Log inSign up
user avatar

Saurabh Srivastava

@_saurabh
Research Scientist @ Nemotron, NVIDIA; **Views are my own** Previously: Head of Research, Code @ Essential AI; 2x YC (W15, S18); PhD + Postdoc in Code Synthesis
San Francisco, CA
Joined November 2008
1,502
Following
1,064
Followers
RepliesMedia
  • user avatar
    Saurabh Srivastava
    @_saurabh
    Jul 24
    @JensenHuang is on X; and his first post: open weights matter. Open weights = collaboration + competition + and surprising we have to say this + safety [known: security through obfuscation is a dead-end]. We have to empower every mathematician, scientist, and defender.
    user avatar
    Jensen Huang
    NVIDIA
    @JensenHuang
    Jul 24
    For my first post, I’m sharing a letter @nvidia signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
    Image
    Image
    Image
  • user avatar
    Saurabh Srivastava
    @_saurabh
    Mar 17
    Community together solved 8/10 of first proof. spec + reviews are the new bottlenecks in both code and math. expect to see solutions to the latter this year. Litt: “I actually expect to be doing the best work I’ve ever done, because I’ll have these amazing tools.” “Current
  • user avatar
    Saurabh Srivastava
    @_saurabh
    Mar 3
    Don Knuth co-solving an open problem with human-AI collaboration. Calling it "Claude cycles" feel's like the right attribution. We should note: Noticing that a narrower version of the problem can be solved is an important cognitive ability! A model identifying the right
  • user avatar
    Saurabh Srivastava
    @_saurabh
    Feb 25
    I agree that some form of this is inevitable by 2027. I conjecture that we might also see progress in what I call automated *hypothesis generation*. Business needs are driving inference (speed/cost) close to zero. That will unlock directions in hypothesis generation.
    user avatar
    Taelin
    @VictorTaelin
    Feb 24
    Replying to @SebastienBubeck
    proofs will be worthless and theorem proving will be fully automated. only definitions will still be human-driven, at least in the nearish term, since AI still harshly lacks out-of-the-box thinking. yet, whenever a human comes up with a new cool definition, the AI will be able
  • user avatar
    Saurabh Srivastava
    @_saurabh
    Feb 24
    @littmath's essay is a good read. Daniel is revising his ETA for autonomous AI mathematicians. We got here by mapping math into (lean) code. Contrary to opinions that code is solved, we are just getting started. Using code as the domain language has outsized benefits. Worth