Keep on to blur preview images; turn off to show them clearly
RT @MSFTResearch: Fara-7B is our first agentic small language model for computer use. This experimental model includes robust safety measur…
making models learn • eXperiments lab • memes and training lores
China isn't easy to travel Alipay won't accept my Revolut WeChat won't let me sign up without someone else using WeChat to verify me >.<
I'd be fucked without X DM'd 10 people my QR code and they verified me Thank you!
RT @_valsai: Despite billions in investment, only 2 models show real capability: GPT 5.1 and Claude Sonnet 4.5 Thinking. Every other model…
Asst professor @MIT EECS & CSAIL (@nlp_mit). Author of https://t.co/VgyLxl0oa1 and https://t.co/ZZaSzaRaZ7 (@DSPyOSS). Prev: CS PhD @StanfordNLP. Research @Databricks.
More people must know Beren. I admire him a lot. He's been a voice of reason during peak AI Doom era, and after some mathematically rigorous blog posts countering Paperclip-type narrative, he founded Zyphra to contribute to AI transparency and decentralization of power. Great man
A vanishingly rare combination of technical chops, intellectual breadth and depth, integrity and courage to change his opinion when error has existential stakes, and the savvy to execute rather than just drive a feel-good «open science» startup into a ditch. Follow Beren.
Most of Reddit is like this. It needs to be burned to the ground and the fields salted. Then they should replace all mods with AI that follows a charter that all the subreddit veteran accounts can vote to ratify and amend in the future with a super-majority.
Former Quant Investor, now building @lumera (formerly called Pastel Network) | My Open Source Projects: https://t.co/9qbOCDlaqM
pretty Indian too, though I would not mind seeing "DeepSeek of the South Asia", clearly there's enough talent for a top-tier lab
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1