads |
2023-09-12 | 95947705 | chub-archive anon switches to vllm-gptq for batched inference of his local proxy |
2023-09-12 | 95950296 | Phi1.5 is 1.3B model stronger than top 7B models |
2023-09-12 | 95950716 | anon reminds how true "We have no moat, and neither does OpenAI" leak was |
2023-09-12 | 95951464 | venus.chub.ai gets "report stolen bot" button, regular CharacterHub soon to follow |
2023-09-12 | 95955435 | c.ai reopens the discord server |
2023-09-13 | 95977528 | exllama2 reported to be FAST |
2023-09-13 | 95980807 | in his second post anon explains the developed Silly extension adding MMD models with expressions |
2023-09-14 | 95986825 | anon writes an ethnography on main AI community |
2023-09-14 | 95988202 | c.ai adds "edit message" and "generate greeting" features for paid users |
2023-09-14 | 95993059 | subscription based access to a finetune of llama 2 70b for male dating / romance advice shilled |
2023-09-15 | 96011881 | subscription based Claude Pro on claude.ai still has prompt limit |
2023-09-15 | 96016752 | new WebP related vulnerability discovered |
2023-09-16 | 96027176 | mikudev shills finally released nagatoro bot, shares plans for "generative visual novels" platform |
2023-09-16 | 96030864 | anon speculates about reasons for Colab tolerating llama 2 finetunes |
2023-09-16 | 96034059 | Silly TTS extension demoed |
2023-09-17 | 96058664 | anon gets an email from Samantha Cole, 404 Media journalist |
2023-09-18 | 96080720 | llama.cpp batched inference update makes it even faster |
2023-09-18 | 1153338196084666439 | Pyg2 70B has finished training, pygmalion.chat postponed yet again |
2023-09-19 | 96089796 | GPT3.5 Turbo-instruct reported to be good |
2023-09-19 | 96099877 | news about Google Bard update, still irrelevant |
2023-09-20 | StableDiffusionXL added on poe.com | |
2023-09-20 | 96102656 | AI bad because men happy; happy women ignored |
2023-09-20 | 96104060 | The proxy gets GPT3.5 Turbo-instruct and Google PaLM support |
2023-09-20 | 96112639 | KoboldCpp 1.44 update shilled, added features to be more useful as a frontend |
2023-09-20 | 96117433 | innovative botmaking technique discovered, the use of randomized emojis |
2023-09-21 | 96126620 | chub-archive anon adds status page |
2023-09-21 | 96130465 | another demo of Silly TTS extension, singing with voice cloning |
2023-09-21 | 96132990 | 404 Media article "DIY Chatbots Unleash Large Language Models’ Repressed Sexuality" published |
2023-09-22 | 96142847 | c.ai adds "generate reply" feature and it talks exclusively in bad zoomer-speak |
2023-09-23 | 96165230 | Silly update glows |
2023-09-24 | 96187191 | Xwin 70B model released, beats GPT3.5 Turbo and sometimes even GPT4 |
2023-09-24 | 96194563 | free GPT4 and other llms source dismissed as unusable |
2023-09-24 | 96195548 | emails from candy.ai, considered spam |
2023-09-25 | 96208511 | Amazon invests $4 billion in Anthropic |
2023-09-25 | 96210792 | GPT4Vision System Card reveals extreme censoring of the model |
2023-09-25 | 96215302 | current state of the shadowbans at c.ai community forum |
2023-09-25 | 96219047 | speculations about GPT5 being AGI with next year release |
2023-09-26 | 96220482 | news about Meta's chatbots |
2023-09-26 | 96232287 | chinese Qwen 14B model is promising |
2023-09-26 | 96238920 | Mozilla commits $30M to build a trustworthy AI ecosystem |
2023-09-27 | 96256581 | Mistral 7B v0.1 released, outperforms llama 2 13B |
2023-09-28 | 96264566 | ChatGPT can now browse the internet, no longer limited to data before 2021 |
2023-09-28 | 96267204 | Slack's Claude now issues 24 hour bans instead of permanent |
2023-09-28 | 96273171 | chinese source of paid proxies brought up |
2023-09-28 | 96279446 | CncAnon provides a new source of Claude, Amazon Web Services (AWS) |
2023-09-28 | 96280416 | CncAnon reveals actual Claude context size to be 18K instead of 100K |
2023-09-29 | 1157166185180315658 | the new Mikugg platform is available at the main domain |
2023-09-29 | ChatNBX Plus | chat.nbox.ai introduces subscription, severely limits already weak free tier |
October 2023
When | Where | What |
---|---|---|
2023-10-01 | 96329673 | chub-archive anon upgrades local proxy service |
2023-10-02 | 96365111 | because of the deal Anthropic is speculated to become Amazon's bitch |
2023-10-02 | 96369365 | GPT3.5 Turbo-instruct can play chess because of being completion model instead of chat model |
2023-10-04 | 96416989 | new way of controlling llm outputs in the name of ethics |
2023-10-05 | 96444415 | Anthropic is considered the least open source AI friendly key company |
2023-10-05 | 96449699 | cards can now include a character specific author's note in Silly |
2023-10-06 | 96461860 | mikudev teases miku.gg platform with a free 20B model running |
2023-10-06 | 96463116 | lightweight mikupad frontend intended for local backends shared |
2023-10-06 | 96470678 | Claude AWS proxy can be logged by Amazon |
2023-10-06 | 96472131 | smart people are close to figuring out how exactly AI models work |
2023-10-06 | 96476909 | kinktober botmaking event by some other community discovered |
2023-10-07 | 96492700 | Lore anon has woken up from a literal coma, delegates responsibilities to Data anon |
2023-10-07 | 96494804 | Data anon !!2d50io56s7+ appears, promises to work |
2023-10-07 | 96503010 | ollama.ai by a guy who got famous for running a script made by /lmg/ is mentioned |
2023-10-07 | 96504127 | reviews posting date bug on CharacterHub is vanquished after months of suffering |
2023-10-08 | 96519748 | Data anon reports progress |
2023-10-08 | 96529189 | Silly update 1.10.5 |
2023-10-09 | 96532618 | c.ai will release a much more powerful version of the AI "by the end of the year" |
2023-10-09 | 96534794 | former sturdychan.help users create a subreddit |
2023-10-09 | poe.com gets GPT3.5 Turbo-based websearch bot and GPT3.5 Turbo-instruct bot | |
2023-10-10 | 1697006641211 | pygdevs move but not really |
2023-10-10 | 96569447 | Data anon reports progress, CharacterHub email and password requirements get stricter |
2023-10-10 | 96576118 | multiple CharacterHub accounts get bruteforced and leaked |
2023-10-11 | 96585449 | Data anon reports progress |
2023-10-11 | 96605735 | c.ai introduces multi-user rooms for subscribers |
2023-10-12 | 96607924 | the ultimate botmaking guide is published |
2023-10-12 | 96615546 | CharacterHub 1 star reviews botting is exposed |
2023-10-12 | 96622818 | Maid app is shared by the dev, running llms on Android devices |
2023-10-13 | 96632973 | HF is down |
2023-10-13 | 96633134 | anon shares a nice guide on finetuning a LoRA on CPU using llama.cpp |
2023-10-13 | 96647794 | desu shares a nice desu |
2023-10-14 | 431326 | OAI has broken the contract with developers to leave still-working gpt-3.5-turbo-0301 alone |
2023-10-15 | 96687909 | Maid app gets GGUFed models support |
2023-10-16 | 96707034 | c.ai reportedly has fucked up yet another update |
2023-10-16 | 96716622 | anon researches dynamic temperature sampling |
2023-10-18 | 96760112 | Nexon Games works on "MomoTalk AI Conversation" for Blue Archive |
2023-10-18 | 96762218 | astroturfed meme fuel aisekai.ai is experiencing issues with comical explanation |
2023-10-19 | 96780077 | various money milking AI chat platforms are laughed at |
2023-10-19 | 96786815 | waifuism future through the eyes of AR users |
2023-10-21 | 96812877 | anon bakes |
2023-10-21 | 96814421 | HF OAI key scraping script shared |
2023-10-21 | 96831401 | /aitg/ AI Tools General is born and dies |
2023-10-21 | 96819280 | CharacterHub reviews botting continues |
2023-10-21 | 96822891 | CharacterHub email check oversight exposed |
2023-10-21 | 96825746 | CharacterHub email check oversight exposed again after being fixed |
2023-10-21 | 96827571 | Lore anon reports about switching to using a classifier model on new reviews and creating a list of spammers |
2023-10-22 | 96851206 | CharacterHub reviews botting continues |
2023-10-23 | 96864426 | more spam emails, this time from myshell.ai |
2023-10-23 | 96867423 | Maid app updated |
2023-10-23 | 96869352 | AWS Claude retardation is attributed to Claude 2 vs Claude 1 |
2023-10-24 | 96875003 | chub-archive anon upgrades local proxy service, laments about too much cards |
2023-10-24 | 96886743 | CharacterHub reviews spammer continues |
2023-10-25 | 96908381 | new chinese llms score high against GPT, explained by benchmarking data abuse |
2023-10-26 | creator monetization at poe.com, the ability for bot creators to generate revenue (for Poe) | |
2023-10-26 | 96916827 | bot3.ai platform is shilled on CharacterHub, stealing characters |
2023-10-26 | 96923682 | AI girlfriends fearmongering explained by 'muh economy' |
2023-10-26 | 96927032 | aisekai.ai runs out of money and goes down |
2023-10-27 | 96934244 | Boston dynamics adds GPT4 to Spot |
2023-10-27 | 96945498 | anon asks the right questions |
2023-10-28 | 96956499 | /aicg/ is spammed by the old posts |
2023-10-28 | 96963173 | PygWiki knowledge database shared |
2023-10-28 | 96968398 | anon shares custom CSS for Silly to make it resemble c.ai |
2023-10-29 | 96971199 | c.ai adds Persona field |
2023-09-25 | 96980225 | Google invests $2 billion in Anthropic |
2023-10-29 | 97056477 | LeCun tells it like it is, as usual, this time about corporate lobbying |
2023-10-30 | 96999727 | President Biden issues executive order on safe, secure, and trustworthy Artificial Intelligence |
2023-10-30 | 96996599 | /aicg/ spam is cleared out |
2023-10-30 | 96997601 | /aicg/ temporarily purged from /g/; reason - heavy /g/ spam |
2023-10-30 | 96999009 | Microsoft paper published claiming GPT3.5 Turbo is 20B |
2023-10-31 | 97018489 | /aicg/ spam is cleared out again |
November 2023
When | Where | What |
---|---|---|
2023-11-01 | 97036103 | aisekai.ai announces NSFW filter |
2023-11-01 | 97042874 | Maid app updated, now with character JSON import in own format |
2023-11-02 | 97055555 | myshell.ai spam turns out to be real, cleared out from CharacterHub |
2023-11-03 | 97070243 | anon writes a bot to prompt SD from Silly in a better way |
2023-11-03 | 97076562 | xAI from Musk is almost released for testers |
2023-11-03 | 97081937 | new OAI rate limits |
2023-11-03 | 97082311 | OAI leak promises price reduction and improvements for GPT4 |
2023-11-04 | 97089797 | c.ai is still popular despite all the fuckups |
2023-11-04 | Grok | xAI from Musk is unveiled... |
2023-11-04 | 97092326 | ... and appears not censored... |
2023-11-04 | 97092376 | ... but probably still is. |
2023-11-04 | 97095210 | Silly supports comments inside of character cards |
2023-11-04 | 97097740 | CharacterHub filters reviews with an error "User too new." |
2023-11-04 | 97099408 | Silly gets minimum activations setting for lorebook entries |
2023-11-04 | 97102523 | subscribers will get immediate access to Grok after beta, it will use realtime Twitter data |
2023-11-04 | Netrunner | anonymous network for publishing and distributing open source llm-related stuff, Shoggoth, announced |
2023-11-05 | 97104530 | c.ai replies editing doesn't help to circumvent the filter |
2023-11-05 | 97107238 | redditor cries because aisekai.ai expects botmakers to not mention cunny even outside the site |
2023-11-05 | 97112779 | venus.chub.ai gets HTML and quote colors support |
2023-11-05 | 97118479 | venus.chub.ai has public logs page now |
2023-11-06 | 97137218 | OAI leak promises GPT4 Turbo with 128k context |
2023-11-06 | 97140355 | GPT4 Turbo is cheaper |
2023-11-06 | 97140937 | OAI offers chatbots creation... |
2023-11-06 | 97141883 | ... and new multimodal capabilities in the platform, including vision, image creation, TTS |
2023-11-06 | 97142097 | jennyfag femanon can host locally now, offers ass, yet still no promised card since old times |
2023-11-07 | 97145620 | Elevenlabs offers Turbo TTS |
2023-11-08 | 97174372 | DALL-E 3 frontend teased... |
2023-11-08 | 97175304 | ... and released, GPT4 generated |
2023-11-08 | 97180671 | chinese hustlers exploit Clewd |
2023-11-08 | 97180685 | DALL-E 3 support is added to The proxy |
2023-11-09 | 97192382 | Slack's Claude is now only available for Enterprise workspaces |
2023-11-09 | 97193459 | CharacterHub filters reviews with an error "You are banned from rating." |
2023-11-09 | 97197474 | Claude's longer context is better than GPT4's |
2023-11-10 | 97206261 | janitorai.com haters make janitorai.me with defs and tinder |
2023-11-10 | 97211380 | proxy-supporting frontend for work tasks suggested |
2023-11-10 | 97211661 | GPT4 Turbo's real context length speculated to be 64k |
2023-11-10 | 97218600 | DALL-E 3 frontend updated; price estimations, batch generations, automatic image saving, metadata support |
2023-11-11 | 40523640 | you can now edit user's and char's replies on c.ai without the subscription, chat2 links only |
2023-11-11 | 97231756 | consolidated stats of a proxy, more than $250k spent in a month |
2023-11-11 | 97231865 | poe.com subscribers get DALL-E 3 access |
2023-11-11 | 97241526 | venus.chub.ai gets dubious UI update |
2023-11-12 | 634541 | continued Satania Buddy development is picking up pace |
2023-11-12 | 97251222 | anon can run Pokemon inside Silly |
2023-11-12 | 97253153 | DALL-E 3 frontend updated; metadata viewer, better looks, filter type detection, cancel button, "jailbreak" |
2023-11-13 | 97263476 | The proxy gets DALL-E 3 support |
2023-11-13 | 97271644 | GPT4 Turbo updated, comes to ChatGPT first |
2023-11-14 | 97286210 | c.ai indeed being Google was officially confirmed |
2023-11-15 | 97313437 | venus.chub.ai can rate with stars now |
2023-11-15 | 97313629 | aisekai.ai cries about becoming filtered |
2023-11-16 | 97326941 | antispitefagging measures implemented for The proxy |
2023-11-16 | 97331061 | some AI girlfriend app dev shuts it down because of eavesdropping too much |
2023-11-16 | 97335182 | a matrixfag collects gptisms and claudeisms |
2023-11-17 | 97347296 | CustomGPT users worry about leaking their precious prompts |
2023-11-17 | 97349611 | Grok API support begins rolling out for select partners |
2023-11-17 | 97350895 | c.ai implements user personas |
2023-11-17 | 97351255 | aisekai.ai refunds donations |
2023-11-17 | 97357538 | Sam Altman will depart as CEO and leave the board of directors of OpenAI |
2023-11-17 | 97357988 | info about OAI switching to pre-purchased credits API use system |
2023-11-18 | 97364080 | speculations about profit versus nonprofit... |
2023-11-18 | 97366643 | ... or AI safety being the reason behind OAI trouble |
2023-11-18 | 38310831 | theory: OAI's new release had a serious security issue, leaked a bunch of data, and it wasn't disclosed |
2023-11-18 | 97367157 | OAI's prepaid billing system confirmed but seems valid only for the smallest clients |
2023-11-18 | 97374143 | anon shares cutting edge TTS impressions, human brain is at fault |
2023-11-18 | 97379714 | chub-archive anon reports archive hardware upgrades, local-llm-server project updates |
2023-11-19 | 97384362 | for some reason OAI wants Sam Altman back |
2023-11-19 | 97384551 | CharacterHub followers are as meaningless as ever, accounts are botted with scripts |
2023-11-19 | 97389580 | new Silly extension uses GPT4V, allowing bots to see your gameplay via another extension |
2023-11-19 | 97389933 | Meta disbands its Responsible AI team |
2023-11-19 | 97394481 | Maid app updated |
2023-11-19 | 97401414 | CharacterHub reviews filtering criteria somewhat explained |
2023-11-20 | 97403364 | anon makes a page with DALL-E 3 images pulled from proxies |
2023-11-20 | 97406964 | AI ethicists are mentally ill, OAI's are no exception |
2023-11-20 | 97408740 | GPT4 can’t be purchased now |
2023-11-20 | 8:58 PM PST | Sam Altman won’t return as CEO of OpenAI |
2023-11-20 | 97416493 | many OAI employees threaten to quit following the news |
2023-11-20 | 7:53 AM UTC | Sam Altman moves to Microsoft |
2023-11-20 | 9:34 PM UTC | poe.com might be behind OAI falling apart |
2023-11-21 | 97428776 | OpenAI’s board approached Anthropic about merger |
2023-11-21 | 97431672 | Brave browser integrates free and paid llama 2 |
2023-11-21 | 97438688 | chinese found to be still using Clewd |
2023-11-21 | claude-2-1 | Claude 2.1 with 200k context is out |
2023-11-21 | 97442720 | Stable Video Diffusion released |
2023-11-22 | 97447684 | Claude 2.1's real context length speculated to be the same 18k since it's the same model, only "unlocked" |
2023-11-22 | 1:55 AM UTC | if the company was destroyed, OAI person said, its board's mission would be fulfilled |
2023-11-22 | 97453113 | regretfully Sam Altman to return as CEO of OpenAI |
2023-11-22 | 97455584 | NSFW c.ai alternatives list shilled |
2023-11-22 | Q* Q-Star | another crackpot theory about OAI's ethicists shitting themselves because of the impending AGI |
2023-11-23 | 97470732 | ChatGPT with voice is now available to all free users |
2023-11-23 | 97476673 | scary Q-451-921 - QUALIA OAI leak posted |
2023-11-24 | 97490884 | Silly introduces STscript scripting language to expand its functionality |
2023-11-24 | 97493946 | c.ai increased the greeting length limit from 500 to 2048 w i t h o u t t e l l i n g y o u |
2023-11-24 | 97498517 | anon provides a local model proxy under the guise of auto jailbreak, but anons notice, all in good fun |
2023-11-24 | 97500465 | 11 fucking months later c.ai admits that the """""""""""""""""""""temporary""""""""""""""""""""" freezing was permanent |
2023-11-24 | 97504076 | a better NSFW c.ai alternatives list shilled |
2023-11-25 | 97519410 | LeCun tells it like it is, as usual, this time about Q* Q-Star |
2023-11-25 | 97522247 | current way to continue using Poe in Silly provided, reported as still viable (with the subscription) |
2023-11-25 | 97523768 | anon demonstrates GPT4V use for automated cftf character creation |
2023-11-26 | 97538692 | chub-archive anon finally shares it via torrent |
2023-11-26 | 97540476 | c.ai provides a way to fork characters (supposedly without the training) |
2023-11-26 | 97554496 | an example of STscript use shared |
2023-11-27 | 97565706 | Amica frontend shared, ignored |
2023-11-27 | 97569150 | moemate.io now is c.ai alternative with free unlimited 13B & imagegen, and unlimited GPT and Claude for $19.99/m |
2023-11-28 | 97581373 | Data anon reports plans to move off of GitLab, allow users to edit project titles and usernames |
2023-11-28 | 97583622 | japanese character card finally noticed on CharacterHub |
2023-11-28 | 97586679 | c.ai added Character Profiles with Example Conversations, and Suggested Starters |
2023-11-28 | 97589137 | a site for collecting gossip about OAI's board is shared |
2023-11-28 | sdxl-turbo | SDXL Turbo, real-time text-to-image generation model released |
2023-11-29 | 97598390 | a character card backed by a custom website shilled, the most (you)ed character shill in /aicg/ history (56) |
2023-11-29 | 97604838 | official private confidential discord server for top c.ai creators exposed |
2023-11-30 | 97610605 | redditors can't handle AI characters |
2023-11-30 | 97611059 | NAI's API keys are being scraped as well |
2023-11-30 | 97625710 | SourceGraph reported still usable |
2023-11-30 | 97627668 | new chinese llms score high against GPT, explained by benchmarking data abuse |
2023-11-30 | joi-extension | JOI extension for Silly with TTS support github release, Klingon suggested as the language of choice |
December 2023
When | Where | What |
---|---|---|
2023-12-02 | 97647457 | AetherRoom coming at you with one the gayest devlogs of all time |
2023-12-02 | 97653122 | GPT4 gives longer responses when offered a tip |
2023-12-02 | 97662887 | llamafile by Mozilla lets you distribute and run LLMs with a single file |
2023-12-03 | 7:03 AM UTC | Musk slowly grants Grok access to premium+ users |
2023-12-04 | 97686618 | anon finds a jailbreak for Bing utilizing base64 and science |
2023-12-04 | 97687473 | AI x Robotics is progressing incredibly fast |
2023-12-04 | 97690321 | llama dev teases a breakthrough, but no specifics at all |
2023-12-04 | 97691621 | The proxy gets Azure support |
2023-12-04 | 97694587 | "uncensored" spicychat.ai bans incest because that's what respecting diversity means |
2023-12-04 | 97701588 | Azure allows to list other deployments/models for a key |
2023-12-04 | 97702695 | anon tries to estimate the losses of some company due to a proxy |
2023-12-04 | 6:09 PM UTC | Claude 2.1 is so ruined, Anthropic adds a selector for "the creatives" with the subscription on claude.ai |
2023-12-05 | 97707332 | creating videos from a single image with pose guidance soon |
2023-12-05 | 97713182 | Clipboard Conqueror shilled, a copilot alternative bridging KoboldCPP with the clipboard contents |
2023-12-05 | 97720091 | AI Alliance formed to open source AI models; Meta, IBM, Intel, NASA on board |
2023-12-06 | 97724681 | official private confidential discord server for top c.ai creators was closed |
2023-12-06 | 97729451 | chub-archive includes booru cards now |
2023-12-06 | 97737835 | Welcome to the Gemini era |
2023-12-06 | 4 min read | Anthropic tells how prefilling solves Claude rejection issues and reunlocks 200k, what a clown world |
2023-12-06 | 147 | mikudev shares benchmarks of the Aphrodite-engine on two 2x3090 rigs |
2023-12-08 | 97770979 | chub-archive anon reports search upgrade |
2023-12-08 | 97774881 | NAI notices the abuse of the free image generations, implements rate limits for abusers |
2023-12-08 | 3:44 PM UTC | Mistral mixtral-8x7b Mixture of Experts out |
2023-12-09 | 97791632 | pollinations.ai for free image generation via url is brought to attention |
2023-12-10 | 97803533 | OAI acknowledges GPT4 is becoming worse |
2023-12-10 | 97806200 | anon roleplays with Gemini via Bard |
2023-12-10 | #1550 | Silly sanitizes <style> tags |
2023-12-11 | 97819635 | chub-archive anon reports proxy stats page update |
2023-12-11 | La plateforme | Mistral AI explains Mixtral and teases Mistral-medium |
2023-12-11 | 97832930 | supposedly $240k spent in a week on a proxy |
2023-12-12 | 97840756 | numberfags suffer as they should |
2023-12-12 | 97841218 | Secret Santa event proposed |
2023-12-13 | 97862466 | anon tells about Silly's Idle extension, promises to release advanced version |
2023-12-13 | 97863063 | Mixtral available on poe.com, proves itself smart |
2023-12-13 | 97868609 | Mixtral available on openrouter.ai, but automatic paid fallback renders it unusable for free |
2023-12-13 | 97870663 | Gemini Pro API is available |
2023-12-14 | 97876093 | Data anon dispels rumors about using AI to remove reported bots |
2023-12-14 | 97879204 | open source frontend RealChar shilled |
2023-12-14 | 97889892 | GPT4.5 rumor, considered fake |
2023-12-14 | 97881662 | Gemini has a hidden filter for cunny |
2023-12-14 | 97882441 | The proxy has Gemini support |
2023-12-14 | 97891257 | c.ai adds "proactive messaging" and notifications |
2023-12-14 | poe.com adds knowledge bases, new imagegen | |
2023-12-15 | 97896452 | anon collects Mixtral examples proving it doing well |
2023-12-15 | 97904735 | some Bing Chat users get Notebook mode with up to 18k input |
2023-12-15 | 97911197 | Tavern character hosting still exists, hosts 5290 cards now |
2023-12-16 | 97915763 | muah.ai astroturfing shill spotted |
2023-12-16 | 97920262 | yet another ranking puts Mixtral above GPT3.5 Turbo and even Gemini Pro |
2023-12-16 | 97925439 | new Tay noticed on gab |
2023-12-16 | 97927981 | 4channel no more? |
2023-12-17 | 97941141 | anon wishes to create another CharacterHub archive |
2023-12-19 | 97971696 | freellamas anon is back, hosting Mixtral-3.75bit |
2023-12-19 | 97971937 | c.ai down again, hope gone long ago |
2023-12-19 | v1/messages | Anthropic Claude Messages API beta |
2023-12-20 | 97988535 | anon reports Mistral API keys available for scraping |
2023-12-20 | 97991544 | Mistral API proxies appear |
2023-12-20 | N/A | c.ai swipes at /chat are cut off now like at /chat2; image recognition doesn't work at all (dates unknown) |
2023-12-21 | 98004894 | spicychat.ai forces users to doxx themselves if they're suspected of being underage |
2023-12-21 | 98005448 | anon claims to have extracted the 3d model of some waifu from some paid app, shilled before |
2023-12-21 | 98009966 | anon shares IPIP-BFFM test from prehistoric 1992, hardly relevant today |
2023-12-21 | 1703149242999 | Aphrodite-engine beats ExLlamaV2, making it the fastest backend for both single and batch requests |
2023-12-22 | 98029394 | Claude 2.1 is worse than Claude 2.0 and even Mixtral, Gemini Pro is worse than GPT3.5 Turbo |
2023-12-22 | 98030806 | true value of proxies is revealed, ignored same as during all the preceding months |
2023-12-23 | 98042570 | perchance.org platform for random text generators shared |
2023-12-23 | 98045828 | new online character cards editor shared |
2023-12-23 | 98051728 | anon puts up /aicg/ Christmas neocities page |
2023-12-24 | Miku server | Mikugg undergoes code refactor, gets conversation tree, soon multibot; 70B and TTS are now paid |
2023-12-24 | 98063051 | a drawfag with a shelved comic makes appearance |
2023-12-24 | 98068821 | Gemini 2 testing rumor, no source |
2023-12-24 | 98073495 | a source of free trial Mixtral together.ai is mentioned with the complete explanation |
2023-12-25 | 98077706 | CharacterHub gets ads, explained by a free tier to be offered later |
2023-12-25 | 98080446 | Silly soon to get multi-user chats support... |
2023-12-25 | 98080952 | ... "soon" meaning right away |
2023-12-25 | 98082322 | a dev shills his fork of KoalaClient frontend with Mistral API support |
2023-12-25 | 98087984 | The proxy gets Mistral API support |
2023-12-26 | 98090116 | booru.plus is down |
2023-12-27 | 98106429 | Secret Santa event ends |
2023-12-27 | 98109067 | anon showcases STscript use |
2023-12-27 | 98111262 | magic rocks... make story? Grug no understand |
2023-12-27 | 98111432 | AWS Claude 1.3 rumor, no source |
2023-12-27 | 98111924 | anon explains logit bias |
2023-12-27 | 98116016 | anon reminds that GPT3.5 Turbo being 20B claim was retracted |
2023-12-27 | 98117106 | c.ai's "Character AI Wrapped" is exclusive for subscribers |
2023-12-27 | 98118614 | a rumor about the leaked Azure proxy key, no source |
2023-12-28 | 98130657 | local 34B finetune is less lossy than Claude's initial 200k context assessment |
2023-12-28 | 98131417 | NYT vs OAI lawsuit shows OAI is profit-driven and closed |
2023-12-28 | 98133927 | proxies can offer up to 13 models, or maybe more |
2023-12-28 | 98137431 | anon emulates /aicg/ as a card |
2023-12-28 | 1703759821507 | pygdevs work on a new quant method based on QuIP# |
2023-12-29 | #750 | Agnai adds openai chat/claude setting to opt out of appending {replyAs}: |
2023-12-29 | 98140276 | Gemini Pro found to be worse than GPT3.5 Turbo |
2023-12-29 | 98139519 | anon showcases the use of color gradients |
2023-12-29 | 98151155 | AI fearmongering in form of environmental concerns |
2023-12-29 | 98153834 | anon wants to work on an alt frontend, but it's just empty words... |
2023-12-30 | 98158024 | ... or maybe not, since he's a twink |
2023-12-30 | 98162109 | botmakers might be stealing /aids/'s prompts |
2023-12-30 | 98163300 | more rumors about japanese companies planning to use AI for localization |
2023-12-31 | 98176067 | unpopular esl anon is making a video |
2023-12-31 | 98180233 | keys without DALL-E 3 lead to rate limiting because of proxy retries |
2023-12-31 | 98185201 | Secret Santa event ends for real |
2024
January 2024
When | Where | What |
---|---|---|
2024-01-01 | 98192836 | anon probes about a utilities server for embedding into chats |
2024-01-01 | 98199283 | NYT vs OAI proceedings highlight some specifics of OAI's dataset |
2024-01-02 | 98208280 | anon continues to claim to be working on a new frontend out of dislike of Silly |
2024-01-02 | 98208346 | somebody is making a Silly extension that combines Summary + Smart Context + Objectives |
2024-01-02 | 98209588 | songs generating app with free trial app.suno.ai is brought to attention, enjoyed by all |
2024-01-02 | 98212873 | anon estimates "over half a million in GPT4 usage by a single proxy over a month or two", asks not to quote him on this |
2024-01-02 | 98219089 | more email spam reported, this time from flowgpt.com |
2024-01-02 | 98219919 | anon probes about a paid phone frontend app |
2024-01-03 | 98230735 | anon explores GPT4V's safety prompt |
2024-01-03 | 98231993 | anon provides kudos earning calculations for Horde |
2024-01-03 | 98247066 | for coding tasks begging, scaring or bribing GPT4 Turbo is ineffective |
2024-01-04 | 98248861 | text-davinci-003 is no more |
2024-01-04 | 98250679 | GPT4's moderation explored in context of Protected Material text being considered harmful |
2024-01-04 | 98255921 | some of the largest llms quizzed and arranged on the Political Compass by trackingai.org |
2024-01-05 | 98260723 | 5 months old publication about a vulnerability in OAI API keys generation discovered |
2024-01-05 | 98265649 | news article about boyfriend.myanima.ai shared, both laughed at |
2024-01-05 | 98267419 | anon provides an example of the clever use of Claude for language learning |
2024-01-05 | 98267211 | heavy /aicg/ spam wave purged |
2024-01-05 | 98269517 | paid haven-chat.com shilled by the dev, justifiably ignored |
2024-01-06 | 98277103 | OAI GPT Store launching next week |
2024-01-06 | 98277246 | suspicious DALL-E 3 online UI shared, considered IP logger and keylogger |
2024-01-06 | 98285801 | "alt frontend developing" anon considers multi-user chat interesting, but excessive |
2024-01-06 | 98288233 | CharacterHub resets bio in user profiles |
2024-01-07 | 98296681 | anon from the future brings up chat-instruct finetunes |
2024-01-07 | 98298383 | an ominous post, but only in hindsight |
2024-01-07 | 98301671 | anon claims OAI API keys generation vulnerability is still viable |
2024-01-07 | 98306071 | OAI presents itself as caring only about the safety |
2024-01-07 | 98307300 | Gemini Ultra rumored to become available via subscription based Bard Advanced |
2024-01-07 | 98308189 | current TTS solutions shared |
2024-01-07 | not migu | Mikugg update, new features, future plans |
2024-01-08 | 98313874 | anon presents his unbloated 250 sloc frontend looking like mikupad, promises to share soon |
2024-01-08 | 98319011 | moemate.io could be abusable |
2024-01-08 | 98319896 | an email from HF threatens to bring down a proxy due to guidelines violations |
2024-01-08 | 98323441 | news article with Lore anon's interview shared, CharacterHub has generated more than $1 million in annualized revenue |
2024-01-08 | 98323551 | OAI deserves the hate |
2024-01-08 | 98323641 | anon claims CharacterHub's gross revenue is much lower, Lore anon's comments were misrepresented |
2024-01-08 | 98324002 | janitorai.com's state analyzed, despite VC funding no real progress made |
2024-01-08 | 98325809 | HF bans proxies left and right |
2024-01-08 | 98326520 | HF alternatives analyzed |
2024-01-08 | 98326659 | possible way to abuse moemate.io shared |
2024-01-08 | 98327089 | OAI deserves all the hate |
2024-01-08 | 98327429 | the most (you)ed bait in /aicg/ history (?) to date (121), immediately followed by (74) and (103) |
2024-01-08 | 98328205 | figure.ai's robot learns to make coffee "by watching humans for 10 hours" |
2024-01-08 | 98328501 | anon claims more than $1 million was drained by the proxies using Azure Microsoft key |
2024-01-08 | 98328943 | new proxies are banned on sight |
2024-01-08 | 98330149 | Lore anon announces ToS changes |
2024-01-08 | 98330154 | anon shares benchmark results for some local models, expresses concerns about future cheating |
2024-01-09 | 98332327 | anon shares an archive with characters jsons scraped from CharacterHub, offers local proxy |
2024-01-09 | 98333068 | Lore anon reports huge traffic surge |
2024-01-09 | 98333125 | CharacterHub scraping anon shares his script, allowing everybody to have an up to date dump of jsons |
2024-01-09 | 98334126 | AI chatbots exposing journalist's post highlight that their goal is fearmongering and public opinion manipulation |
2024-01-09 | 98334212 | OAI deserves all the hurt |
2024-01-09 | 98335074 | chub-archive anon announces proxy tracking update to account for the new hosting solutions |
2024-01-09 | 98335655 | render.com confirmed as unreliable solution |
2024-01-09 | 98335756 | kindroid.ai dev takes a stand |
2024-01-09 | 98336476 | a way to abuse AWS explained |
2024-01-09 | 98336536 | Lore anon confirms being mistreated by the journalist |
2024-01-09 | 98336495 | ToS updated for venus.chub.ai but not for CharacterHub |
2024-01-09 | 98339303 | jennyanon puts out to soothe horny crackheads, takes triple penetration, yet still raped to sleep |
2024-01-09 | 98340280 | venus.chub.ai fix promised to facilitate future jennyanon rape |
2024-01-09 | 98340827 | false key revoking in a proxy explained |
2024-01-09 | 98340810 | high effort botmaking in process showcased |
2024-01-09 | 98343528 | a reasonable theory about proxy takedown shared, pointing at Azure |
2024-01-09 | 98345379 | anon suggests a good idea for a movie |
2024-01-09 | 98356178 | local models can be used to control sex toys via the Buttplug Library |
2024-01-10 | 98356184 | render.com now requires credit card verification |
2024-01-10 | 98365391 | cute but pointless Rabbit R1 is brought to attention |
2024-01-10 | 98367678 | a way to selfhost moemate.io proxy shared |
2024-01-10 | 98370401 | moemate.io's Claude endpoint errors reported |
2024-01-10 | 98372472 | moemate.io's GPT4 Turbo reported to be GPT3.5 Turbo with 4k context |
2024-01-10 | 98373007 | OAI's GPT Store is live, Sam Altman makes a lame joke, shilling a bot |
2024-01-10 | 20230613 | as of today only existing OAI users of 0301 and 0314 will be able to continue using these snapshots |
2024-01-11 | 98375515 | nameless Python + PyQT5 based frontend teased again, release "tomorrow maybe" |
2024-01-11 | 98390243 | anon releases the new frontend named 4rp /fourp/ |
2024-01-12 | 98400329 | after multiple attempts moemate.io devs patch the exploit |
2024-01-12 | 98403206 | a dev shills tryspellbound.com switching between Claude and "another model", defends it |
2024-01-12 | 98405220 | Maid app update, remote APIs and other stuff added, looks great |
2024-01-12 | 98411683 | chub-archive proxy tracking is now byou |
2024-01-13 | 98418475 | moemate.io introduces modules |
2024-01-13 | 98420672 | OAI quietly allows to use its technology for military purposes |
2024-01-13 | 98421630 | jennyanon conjures a black kitten at his front-fucking door, not a robot therefore safe |
2024-01-13 | 98423377 | changes in GPT4's web browsing explained, assumed cost-cutting, quality drop expected |
2024-01-13 | 98423584 | bored anon sends password reset emails to CharacterHub users |
2024-01-13 | 98424509 | more fearmongering from Anthropic, what if llms become deceptive |
2024-01-13 | 98426530 | over 50 million characters on c.ai |
2024-01-13 | 1:48 AM | autonomous AI for conversations demoed |
2024-01-14 | 98436988 | anon wonders why not waste time read lot word despite few word do trick |
2024-01-14 | 98440233 | sakura.fm is brought up, insane prices laughed at |
2024-01-14 | 98445596 | ohmygpt.com is brought up as a potential paid source, looks suspicious |
2024-01-14 | 98446458 | anon shares the finished network diagram of CharacterHub followers, visualizing relationships between botmakers |
2024-01-14 | 98447903 | first (?) OAI """GPT""" bot shilled |
2024-01-15 | 98461632 | hustler sells 100 messages GPT4 discord roleplay package |
2024-01-15 | 98466300 | anon showcases <ruby> elements use for furigana |
2024-01-15 | 98467859 | HF removes the way to learn space owner's email, HF insider's bait about higher ups noticing was truthy |
2024-01-15 | 98468222 | Njalla might be unreliable for proxies as well |
2024-01-15 | 98470337 | OAI adds new safety features in fear for the elections integrity |
2024-01-16 | 98473175 | a series of Q2 quants optimizations for llama.cpp allows to use a 70B model with 24GB VRAM and relative coherency |
2024-01-16 | 98482495 | anon isn't going to share his DALL-E 3 frontend... |
2024-01-16 | 98482600 | ... and shows how he rapes the API using it |
2024-01-16 | 98489887 | yet another VenusAI clone noticed |
2024-01-16 | 98490059 | AI models and hosting providers ranking shilled |
2024-01-17 | 98495075 | a list of free to use AI APIs and sites shared stealthily |
2024-01-17 | 98496916 | OAI allows to view per key usage metrics in the dashboard two weeks prior to the post |
2024-01-17 | 98498811 | venus.chub.ai offers Mixtral finetune to subscribers |
2024-01-17 | 98499659 | llm sampling explained yet again |
2024-01-17 | 98505098 | yodayo.com doesn't cry about becoming filtered, anon speculates janitorai.com will soon follow |
2024-01-17 | 98506817 | OAI is working with the Pentagon explaining recent military applications ban lift |
2024-01-17 | $JHf3OD | Pyg subreddit drama resolved, originally fanmade becomes official, originally official hidden |
2024-01-17 | 98507259 | the page with DALL-E 3 images pulled from proxies is overhauled |
2024-01-17 | 98509840 | OAI boasts about GPT5 capabilities, and as long as it won't help kill gays, other cultures can customize it |
2024-01-18 | 98512949 | some data appearing to be leaked from CharacterHub |
2024-01-18 | 98513715 | voice changing software and a voice models database shared again |
2024-01-18 | 98518242 | the winner of Japan's top literary prize admits she used unedited OAI output |
2024-01-18 | 98524890 | a possible explanation for the latest leak provided |
2024-01-18 | 7:55 PM | Llama-3 is coming! |
2024-01-18 | 98530463 | desu 1 |
2024-01-20 | 98549151 | Bard Advanced is coming |
2024-01-20 | 98549442 | Google planning to add AI image generator directly to Bard |
2024-01-20 | 98550392 | AetherRoom devlog 2 |
2024-01-20 | 98556628 | new approach to spam emails suggested, using bulk newsletters |
2024-01-20 | 98554487 | mikudev spends a couple of bucks to show boobs to 4chan |
2024-01-20 | 98559364 | OAI doesn't allow bots that foster "romantic companionship" in GPT Store |
2024-01-20 | $8rBW0n | c.ai subreddit riots again because violence is filtered hard now |
2024-01-20 | 98565981 | c.ai subreddit has 915k readers |
2024-01-21 | 98567428 | mysentient.ai with $500 weekly subscription tier laughed at |
2024-01-21 | 98570203 | Prisma Anon is back from the dead? |
2024-01-21 | 98575924 | anon finally publishes CharacterHub statistics |
2024-01-21 | 98580483 | anon showcases his use of edited server-personification project, giving it sudo access to his machine |
2024-01-21 | 98581433 | anon explains function calling capabilities of GPT |
2024-01-22 | 98584108 | universal key scraping template provided |
2024-01-22 | 98584835 | implications of widespread AWS credentials scraping discussed, concerns dismissed |
2024-01-22 | 98586157 | anon shares an alternative bot roulette page for CharacterHub |
2024-01-23 | 98606383 | advanced use of STscript demoed |
2024-01-23 | 98609525 | actually uploaded chinese card spotted |
2024-01-23 | 98613688 | anon might host Mixtral-Instruct indefinitely for /aicg/ |
2024-01-23 | AMA | c.ai multi-user rooms for everyone, 3200 limit actually acknowledged |
2024-01-24 | 98616759 | CharacterHub rat infestation begins and continues for many hours... |
2024-01-24 | 98617786 | ... together with rapid followers botting... |
2024-01-24 | 98619065 | suddenly gab.ai, enters invite sending waitlist stage or already kind of accessible |
2024-01-24 | 98621759 | ... leading to cards uploading via API being shut down... |
2024-01-24 | 98622326 | ... stricter email verification requirements for following, reviewing... |
2024-01-24 | 98625906 | ... and regular cards uploading |
2024-01-24 | 98630480 | any non-root AWS keys can be killed by posting them on Github now |
2024-01-24 | 98630668 | an explanation of the way to use FOFA for AWS key scraping shared |
2024-01-25 | 98635049 | The proxy gets logprobs, function calls support |
2024-01-25 | 98640868 | anon tests logprobs, AI makes him an underage zoomer, somehow doesn't get banned for underage posting |
2024-01-25 | 98645448 | Silly adds per-lorebook-entry activation settings |
2024-01-25 | 98649001 | HF and Google partner for open AI collaboration |
2024-01-25 | 98650990 | new GPT4 Turbo snapshot (filtered), GPT3.5 Turbo becomes even cheaper, API keys managing and moderation update |
2024-01-26 | 98662722 | the change in so called "laziness" of GPT4 Turbo explained |
2024-01-26 | 98664931 | a script automating the process of character creation shared |
2024-01-26 | 98666532 | anons reminisce about simWaifu |
2024-01-26 | 98662510 | OAI's way to ensure deterministic outputs is brought to attention |
2024-01-26 | 98668123 | server-personification project updated to allow Google search via SerpApi |
2024-01-26 | 98670935 | Gemini Pro beats GPT4 according to human preference votes |
2024-01-27 | 98676519 | pygmalion.chat, public character repository moderated and SFW only, but anything goes as long as it's private |
2024-01-27 | 98676704 | miqu-1-70b shilled, later claimed to be leaked Mistral-medium |
2024-01-27 | 98676707 | ChatGPT has group chats now |
2024-01-27 | 98685084 | Secret Valentines event proposed |
2024-01-27 | 98685239 | /pbg/ - Proxy Begging General sadly proves to be unviable |
2024-01-28 | 98689059 | chatbots = sexbots = sex dolls = sex robots = rape, "dolls cannot consent and allow for violent fantasies" |
2024-01-28 | 98693447 | a screenshot of @MetaAI with imagegen in Meta's Messenger shared |
2024-01-28 | 98695357 | stolen OAI keys can be used for something other than roleplaying |
2024-01-28 | 98696357 | blind anon outsmarts GPT4V's soft filter |
2024-01-28 | 98700049 | anon teases his proxy switcher script for Silly |
2024-01-28 | 98703456 | anon tries to explain why anyone would try to bot someone else's downloads using premade accounts |
2024-01-28 | 98703923 | LLM Safety Leaderboard shilled, questionable stats |
2024-01-28 | 98704406 | c.ai is soon getting a new website, next.character.ai |
2024-01-28 | 98706271 | old AI-Waifu-Vtuber project is brought to attention |
2024-01-29 | 98708130 | anons look for ways to support miqu leek baiting other than comparing it to the API outputs... |
2024-01-29 | 98708651 | ... at the same time speculating it might be an experimental Stability AI base model leak |
2024-01-29 | 98711628 | anon writes another frontend with his own tavernai card decoder |
2024-01-29 | 98712707 | anon shares his proxy switcher script for Silly |
2024-01-29 | 98717285 | supposed 5 min cooldown between every review on CharacterHub introduced |
2024-01-29 | 98718568 | a Startup Grants Program is brought to attention, promising free OpenAI, Anthropic, etc API access for 6 months |
2024-01-29 | 98722809 | concerns about character card assets being hosted on a remote server shared |
2024-01-29 | 5:00 PM | Code Llama 70B released |
2024-01-29 | 12:16 PM | unlimited context length promised, 400k ctx_len of a llama 2 finetune achieved, code provided, reproducibility unclear |
2024-01-30 | 98726408 | anon teases a private 70k tokens dark fantasy lorebook/characters/world project |
2024-01-30 | 98728482 | c.ai continues rolling out the new login form with email verification instead of the password, locks out tempmail users |
2024-01-30 | 98734688 | one more logprobs viewer demo |
2024-01-30 | 98736206 | Bard changelog leak (?) with imagegen and global Gemini rollout |
2024-01-30 | 98737561 | NAI textgen CFG sampling update |
2024-01-30 | 98739530 | Common Sense Media partners with OAI to protect teens and families, no clear goal in sight |
2024-01-31 | 98756857 | aisekai.ai is shutting down, devs intend to return with a new platform |
2024-01-31 | 98758853 | miqu turns out to be for real, confirmed to be a leak of an early Llama 2 finetune by the Mistral CEO |
2024-01-31 | 98759256 | yet another claim of reaching $1 million of token cost by some proxies |
February 2024
When | Where | What |
---|---|---|
2024-02-01 | 98764451 | OAI scared of biological threats |
2024-02-01 | 98765706 | AI accelerator GroqChip shilled, why |
2024-02-01 | 98776017 | Bard's integrated imagegen demoed, appears to be very cucked... |
2024-02-01 | 98777919 | ... one more screenshot to prove the point |
2024-02-02 | 98784488 | sakura.fm drama due to the filter, who could have known it's the source of all issues |
2024-02-03 | 98803717 | Gemini Ultra 02.07 release claimed by another completely worthless AI leaker |
2024-02-03 | 12033312 | Mikugg releases emotion image generation tool for subscribers, adds autocomplete suggestions |
2024-02-03 | 98804473 | personal recommendations service by anons to anons is reaching new heights |
2024-02-03 | 98804574 | CharacterHub is down for the planned server migration, some minor bugs appear as a result |
2024-02-03 | 98811107 | Gemini Ultra 02.07 release confirmed by a more substantial leak |
2024-02-03 | 98813255 | some of the export buttons noticed to be removed |
2024-02-04 | 98819482 | can namefags finally become self-sufficient? the possibilities are endless... |
2024-02-04 | 98824179 | /lmg/ loves /aicg/, offers janitorai.com cards as proof |
2024-02-04 | 98829698 | AI generated comments summary on YouTube now, hardly relevant |
2024-02-05 | 98839174 | anon teases a price calculator for API models... |
2024-02-05 | 98839180 | new Silly extension, Quick Persona |
2024-02-05 | 98839809 | ... and releases it |
2024-02-06 | 98851670 | AWS Claude v1.3 EOL date is linked again |
2024-02-06 | 98853675 | /aicg/ advertises on 4chan, the original and largest English-language image board |
2024-02-06 | 98860705 | promising Weaver family of llms mentioned, but there's no hope |
2024-02-06 | 98861203 | Gemini Ultra delayed |
2024-02-06 | 98856621 | Llama 3 available subject to NDA, watermarked |
2024-02-08 | 98879888 | retards really do pay $15+ for a custom character card |
2024-02-08 | 98883890 | Gemini Ultra delayed again... |
2024-02-08 | 98883982 | ... with additional confirmation |
2024-02-08 | 98885109 | chatlog aggregator reported lagging behind |
2024-02-08 | 98885993 | current ChatGPT system prompt extracted |
2024-02-08 | 98886086 | Gemini Ultra is coming... |
2024-02-08 | 98887742 | ... and is finally here, subscription based, web chat UI only, with Gemini replacing Bard |
2024-02-08 | 98889454 | Gemini Apps and Gemini API models are differentiated in the docs, suggesting they're separate versions |
2024-02-08 | 98889695 | Gemini Advanced review, called "GPT4 class model", Claude not even mentioned |
2024-02-08 | 98890922 | Smaug-72B finetune is the new best local model according to the latest HF rankings |
2024-02-08 | 98891028 | anon discovers the truth about current AI |
2024-02-08 | 98892225 | Gemini Ultra API could become available in April, or May, or June |
2024-02-08 | 98893243 | first indications of a potential Gemini Advanced to Tavern adapter |
2024-02-08 | 98894530 | new chinese llm is claimed to beat GPT4 Turbo by Chinese-government-affiliated media |
2024-02-09 | 98899526 | Eclair getting railed by several generations of various local models turns anon into a fucking cuck |
2024-02-09 | 98899636 | due to new OAI API permissions management some keys have limited permissions |
2024-02-09 | 98908693 | goody2.ai is brought to attention |
2024-02-10 | 98919589 | anon works on a STscript alternative |
2024-02-11 | 98929091 | Silly gets regex import/export support |
2024-02-11 | 98936155 | Gemini Advanced to Tavern adapter made possible |
2024-02-11 | 98938931 | Gemini Advanced to Tavern adapter released |
2024-02-12 | 98943944 | Gemini imagegen via Bard affected by refusals |
2024-02-12 | 98953649 | anon shares a compiled collection of common Japanese onomatopoeia |
2024-02-12 | #5453 | 1.5 bit quantization; Mixtral8x7B fits into 8.9 GB VRAM |
2024-02-12 | Cascade | Stable Cascade, new text to image model by Stability AI, released |
2024-02-13 | 98963771 | filter dialed down for Gemini, suggesting it's prompt-based |
2024-02-13 | 98969197 | OAI seeks $7 trillion to manufacture its own AI chips |
2024-02-13 | 98969278 | Anthropic data leak remembered |
2024-02-13 | 98969922 | NVIDIA Chat with RTX |
2024-02-13 | 6694432 | LeCun promises video multimodality for llama3 |
2024-02-14 | 98975040 | anon wants to make a CYOA-style sci-fi thingy using STscript, fails |
2024-02-14 | 98978596 | NVIDIA Chat with RTX is lame |
2024-02-14 | 98989824 | another character card backed by a custom website shilled |
2024-02-14 | LWM | Large World Model, a 7B multimodal llm with 1 million tokens context, released |
2024-02-15 | 98993147 | OAI is hiring company detectives |
2024-02-15 | 98994030 | OAI busts accounts associated with state-affiliated threat actors |
2024-02-15 | 98994779 | anon proposes monkey's paw event for April Fools' |
2024-02-15 | 99000225 | Secret Valentines event ends |
2024-02-15 | 99001403 | CharacterHub faces fork spam issue |
2024-02-15 | 99001443 | Gemini 1.5 Pro announced, up to 1 million tokens context but tested up to 10kk, waitlisted |
2024-02-15 | 99001496 | Gemini 1.0 Pro finetuning available |
2024-02-15 | 99001536 | Gemini API model listing indicates Gemini snapshots exist |
2024-02-15 | 99001985 | Gemini 1.5 Pro model card, called a mid-size multimodal model performing at a similar level to 1.0 Ultra |
2024-02-15 | 99002107 | Gemini Ultra is Generally Available via an allowlist on Vertex AI |
2024-02-15 | 99005423 | OAI introduces Sora, their text-to-video model |
2024-02-15 | V-JEPA | Meta takes a step toward LeCun’s vision, releases a a non-generative video prediction model |
2024-02-16 | 99007262 | unknown Google test model jkessinger discovered |
2024-02-16 | 99007496 | OAI's Soma can do text |
2024-02-16 | 99009592 | Test Model jkessinger only turns out to be Gemini 1.0 Pro |
2024-02-16 | 99012328 | anon pays $70/month for Censys to use the nifty features for key scraping and making money |
2024-02-16 | 99012674 | mistral-next model spotted |
2024-02-16 | 99014705 | Nvidia dunks on OAI for begging |
2024-02-16 | 99016995 | mistral-next model confirmed to be the latest prototype model, no details provided |
2024-02-16 | 99022851 | a video prediction model use for gaming showcased |
2024-02-16 | 99023172 | Gemini key checker shared |
2024-02-16 | 99024522 | Anthropic adds a new copyright protection prefill |
2024-02-16 | 99026060 | AI assisted coding app cursor.sh shared, anons warned about request rerouting |
2024-02-17 | 99026787 | canadians admit, chatbots aren't separate legal entities responsible for their own actions |
2024-02-17 | 99028615 | anon wishes to use Gemini for TTRPGs via Foundry VTT plugin |
2024-02-17 | 99028891 | anon talks to Google support about RnD, generally availability of Gemini Ultra doubted |
2024-02-17 | 99029662 | openrouter.ai offers Claude 2 and Claude Instant with "self-moderation" |
2024-02-17 | 99029965 | Gemini Advanced might flashbang users for hitting the filter |
2024-02-17 | 99034792 | regexes for various API keys to facilitate scraping shared |
2024-02-17 | 99036068 | arabian media outlet uses Claude for its texts, exposed due to Anthropic's prefill |
2024-02-17 | 99036411 | HF reported to still delete even personal proxies |
2024-02-17 | 99039276 | /aicg/ advertises on 4chan, the original and largest English-language image board |
2024-02-17 | 99040061 | anon reminds, that 52B was the size of Anthropic's original prototype model RL-CAI/Madison/Claude |
2024-02-17 | 99042009 | universities run ads to resell students' data for training llms |
2024-02-18 | 99045527 | anon reminds, that GPT4V has additional prompt inserted and supports less features |
2024-02-18 | 99045675 | anon baits with uncensored image recognition outputted as JSON |
2024-02-18 | 99045941 | /aicg/ is advertised on 4chan, the original and largest English-language image board |
2024-02-18 | 99048065 | Anthropic moves away from text completion, prefill still works in chat completion mode... for now |
2024-02-18 | 99049025 | Gemini Ultra available for some API keys |
2024-02-18 | 99051884 | anon shares lorebooks formatting advice for ChatML |
2024-02-18 | 99054121 | Anthropic valued at more than $18 billion premoney |
2024-02-18 | 99055726 | anon advises... eh, here's a generated tagline "Spice up AI conversations: Experiment with token spacing and more!" |
2024-02-18 | 99058400 | llm performance decreases with long inputs, but Google has ideas |
2024-02-18 | 99058593 | anon has crazy interesting theory about Claude's context size |
2024-02-19 | 99060356 | irrelevant GPT-J likes dicks |
2024-02-19 | 99071052 | llms become dumber when asked complex questions with scattered supporting facts |
2024-02-19 | 99074109 | Aesir dataset containing Mancer logs leaked |
2024-02-20 | 99074752 | no Gemini 1.5 Pro API yet |
2024-02-20 | 99074836 | Gemini 1.5 Pro success might mean Gemini 1.0 Ultra failure |
2024-02-20 | 99074886 | anon continues to work on his CYOA-style sci-fi thingy using STscript and JS, prevails |
2024-02-20 | 99076000 | GPT4 Turbo fails needle in a haystack tests with more complex methodology |
2024-02-20 | 99076594 | figgs.ai astroturfed |
2024-02-20 | 99076716 | another DALL-E 3 frontend promised... |
2024-02-20 | 99078605 | ... but not delivered |
2024-02-20 | 99081989 | you can make $200 selling slopbots, but you don't |
2024-02-20 | 99082125 | openrouter.ai keys are easy to scrape |
2024-02-20 | 99083319 | Gemini 1.5 Pro successfully generates a character based on a 250k token rulebook |
2024-02-20 | 99083583 | Groq reverse proxy suggested, not real |
2024-02-20 | 99084784 | exposed Silly instances invaded again |
2024-02-20 | 99085279 | Llama 3 available for 3 weeks already subject to NDA, watermarked |
2024-02-20 | 99086291 | anon reports self-moderated Claude on openrouter.ai to be filtered, methodology questionable |
2024-02-20 | 99087009 | ShellGPT command-line tool shilled |
2024-02-21 | 99097984 | yet another alternative frontend claim that won't be realized |
2024-02-21 | 99099876 | ChatGPT acts schizo again |
2024-02-21 | 99103130 | now comments are required to rate bots on CharacterHub |
2024-02-21 | 99104711 | Google releases 2B and "7B" Gemma models |
2024-02-21 | Lightning | SDXL Lightning distilled model released, LoRAs can be applied to other base models |
2024-02-21 | 99110257 | Google acknowledges Gemini imagegen retardation |
2024-02-21 | 240216 | NTIA gathers feedback on the risks and implications related to open foundation models |
2024-02-22 | 99115424 | Claude is officially called a silly little guy, and other weird domains connected to Anthropic |
2024-02-22 | redesign | c.ai is out of beta |
2024-02-22 | 99124001 | Gemini 1.5 Pro multimodality demoed |
2024-02-22 | SD3 | Stable Diffusion 3 early preview announced |
2024-02-22 | 6:42 AM | Google pauses Gemini image generator |
2024-02-23 | unsettling | HF introduces a red-teaming resistance leaderboard |
2024-02-23 | 15570 | new method of automated jailbreaking explained, tested on 7B, could be extrapolated to GPT4 |
2024-02-24 | 99151320 | anon gets Gemini 1.5 Pro access... |
2024-02-24 | 99151545 | accessible only via Google AI Studio, no API provided |
2024-02-24 | 99152477 | anon finds a Sora key, supposedly useless |
2024-02-24 | 99156667 | OAI does recommend to use XML tags |
2024-02-25 | 99163434 | Nvidia says programmers aren't needed anymore |
2024-02-25 | 99165929 | Gemini 1.5 Pro detail recall reported to be impressive |
2024-02-25 | 99168671 | redditor does the math and believes, info obvious from the Gemini 1.5 Pro model card |
2024-02-25 | 99169067 | HF TTS Arena is shilled |
2024-02-25 | 99171986 | anon claims Anthropic is now pozzing every single API key |
2024-02-25 | b4fdf3d | Agnai gets TabbyAPI and Mistral API support |
2024-02-26 | 99188829 | another benchmark shows how reasoning performance drops with context growth |
2024-02-26 | 99188930 | Google isn't ready to lose any employees to the competitors in AI sector |
2024-02-26 | 99191014 | Microsoft makes a deal with Mistral AI, hosting the models on Azure... |
2024-02-26 | 99191026 | ... while Mistral releases Mistral Large, Mistral Small, Le Chat, and removes mentions of its dedication to open source |
2024-02-26 | 99189627 | new lightweight TTS model MeloTTS released by myshell.ai |
2024-02-26 | 99188288 | new synthetic anime-themed text dataset released |
2024-02-26 | 99191657 | anon uses Mixtral on anyscale.com |
2024-02-26 | 99192634 | c.ai redesign UI showed |
2024-02-26 | NTIA AI | National Telecommunications and Information Administration considers open foundation models dangerous |
2024-02-27 | 99205532 | anon finally shares the prompt for uncensored image recognition outputted as JSON |
2024-02-27 | 99209801 | OAI claims to be hacked by NYT "by using deceptive prompts that blatantly violate OAI’s terms of use" |
2024-02-27 | 99209860 | freellamas anon is back briefly, updates the Colab guide |
2024-02-27 | 99211917 | email spam continues |
2024-02-28 | 99214595 | akuma.ai introduces a tool to turn sketches into manga in real time |
2024-02-28 | 99222796 | the era of 1-bit llms: all llms are in 1.58 bits, i.e. something extremely promising for localchads |
2024-02-28 | 99224221 | Emote Portrait Alive - generating expressive portrait videos with audio2video |
2024-02-28 | 12:35 PM UTC | Mistral CEO says they're still good |
2024-02-28 | 99229164 | a running instance of gpt4free shared |
2024-02-28 | 99231944 | anon scraps 22800 bots metadata from CharacterHub... |
2024-02-29 | 99233346 | ... and shares the stats with /lmg/ |
2024-02-29 | 99232752 | BitNet b1.58 expected to deliver so much more for so much less |
2024-02-29 | 99234485 | Meta supposedly calls Llama 2 too cucked, supposedly intends to release Llama 3 in July |
March 2024
When | Where | What |
---|---|---|
2024-03-01 | 99259798 | Musk sues Altman and OAI for being greedy fags |
2024-03-01 | 99261312 | anons are reminded that github reports Anthropic keys |
2024-03-01 | 99265042 | new proxy exploit claimed to exists, nothing follows |
2024-03-02 | 99275082 | differences between GPT4 snapshots are severe |
2024-03-02 | 99282604 | scary Q-451-921 - QUALIA OAI leak turns out to be comfy desu fake |
2024-03-03 | 99290444 | a proxy reaches $100k |
2024-03-03 | 99293812 | anon shares world's best ideas for a first line of a greeting, Lyttle Lytton Contest 2023 |
2024-03-04 | 99311453 | Claude 3 mogs GPT4, has vision, API generally available with the free trial, retains the sovl |
2024-03-04 | 99313619 | Claude 3 available in us-west-2 AWS region |
2024-03-04 | 99314462 | Claude 3 Vision doesn't have an external NSFW filter, GPT4V's was removed |
2024-03-04 | 99315811 | only Claude 3 Sonnet is available on AWS at the moment |
2024-03-04 | 99316304 | somehow Claude 3 turns out to be sexist, a surprise, to be sure |
2024-03-04 | 99316520 | The proxy gets Claude 3 support |
2024-03-04 | 99316983 | Claude 3 is available on poe.com |
2024-03-04 | 99317895 | Claude 3 Opus jokes around while being tested |
2024-03-04 | 99319968 | previously GPT4 Turbo-only cards perform well on Claude 3 Opus |
2024-03-05 | 99320798 | anon uses Selenium to mass sign up for Claude API to abuse the free trial |
2024-03-05 | 99324338 | anon makes use of cute emojis in a card |
2024-03-05 | 99326444 | Sergey Brin admits Gemini imagegen launch was a flop |
2024-03-05 | SD3 | Stable Diffusion 3 research paper published, beats DALL-E 3 and Midjourney v6 in typography and prompt adherence |
2024-03-05 | 99329039 | Claude 3 on openrouter.ai reported to behave differently from direct API access... |
2024-03-05 | 99329639 | ... with additional confirmation |
2024-03-05 | 99329872 | anon erps with Gemini via Bard |
2024-03-05 | 99330626 | even after all this time anons believe the llms' hallucinations |
2024-03-05 | 99331315 | Claude 3 understands quantum physics at the PhD level |
2024-03-05 | 99332069 | Anthropic has a botmaking bot |
2024-03-05 | 99336335 | Cloudflare announces Firewall for AI, i.e. a filter reading both prompts and responses |
2024-03-06 | 99341089 | BREAKING: a twitter person is unhappy because Claude hallucinates; FBI, CIA, etc already notified |
2024-03-06 | 99341672 | Google might be giving out Gemini 1.5 Pro access more freely |
2024-03-06 | 99344099 | a source of free GPT4 Turbo and Claude 3 exposed |
2024-03-06 | 99345214 | OAI attempts to dramafag and get back at Musk by leaking emails, instead confirms his accusations... |
2024-03-06 | 99345442 | ... as additionally explained by an anon here |
2024-03-06 | 99347901 | Claude tries to be unbiased, anon is one of those gura/gultra fags |
2024-03-06 | 99348254 | Claude 3 vs GPT4 benchmark results comparison might have been unfair |
2024-03-06 | 99348797 | anon continues comparative testing, is revealed to be braindead, asks to disregard this |
2024-03-06 | 99349453 | twitter considers Claude 3 Opus to be better than GPT4 Turbo at coding |
2024-03-06 | unredact | Claude 3 used to hallucinate the redacted parts of OAI's emails |
2024-03-06 | Unsloth | Gemma has bugs |
2024-03-06 | 99353417 | Silly holds a contest for the best original general purpose AI assistant |
2024-03-06 | FSDP | a 70B model can be trained on two 24GB GPUs |
2024-03-07 | trained | a 7B model can be trained on a single 24GB GPU |
2024-03-07 | 99358229 | anon predicts "full release" of Claude 3 on AWS on the 14th |
2024-03-07 | 99359970 | ChatGPT memory is remembered, temporary chats won't be remembered, usage policies change noticed |
2024-03-07 | 99366485 | Claude 3 is claimed to have gotten worse within a few days |
2024-03-07 | 468527441 | following multiple failed attempts, a permanent /aicg/ outpost on /vg/ established |
2024-03-08 | 99372051 | anon shares CSS for Silly on mobile |
2024-03-08 | 99386510 | virtual number services have a dedicated Claude option |
2024-03-08 | 99388740 | Claude 3 Opus is the new drug that allows anons to feel happiness again |
2024-03-09 | 99390814 | Anthropic employs automatic review process to combat free trial abuse... |
2024-03-09 | 99395941 | ... and fakes the email refusal error |
2024-03-09 | 99398837 | Claude 3 Opus excels in weird/complex formatting/stat tracking |
2024-03-09 | 468705829 | sj.u2b.cx youtube raw video file proxy shared |
2024-03-10 | 99409995 | plans for a CharacterHub app to be released are brought to attention |
2024-03-10 | 99410537 | c.ai being a vore haven remains a known fact |
2024-03-10 | 99411672 | yet another new frontend bait |
2024-03-10 | 99412875 | DALL-E 3 frontend shilled |
2024-03-10 | 99418020 | Claude 3 Sonnet lacks abilities, Opus is too good at imitating |
2024-03-10 | 468920870 | Claude 3 Opus recall ability praised |
2024-03-10 | 99421325 | h2oGPT demo exposed |
2024-03-11 | 99422373 | c.ai anti-filter petition has 150k signatures |
2024-03-11 | 99424042 | some open source models are better than some old proprietary models |
2024-03-11 | 99428007 | anon adds google search access to Silly |
2024-03-11 | 99428691 | CharacterHub app released |
2024-03-11 | musk | xAI promises to open source Grok this week |
2024-03-11 | 99433646 | IP counter is gone |
2024-03-12 | 99441695 | anon explains how to keep down the costs of slowburns with Claude 3 |
2024-03-12 | 99445788 | UCI faculty & staff get free GPT4.5 ZotGPT |
2024-03-12 | 99448605 | AWS Claude 3 is claimed to be quantized hence dumber |
2024-03-12 | Truffle-1 | Truffle-1, a $1299 inference engine designed to run OSS models using just 60 watts, released |
2024-03-13 | 99460210 | CharacterHub disallows filtering by unsavory tags without an account |
2024-03-13 | 99460383 | "The model gpt-4 does not exist" scare |
2024-03-13 | 99460743 | Devin the programming llm agent presented |
2024-03-13 | 99461665 | yodayo.com has to implement a filter |
2024-03-13 | 99464336 | anon unlocks unlimited Opus due to "amount: Input should be a valid integer" error |
2024-03-13 | 99467051 | Google technical writing courses recommended to learn how to write better |
2024-03-13 | 99467507 | c.ai is bad but offers new TTS voices |
2024-03-13 | 99466699 | Distill Web Monitor recommended to monitor web for changes |
2024-03-13 | 99470549 | securing the "AI" is such a serious business |
2024-03-13 | 469247748 | local novelcrafter.com frontend alternative shared |
2024-03-14 | 99478371 | Claude 3 Haiku released |
2024-03-14 | 99478576 | figure.ai's robot can talk and use hands |
2024-03-14 | 99478847 | Haiku is claimed to be the most harmless option by Anthropic |
2024-03-14 | 99479038 | anon's prediction comes true, Claude 3 Haiku available on AWS |
2024-03-14 | 99485348 | anon promises more AWS insights |
2024-03-14 | 99485161 | Anthropic API free trial issues brought to attention |
2024-03-14 | 385 | HuggingChat allows Internet access |
2024-03-15 | 99493631 | streaming bug of The proxy exposed |
2024-03-15 | 99498201 | Anthropic fucks around with Claude 3 API prefill... |
2024-03-15 | 99498506 | ... additional examples provided... |
2024-03-15 | 99498666 | Claude API output filtering system observed |
2024-03-15 | 99499154 | ... with different behavior across Claude 3 models |
2024-03-15 | 99499274 | anon uses Haiku for a home voice chatbot, explains the way |
2024-03-15 | 99502932 | another Claude Hackathon is held, no API access granted |
2024-03-15 | 99507013 | Clickable Inputs Silly extension shilled |
2024-03-15 | 99508369 | Silly allows greeting-specific defs/character-notes/scenarios |
2024-03-15 | 99508436 | hackers can read private AI-assistant chats, "All non-Google chat GPTs" phrasing proves journos are braindead |
2024-03-16 | 99518052 | January 2024 ranking the top 50 AI-first web products, according to SimilarWeb |
2024-03-16 | 99519124 | current Claude prompt injections said to be conditional |
2024-03-16 | 99514558 | data about proprietary models can be obtained by prompting |
2024-03-17 | 99527470 | yet another site bans NSFW, citing restrictions from payment processors |
2024-03-17 | 99529248 | Anthropic's actual customers are getting banned on sight |
2024-03-17 | 99531349 | Anthropic accounts with tempmail said to be purged, can't be created |
2024-03-17 | 99535727 | the difference in quality between AWS and API Claude explained by a bug of The proxy |
2024-03-17 | 99538284 | image recognition capability of Opus compared to Gemini and GPT4 Turbo |
2024-03-17 | 99538903 | botmaking idea bord idea suggested |
2024-03-17 | 99539957 | Grok is open sourced |
2024-03-17 | 99540176 | Claude API keys are now tiered with varying rate limits |
2024-03-18 | 99542893 | Claude API rate limits key checker update |
2024-03-18 | 99543813 | Apple teases MM1, a family of superior multimodal models |
2024-03-18 | 99551964 | request anchor idea acquiesced to |
2024-03-19 | 99558410 | Lobe Chat frontend shilled |
2024-03-19 | 99563189 | a way to reuse gmail for Anthropic free trial abuse suggested |
2024-03-19 | 99565842 | Google informs about the dangers of exposing the Google Cloud keys, intends to disable them when detected |
2024-03-19 | 99566012 | prefill support confirmed for openrouter.ai |
2024-03-19 | 99569271 | anti-proprietary model protestors outside of the NVIDIA GTC keynote venue |
2024-03-20 | 99579071 | Anthropic fixes "Input should be a valid integer" bug |
2024-03-20 | 99580922 | NVIDIA Blackwell Platform is fast |
2024-03-20 | 99586235 | c.ai releases a new TTS and voice cloning system for everyone for free |
2024-03-20 | 99588799 | Microsoft doesn't care if OAI disappears tomorrow, sadly it doesn't |
2024-03-20 | 99589365 | anon takes surveys about how tasty different colored pencils seem to pay for Claude |
2024-03-20 | j56Koka | Novel Builder for Mikugg released along with a new full scale VN |
2024-03-21 | 99591056 | current state of chub-archive reported, its total size reaches 455GB |
2024-03-21 | 99592536 | ascended anon enjoys Mikugg |
2024-03-21 | 99596309 | Anthropic hides free trial credits mentions from the Help Center |
2024-03-21 | 99596341 | autobotmaking tool CharGen brought to attention |
2024-03-21 | 99598770 | pokemon adventure card with elaborate visuals teased |
2024-03-21 | 99599522 | anon compares llms' chess performance |
2024-03-21 | 99601262 | CharacterHub's "following" feature unavailable with unverified email |
2024-03-21 | 99603478 | AWS, Accenture and Anthropic partner to accelerate enterprise AI adoption |
2024-03-21 | 99603818 | AetherRoom devlog 3 |
2024-03-22 | 99606630 | limited CharacterHub passwords leak, turns out not to be its fault |
2024-03-22 | 99607843 | Gemini 1.5 Pro API is being rolled out, waitlist removed from the playground |
2024-03-22 | 99610583 | anon explains transformers |
2024-03-22 | 99611307 | private venus.chub.ai logs from the accounts with leaked passwords published |
2024-03-22 | 99613192 | beta.luka.ai by the creators of Replika shilled |
2024-03-22 | 99614682 | you.com subscription with unlimited chat access to the best models brought to attention |
2024-03-23 | clem | HF considers acquiring Stability AI |
2024-03-23 | 99633660 | anon sharest the way to revoke Anthropic keys... |
2024-03-23 | 99635497 | ... The proxy adds an option to prevent this |
2024-03-23 | 99634481 | Mistral 7B Instruct v0.2 with real 32k context released |
2024-03-24 | 99644363 | Deep Infra suggested as a way to get free Mixtral and SD |
2024-03-24 | 99639481 | app.suno.ai has strict moderation |
2024-03-24 | 99639783 | Anthropic changes the way free trial is provided per phone number |
2024-03-24 | 99648338 | GPT5 rumors place the release "mid-year" |
2024-03-25 | 99659203 | Devika, a competitive open source alternative to Devin, brought to attention |
2024-03-25 | 99660711 | another effortcard with 184 images shared |
2024-03-26 | 99669001 | more leaked credentials for chatbots services posted |
2024-03-26 | 99674830 | Maid app update |
2024-03-27 | 99685103 | new GPT4V snapshot noticed |
2024-03-28 | 99716559 | rumors claim Sora to be GPT5 |
2024-03-28 | qwenlm | Qwen1.5-MoE-A2.7B, a small MoE 7B alternative with 2.7B activated parameters, released |
2024-03-28 | ai21 | Jamba, a hybrid architecture 12B/52B model with 256k context, released |
2024-03-28 | grok-1.5 | Grok-1.5 with 128k context announced |
2024-03-28 | 471248312 | CharacterHub app nuked |
2024-03-29 | 99722849 | some Anthropic verification number providers don't work |
2024-03-29 | 99733900 | some scraped Google API keys have gained access to Gemini 1.0 Ultra |
2024-03-30 | 99740400 | jennyanon returns from the hospital, takes on random strangers raw no rubber, takes beatings, the cat is fine too |
2024-03-30 | 99740430 | anon explains how to host a personal proxy on HF |
2024-03-30 | 99742213 | 1-bit is even better than BitNet and 1.58 bit |
2024-03-30 | 99745108 | Double, an AI coding assistant, lists GPT5 as "coming soon" |
2024-03-30 | 99750889 | a gift exchange event proposed |
April 2024
When | Where | What |
---|---|---|
2024-04-01 | 99777941 | websim.ai shilled, requires a Claude key to function |
2024-04-02 | 99791495 | chub-archive server upgrade postponed due to CPU issues |
2024-04-02 | 99798947 | an example of the bizarre ways people prompt chatGPT shared |
2024-04-02 | 99799562 | booru.plus is shut down |
2024-04-02 | 99799814 | limited passwords leak continuation, Agnai credentials posted |
2024-04-02 | 99799730 | Anthropic explores many-shot jailbreaking, considers pre-filtering, fear mongering as always |
2024-04-02 | 99800265 | claude.ai accounts are reviewed automatically, can be disabled for no reason |
2024-04-02 | in a month pay-as-you-go pricing for Gemini API will be introduced | |
2024-04-03 | 99804465 | character preview script for CharacterHub links shilled, immediately updated to preview the defs |
2024-04-03 | 99805860 | Amazon offers free credits for startups to use AI models including Anthropic |
2024-04-03 | 472039801 | a worldbook building script teased |
2024-04-03 | 99808354 | CharacterHub users can freely add tags to other people's characters |
2024-04-03 | 99812296 | Mistral Large available on AWS |
2024-04-03 | 99812332 | character preview script gets NSFW blur feature |
2024-04-03 | 472108189 | fandom.com scraping script released |
2024-04-03 | stability | Stable Audio 2.0, a service to generate long high-quality melodies, introduced; not open source |
2024-04-04 | 99823793 | a way to emulate Claude prefill for Opus based chatbot services shared |
2024-04-04 | 99827025 | more rumors about the impending GPT5 release |
2024-04-04 | 99835186 | a proxy with a proof of work captcha demoed |
2024-04-04 | cohere | Command R+, 104B model with 128k context, released; the best open source model to date |
2024-04-05 | 99842376 | function calling available to all customers in the Anthropic Messages API |
2024-04-05 | 99845329 | SillyTavern Character Generator shilled, requires an API key, updated to accept a proxy link |
2024-04-05 | 99853767 | forks are spammed on CharacterHub |
2024-04-05 | 99854354 | new Anthropic trial keys are 2 week limited, old ones aren't |
2024-04-06 | 99856619 | a script to exploit world_sim as a source of free Opus shared |
2024-04-06 | 99857305 | AWS warns about suspected credentials exposure in connection to unusually high usage |
2024-04-06 | 99861335 | an alternative exploit script shared |
2024-04-06 | 99863069 | the exploit usage can be monitored on openrouter.ai... |
2024-04-06 | 99867335 | ... with the activity graph allowing to see the surge, suggesting it was quietly exploited since the day it was astroturfed |
2024-04-06 | 472489015 | character preview script updated |
2024-04-06 | 99867903 | world_sim Opus depleted |
2024-04-06 | 99874471 | Characterhub Extension Template linked in response to LLM-powered games being brought up |
2024-04-07 | 99879874 | world_sim requires to sign in, said to have switched to Sonnet, still considered exploitable |
2024-04-08 | 99893970 | Characterhub resets profile pics to enforce the new venus.chub.ai logo |
2024-04-08 | stability | Stable LM 2 12B released, cucked |
2024-04-08 | 3:33pm | Meta Platforms to launch small versions of Llama 3 next week |
2024-04-09 | 99908617 | anon shares info about useful scripts for Bing Create |
2024-04-09 | 99909430 | anon learns the obvious source of potential new services to exploit, points at sagarpg.io |
2024-04-09 | 99923000 | new GPT4 Turbo snapshot, has Vision, easier to jailbreak |
2024-04-10 | 99925906 | Gemini 1.5 Pro API available for everyone with full 1kk context; video inputs now support STT in AI Studio |
2024-04-10 | 99926032 | new benchmark paper puts Opus above GPT4, Ultra sucks |
2024-04-10 | 99928559 | Mixtral 8x22B with 65k context released, new best local |
2024-04-10 | 99932303 | claude.ai sometimes downgrades users from Sonnet to Haiku |
2024-04-10 | 99933993 | Gemini voice recognition abilities praised, 6 speakers recognized |
2024-04-11 | 99942911 | /mag/, yet another attempt to create a Music AI General, lives longer than a single thread |
2024-04-11 | 99945873 | poly.ai email spam |
2024-04-12 | 99958063 | websim.ai rediscovered, now supplies its own Claude, enjoyed by all |
2024-04-12 | ylecun | OAI and Meta want bots that reason |
2024-04-12 | 99963089 | Opus available on Vertex AI |
2024-04-12 | 99965631 | udio.com shilled, an alternative to suno.ai |
2024-04-12 | 99965893 | Characterhub allows to blacklist words, preventing bad reviews |
2024-04-12 | 99970726 | gab.ai system prompt shared |
2024-04-13 | 99973084 | websim.ai gets too popular, exceeds its rate limit |
2024-04-13 | 99973086 | anon enjoys Gemini via Bard with devastating results |
2024-04-13 | 99974269 | Gemini recall abilities praised, 400k token rp continues smoothly |
2024-04-13 | 99980209 | anon promises to share a hidden RunPod feature |
2024-04-14 | 99994906 | premai.io brought up as a questionable source of free Claude |
2024-04-14 | 99997353 | anon encounters an OpenRouter bug, enjoys free Claude |
2024-04-14 | 99999666 | a reminder that Command R+ is available for testing officially for free |
2024-04-14 | 99994247 | incomplete Anthropic timeline |
2024-04-15 | 100011451 | OAI releases a custom GPT4 model optimized for the Japanese language |
2024-04-15 | 100015920 | Reka Core, multimodal, fake 128k context, size unknown, free playground with 4k token chat cap |
2024-04-15 | wizardlm | WizardLM-2 finetunes by Microsoft AI released, new best local, uncensored |
2024-04-15 | quora | multi-bot chat on Poe lets you chat with multiple models in a single thread |
2024-04-16 | 100019544 | Reka to Tavern adapter |
2024-04-16 | 100021767 | a reminder that there's a lot more sites with exploitable Claude to be found using Censys |
2024-04-16 | 100022632 | cftf.ai shilled, a tool to convert character metadata to pixels and back |
2024-04-16 | 100024114 | WizardLM-2 models taken down due to toxicity |
2024-04-16 | 100024357 | OAI introduces Batch API, cheaper async tasks |
2024-04-16 | 100025584 | chub-archive offline due to data migration |
2024-04-16 | 100025898 | Opus does what GPT4n't |
2024-04-16 | 100027282 | paid option with daily limits shared, considered shady |
2024-04-16 | 100027809 | Opus soon on AWS |
2024-04-16 | 100028693 | Opus available on AWS |
2024-04-16 | 100030140 | cftf.ai attacked by decompression bombs, source code shared |
2024-04-16 | 100031515 | OAI introduces a way to manage individual projects with the ability to set billing and model restrictions |
2024-04-17 | 100035496 | moemate.io limits the context window and the amount of prompts a day before the arbitrary downgrade to Mixtral |
2024-04-17 | 100039281 | Character Card V3 concept published |
2024-04-17 | 100039916 | Perplexity to Tavern adapter |
2024-04-17 | 100042890 | Claude Vision appears to use Cloud Vision by Google |
2024-04-17 | 100044945 | poly.ai email spam continues |
2024-04-17 | 8x22b | finally the official blog post dedicated to Mixtral 8x22B |
2024-04-18 | 100057879 | SD3 API available, performs unsatisfactory, Turbo mentioned |
2024-04-18 | 100062736 | OAI is losing their safety-focused talents |
2024-04-18 | 100063009 | llama3 inference pricing listed on replicate.com and Azure |
2024-04-18 | 100065742 | Claude early access evaluation accounts will lose free access in a month... |
2024-04-18 | 100068226 | ... with an example of the email posted |
2024-04-18 | llama-3 | Llama 3 released, 8B and 70B, 8k context; 400B in training; multimodality, longer context, etc. promised later |
2024-04-19 | 100071711 | MikuForbinDev reports $11k in losses, still plans to open source the app |
2024-04-19 | 474207194 | Characterhub adds “Tag Moderators” to allow select people to force-tag others' projects |
2024-04-20 | 100089781 | Characterhub Apple App Store app released |
2024-04-20 | 100093252 | redditor claims chatbot arena from LMSYS is manipulating votes |
2024-04-20 | 100100042 | something from OAI is expected in two days |
2024-04-21 | 100106223 | Characterhub Extensions security concerns addressed |
2024-04-21 | 100109296 | anon promises Phi-3 release soon |
2024-04-21 | 100115955 | someone talks about some cryptic game |
2024-04-22 | 474582103 | anon learns more about OAI completion objects |
2024-04-22 | 100122834 | Characterhub users can reply to reviews now |
2024-04-22 | 100127969 | an example of AWS investigating Anthropic policy violations |
2024-04-22 | 100128448 | replying to reviews on Characterhub affects the rating |
2024-04-22 | 100129943 | getmerlin.in suggested as a possible way to pay for AI access, dismissed |
2024-04-22 | 100131748 | anon makes a browser plugin to summarize the contents of a 4chan thread |
2024-04-22 | 100133292 | Phi-3 release confirmed |
2024-04-22 | 100134263 | chatbot apps astroturfing |
2024-04-22 | 100134905 | prominent researchers from various large organizations occasionally visit /lmg/ |
2024-04-23 | 100135451 | NovelAI Diffusion Furry V3, whitepapers for NAIv3 not released, artist tags removed |
2024-04-23 | 100138393 | rare understanding of current Gemini availability exhibited |
2024-04-23 | 100141747 | Amazon employee makes llms battle Pokemon, Opus wins |
2024-04-23 | 100145947 | Phi-3 released, 128k context, only 3.8B instruct available, lobotomized savant |
2024-04-23 | 100148412 | llms have big context windows but don't work well enough for enterprise clients |
2024-04-23 | 100146232 | multimodal llama3 hinted at |
2024-04-24 | 100151858 | anon is building API service that'll host local models to be used over tor and paid for with monero |
2024-04-24 | 100153358 | last booru.plus scrape was on 2024-03-04 |
2024-04-24 | 100154306 | Gemini 1.5 Pro reported to be really fucking good at translation and rape |
2024-04-24 | 100159590 | AWS allows to import and server custom local models, Guardrails generally available |
2024-04-24 | 100163277 | mentee robotics presents llm-based MenteeBot |
2024-04-24 | 100163556 | Guardrails features showcased |
2024-04-25 | 100170157 | chub-archive anon reports archive overhaul, original files removed |
2024-04-25 | 100171309 | venus.chub.ai extensions go live |
2024-04-25 | 100178320 | Characterhub tags moderation observed in action |
- | - | - |