Cristian Garcia
@cgarciae88
JAX/Flax @ Google DeepMind | Open Source | 🇨🇴
MNIST in Flax NNX Create a CNN for classification with Flax NNX by subclassing nnx.Module: 🧵
learn jax kids
we can only surmise that it costs $100M to poach an openai researcher but $1B to poach a gdm researcher because that’s how much it costs to convince someone to switch from JAX to pytorch
for an alphago moment you have to surpass humans. prove collatz, riemann, P = NP. IMO is more like SWE bench for maths.
its absolutely crazy how investors saw deepseek r1--an open source model that showed how to scale compute at inference time--and decided to sell their nvidia stocks
These agentic teams based models like Grok Heavy, the Gemini Deep Think that just won gold, and the next gen from OpenAI are all going to use about fifteen times more tokens than current systems. This is why Pro plans are north of $200. Essentially; Jensen wins again.
Advanced version of Gemini Deep Think (announced at #GoogleIO) using parallel inference time computation achieved gold-medal performance at IMO, solving 5/6 problems with rigorous proofs as verified by official IMO judges! Congrats to all involved! deepmind.google/discover/blog/…
founders are the process moves money from VCs to token providers
officially joined the $10k claude code club
this has to be called The Scale Game I just need a few investors (@MrBeast??) prepare for season 1
nvidia could do the most viral ai competition in history: start with 10,000 researchers and give each a free gpu to work on a public leaderboard but do rounds of elimination where the winners take the remaining hardware. the final winner gets all the gpus for a year.
be @dhh > reject react > reject windows > reject the cloud > reject build systems > reject types > reject AI > reject venture capital > fight apple and win > promote aesthetic code as a virtue > start racing as a hobby
deepmind and openai getting gold at IMO is cool and all but why did dandadan change the banger intro for season two?
Terrance is right: LLMs are less efficient than humans on a per-token basis but crush humans on throughput (btw GDM also got gold)
Terence Tao on the supposed Gold from OpenAI at IMO
