antoine ross

on a journey to build great things with a box of scraps

founder of

[antoine ross @antoineross__]

Pinning this right before I become successful from the Philippines

[Neetbux @NeetbuxSOL]

@amandyks > born in nyc

yeah super hard to become successful being born in the richest and most opportunity rich city on earth lmao

[antoine ross @antoineross__]

@thsottiaux ssh connections parity on history

if I ssh to macmini, add a conversation there

I should also be able to see that conversation on the macmini's codex app. today I cannot do that

it's very annoying. to get devs out of tmux hell you guys have to fix this

[antoine ross @antoineross__]

@thsottiaux when you connect to an ssh it does not persist the session remotely, but rather locally. So you can't find it from the codex app on the actual remote

speaking for a macmini user

[antoine ross @antoineross__]

i love codex sm, genuinely

moved from copilot to this and never looking back

[Tibo @thsottiaux]

The Codex usage limits have been reset for all paid ChatGPT subscriptions. You should be back to 100% weekly and 100% hourly limits.

Let the tokens do incredible things today and have fun.

[antoine ross @antoineross__]

@thsottiaux i love you

[Andrej Karpathy @karpathy]

Fireside chat at Sequoia Ascent 2026 from a ~week ago. Some highlights:

The first theme I tried to push on is that LLMs are about a lot more than just speeding up what existed before (e.g. coding). Three examples of new horizons:

1. menugen: an app that can be fully engulfed by LLMs, with no classical code needed: input an image, output an image and an LLM can natively do the thing.
2. install .md skills instead of install .sh scripts. Why create a complex Software 1.0 bash script for e.g. installing a piece of software if you can write the installation out in words and say "just show this to your LLM". The LLM is an advanced interpreter of English and can intelligently target installation to your setup, debug everything inline, etc.
3. LLM knowledge bases as an example of something that was *impossible* with classical code because it's computation over unstructured data (knowledge) from arbitrary sources and in arbitrary formats, including simply text articles etc.

I pushed on these because in every new paradigm change, the obvious things are always in the realm of speeding up or somehow improving what existed, but here we have examples of functionality that either suddenly perhaps shouldn't even exist (1,2), or was fundamentally not possible before (3).

The second (ongoing) theme is trying to explain the pattern of jaggedness in LLMs. How it can be true that a single artifact will simultaneously 1) coherently refactor a 100,000-line code base *and* 2) tell you to walk to the car wash to wash your car. I previously wrote about the source of this as having to do with verifiability of a domain, here I expand on this as having to also do with economics because revenue/TAM dictates what the frontier labs choose to package into training data distributions during RL. You're either in the data distribution (on the rails of the RL circuits) and flying or you're off-roading in the jungle with a machete, in relative terms. Still not 100% satisfied with this, but it's an ongoing struggle to build an accurate model of LLM capabilities if you wish to practically take advantage of their power while avoiding their pitfalls, which brings me to...

Last theme is the agent-native economy. The decomposition of products and services into sensors, actuators and logic (split up across all of 1.0/2.0/3.0 computing paradigms), how we can make information maximally legible to LLMs, some words on the quickly emerging agentic engineering and its skill set, related hiring practices, etc., possibly even hints/dreams of fully neural computing handling the vast majority of computation with some help from (classical) CPU coprocessors.

[antoine ross @antoineross__]

interesting

[Deedy @deedydas]

This is the single best read on World Models and one of the most important reads in AI.

$10B has flowed into "world models" in the last 18mos, from Yann LeCun to FeiFei Li. The promise is, like LLMs, world models will provide the data it takes to scale robotics foundation models, and solve robotics.

..but the word has been abused to mean one of many things.

This post unpacks:
– What 5 traits makes a world model?
– How do the different approaches stack up?
– What is it used for within and beyond robotics?
– Where is the opportunity?
– Citations to research, news and blog posts

Companies / products in the space include:
– BigCo products: Google Genie, Tesla Optimus, Nvidia DreamDojo, DreamZero, Microsoft Muse
– Pure world model: AMI Labs, World Labs, Runway, Rhoda, Decart, Spaitial, Odyssey, Embo, Dream Labs, OneWorld
– Robot foundation model cos: Skild, Physical Intelligence, Figure, Mind

Very likely one of the seminal technologies of the next decade.

[antoine ross @antoineross__]

I'm using rust for my trading engine and the ability to catch and enforce things from runtime is 🤌🏽

[antoine ross @antoineross__]

thank god I didn't give in to the "it just felt right" allegations from ziggers

rust ftw

[antoine ross @antoineross__]

hello testing

[antoine ross @antoineross__]

@paulg or that the AI tends to answer in slop and this is a correction from the users that concise is always better

don't forget: humans don't want to read poetry from AI.

[antoine ross @antoineross__]

cramming an entire hackathon in a day:
127 commits in 24 hours 😄

[antoine ross @antoineross__]

entered a hackathon:

[antoine ross @antoineross__]

devpost: https://t.co/kakLBUDixv

pr: https://t.co/9e2DRNZgXN

[antoine ross @antoineross__]

entered a hackathon:

[antoine ross @antoineross__]

@mitchellh why is it hanging on a thread? far as I know nothing beats Github. still free, still gives you action hours per month. nothing comes close

[antoine ross @antoineross__]

@paulg @SaadMaan15 quality is harder to come by as a byproduct

[antoine ross @antoineross__]

work compliance course on smoking cessation is perhaps the most retarded company policy ever invented

[ирис @saltwatermp3]

Ultimately I have to be brave

[antoine ross @antoineross__]

@AndrewMayne it sounds cool

[antoine ross @antoineross__]

@DavidSHolz @leothecurious abundance of purpose is cheap. as long as there are starving kids in poor countries your abundance of purpose will only exist in a vacuum. AI will not solve this, we know that much

[antoine ross @antoineross__]

@dpaleka the Chinese labs will save us