Porgimus Prime
184 posts
nobody thinks about the consequences in this industry, we are burning rivers to train these models
elie@eliebakouch
new mistral model: 128B dense with an arch from 3 years ago (llama 2), very low context (128k), priced higher than deepseek v4 pro (1.6T total params, 1M context) and every other oss model that outperforms it this is very sad
English
@ajambrosino thats like... after summer
its not even summer yet
it'll be a whole new school year...
English
this (seriously) might be the most advance notice on anything i've ever seen from the company, internally or externally
OpenAI@OpenAI
OpenAI DevDay is back. San Francisco September 29
English
for those of you who aren't aware of the goblins here's an article published by Open AI (goblin city)
English
@BLUECOW009 @zoriya_dev first rule of goblin club- don't talk about goblin club
English
@zoriya_dev openai themselves posted several things about how gpt kept saying goblin
English
have been doing price prediction for weeks now, this is hard work
English
@drawsgood this is EXACTLY what i was looking for literally word for word
English
@mrdavenport this would fit in so well for a project im working on atm😭
English
@BLUECOW009 we also should be weary of the goblin takeover that is happening right in front of our eyes.
English
AI is already very dangerous, the internet might not survive next iterations of models:
>generalized attacks on infrastructure such as github
>AI labs themselves getting pwnd
>massive increase in email spam
>????
English
@BLUECOW009 Then u realise that u need to make a goblin home in Hermes now
English
@BLUECOW009 Bro only wants to post about goblins now. Is that you… GPT 5.5?
English
When u look back at yourself when u started ur project not realising u have 100+ godfiles
GIF
English
Hot Take: If Sonnet 1M context becomes the default Sonnet model in Claude Code (and we do not have to pay for extra usage) Claude Code will become the preferred CLI over Codex again.
English
