42school student | AI engineer @CMACGM | 26.
Lebanon
Joined April 2022
- Tweets2.3K
- Following592
- Followers444
- Likes2.6K
If money wasn’t a problem, I’d probably spend my whole life studying.
A master’s in philosophy, then astrophysics, ancient history, psychology, archaeology… whatever makes me understand this weird existence a little better.
Not for a career. Not for a degree.
Just because there’s so much to know, and one lifetime already feels too short.
AI rewriting its own architecture is crazy.
Over two weeks, Prime Agent orchestrated a swarm of over 2,000 agents to rewrite itself in Rust.
With 10,000+ sandboxes, 200B+ GLM-5.3 tokens, and 16,000 agent-to-agent messages, this was our largest test yet of Prime Agent’s multi-agent capabilities and our infrastructure powering these agent swarms.
Prime Agent now reaches usable input ~13× faster and uses 83% less startup memory than before.
Try telling someone in 2016 that AI would be writing most of the world's code and handling much of the work we do on computers.
Now imagine trying to explain 2036 to someone today.
We truly live in wild times.
What happened to this perspective?
If this Karpathy interview doesn't pop the ai bubble,
nothing will.
10 brutal quotes:
1. LLMs don’t work yet
They don’t have enough intelligence, they’re not multimodal enough, they can’t use computers, and they don’t remember what you tell them.
They’re cognitively lacking. It’ll take about a decade to work through all of that.
2. When you boot them up, they always start from zero
They have no distillation phase, no process like sleep where what happened gets analyzed and written back into the weights.
3. What’s stored in their weights is only a hazy recollection of the internet
It's just a compressed blur of 15 trillion tokens squeezed into a few billion parameters. Their context window is just short-term working memory.
4. They’re good at imitation, terrible at going off the data manifold
Too much memory, not enough reasoning.
We need to strip away the memorized knowledge and keep the cognitive core: the algorithms, the magic of intelligence, problem-solving, strategy.
5. We’ve probably recreated a cortical tissue, pattern-learning and general, but we’re still missing the rest of the brain
No hippocampus for memory.
No amygdala for instincts.
No emotions or motivations.
6. They memorize perfectly but generalize poorly
If you give them random numbers, they can recite them back. No human can do that.
That’s the problem: humans forget just enough to be forced to find patterns.
7. Anything truly new, code that’s never been written before, ideas that have no template; they stumble
They’re still autocomplete engines with perfect recall and no understanding. Until we find that cognitive core, intelligence stripped of memory but full of reasoning, they’ll stay brilliant mimics, not minds.
The best skill to have in 2026 is to be a visionary, to understand accurately how the world will work in 10 years, what technologies will dominate, what people gonna use the most.
So i've been building a shell from scratch in C at 42school for the past few weeks and i've been wondering why does the terminal even exist in the age of AI, why aren't there just a simple AI chatbot interface linux distros out there, would you use that kind of OS?
Is there a need for an AI Operating system?
No desktop.
No settings.
No shell.
Just a chatbot interface.
People that will enter AI in 2025 will be rich,
just like people that entered the internet in 1995,
just like people that entered personal computing in the 1970s,
just like people that entered the automobile/assembly line industry in the 1910s,
just like people that entered the electricity industry in the 1880s,
just like people that entered the railroad/steam engine industry in the 1830s,
just like people that entered the industrial revolution/textile manufacturing in the late 1700s,
just like people that entered the agricultural revolution/farming in the Neolithic era.
Jason retweeted
One of my favorite lessons I’ve learnt from working with smart people:
Action produces information. If you’re unsure of what to do, just do anything, even if it’s the wrong thing. This will give you information about what you should actually be doing.
Sounds simple on the surface - the hard part is making it part of your every day working process.
