Researcher interested in software engineering, decision science and data management. I also love sushis 🍣 and AI debates. https://xitter.cf/t.co/gMBboyWhGA
Montreal
Joined December 2010
- Tweets1.4K
- Following708
- Followers123
- Likes2.9K
Watching Claude clicking "Im not a robot" and successfully bypassing Captcha 100% of the time is quite funny
How many times you use Claude Code and you realize "it just gets it"? If you still think (gen) AI is just a stochastic parrot, you are the one who are not getting it
looks like @ylecun is doing marketing on things RL community has formalized for decades...
arxiv.org/abs/2603.15381?fbc…
John is a legend
Imagine you're John Carmack
you're 22 years old and you just wrote a 3D engine in assembly that runs at 35fps on a 486
Doom drops. Quake drops. Half the planet is playing your code.
you're the reason GPUs exist. you're the reason your friend Jensen has a yacht today.
then in 2009, you sell id Software. people call it betrayal. you call it "they made an offer I couldn't refuse."
VR obsession. Oculus. Meta buys it for $2B. you're CTO.
but Meta thinks you're a liability. your demos are "too intense." your emails are "too long." your focus on frame timing is "slowing us down."
2022. they push you out. not fired officially. just "restructured."
the media writes "end of an era." some crypto bro calls you "washed up."
silicon valley moves on.
but you don't.
you don't write a book. you don't start a podcast. you don't collect speaking fees.
you go completely quiet.
you take the money. you buy a warehouse in Texas. you hire 10 engineers. and you start coding.
not games. not VR.
AGI.
two years. radio silence. no tweets. no conference talks.
while everyone's debating ChatGPT, you're debugging CUDA kernels at 3AM, testing world models.
then in 2025, Keen Technologies pivots hard. you're not "exploring" anymore. you're building it.
here's what people get wrong:
everyone calls it a comeback. a redemption arc. "revenge on Meta."
it's none of that.
you're a 54-year-old engineer who still codes 12 hours a day because you genuinely can't stop.
most CTOs would have bought an island. most legends would have written memoirs.
you just kept typing.
the most dangerous person in any codebase is the one who goes quiet and never stops shipping commits.
karma doesn't need to be real.
but obsession is.
welcome back, Carmack.
More compute scales models but also allow more people to try different approaches. So we will progress both vertically or horizontally in the next couple of years. We are still at the beginning of this new Industrial Revolution.
AI is the only new kind of software that justify (natural obsolescence) building better PCs for everyone without reducing prices at some point.
When you don't even use your AI to validate your slides... Literally a drag-n-drop in GPT5 @OpenAI
I dont understand why people think LLMs as a Markov chain is new insight. Isn't it what we teach in deep NLP courses?
I wonder when people are going to realize that being able to point out the flaws of AI is not proof they (and any human) have a greater intelligence
The ability of o3 to automatically decide to run code to analyse text documents (so not in the code domain) to verify statements you told him is quite mind-blowing
Any1 saying "LLMs don't reason" because of the recent 🍎 paper is a big red flag for me. If you read the paper, all they say is that if you look at complexity using toy tasks, thinking models don't scale to complexity. And btw, human reasoning don't scale well on many tasks too..
After using them quite extensively, I can say that "reasoning" models are clearly better are using tools than early 2024 models. This is super encouraging.
I am pretty sure than anyone who think/say they are to old to code, never liked coding in the first place
