
42school student | AI engineer @CMACGM | 26.
Lebanon
Joined April 2022
- Tweets2.3K
- Following595
- Followers436
- Likes2.6K
Jason retweeted
Introducing: Self-Improving Memory ๐ง โจ๐
We let the model discover its own memory skills, then train them into its weights:
- Meta-optimizes memory from the model's own runs with large gains.
- Distilling these gains works too.
Blog post: ๐ facebookresearch.github.io/Rโฆ
๐งต1/6
How's Elon Musk not on that list?
Top 100 Most Influential People in History ๐
1. ๐ธ๐ฆ Muhammad
2. ๐ฌ๐ง Isaac Newton
3. ๐ต๐ธ Jesus Christ
4. ๐ณ๐ต Buddha
5. ๐จ๐ณ Confucius
6. ๐น๐ท St. Paul
7. ๐จ๐ณ Ts'ai Lun
8. ๐ฉ๐ช Johann Gutenberg
9. ๐ฎ๐น Christopher Columbus
10. ๐ฉ๐ช Albert Einstein
11. ๐ซ๐ท Louis Pasteur
12. ๐ฎ๐น Galileo Galilei
13. ๐ฌ๐ท Aristotle
14. ๐ฌ๐ท Euclid
15. ๐ช๐ฌ Moses
16. ๐ฌ๐ง Charles Darwin
17. ๐จ๐ณ Shih Huang Ti
18. ๐ฎ๐น Augustus Caesar
19. ๐ต๐ฑ Nicolaus Copernicus
20. ๐ซ๐ท Antoine Laurent Lavoisier
21. ๐ท๐ธ Constantine the Great
22. ๐ฌ๐ง James Watt
23. ๐ฌ๐ง Michael Faraday
24. ๐ฌ๐ง James Clerk Maxwell
25. ๐ฉ๐ช Martin Luther
26. ๐บ๐ธ George Washington
27. ๐ฉ๐ช Karl Marx
28. ๐บ๐ธ Orville Wright and Wilbur Wright
29. ๐ฒ๐ณ Genghis Khan
30. ๐ฌ๐ง Adam Smith
31. ๐ฌ๐ง Edward de Vere
32. ๐ฌ๐ง John Dalton
33. ๐ฌ๐ท Alexander the Great
34. ๐ซ๐ท Napoleon Bonaparte
35. ๐บ๐ธ Thomas Edison
36. ๐ณ๐ฑ Antony van Leeuwenhoek
37. ๐บ๐ธ William T. G. Morton
38. ๐ฎ๐น Guglielmo Marconi
39. ๐ฆ๐น Adolf Hitler
40. ๐ฌ๐ท Plato
41. ๐ฌ๐ง Oliver Cromwell
42. ๐ฌ๐ง Alexander Graham Bell
43. ๐ฌ๐ง Alexander Fleming
44. ๐ฌ๐ง John Locke
45. ๐ฉ๐ช Ludwig van Beethoven
46. ๐ฉ๐ช Werner Heisenberg
47. ๐ซ๐ท Louis Daguerre
48. ๐ป๐ช Simon Bolivar
49. ๐ซ๐ท Rene Descartes
50. ๐ฎ๐น Michelangelo
51. ๐ซ๐ท Pope Urban II
52. ๐ธ๐ฆ 'Umar ibn al-Khattab
53. ๐ฎ๐ณ Asoka
54. ๐ฉ๐ฟ St. Augustine
55. ๐ฌ๐ง William Harvey
56. ๐ณ๐ฟ Ernest Rutherford
57. ๐ซ๐ท John Calvin
58. ๐จ๐ฟ Gregor Mendel
59. ๐ฉ๐ช Max Planck
60. ๐ฌ๐ง Joseph Lister
61. ๐ฉ๐ช Nikolaus August Otto
62. ๐ช๐ธ Francisco Pizarro
63. ๐ช๐ธ Hernando Cortes
64. ๐บ๐ธ Thomas Jefferson
65. ๐ช๐ธ Queen Isabella I
66. ๐ฌ๐ช Joseph Stalin
67. ๐ฎ๐น Julius Caesar
68. ๐ซ๐ท William the Conqueror
69. ๐จ๐ฟ Sigmund Freud
70. ๐ฌ๐ง Edward Jenner
71. ๐ฉ๐ช William Conrad Rontgen
72. ๐ฉ๐ช Johann Sebastian Bach
73. ๐จ๐ณ Lao Tzu
74. ๐ซ๐ท Voltaire
75. ๐ฉ๐ช Johannes Kepler
76. ๐ฎ๐น Enrico Fermi
77. ๐จ๐ญ Leonhard Euler
78. ๐จ๐ญ Jean-Jacques Rousseau
79. ๐ฎ๐น Niccolo Machiavelli
80. ๐ฌ๐ง Thomas Malthus
81. ๐บ๐ธ John F. Kennedy
82. ๐บ๐ธ Gregory Pincus
83. ๐ฎ๐ถ Mani
84. ๐ท๐บ Lenin
85. ๐จ๐ณ Sui Wen Ti
86. ๐ต๐น Vasco da Gama
87. ๐ฎ๐ท Cyrus the Great
88. ๐ท๐บ Peter the Great
89. ๐จ๐ณ Mao Zedong
90. ๐ฌ๐ง Francis Bacon
91. ๐บ๐ธ Henry Ford
92. ๐จ๐ณ Mencius
93. ๐ฎ๐ท Zoroaster
94. ๐ฌ๐ง Queen Elizabeth I
95. ๐ท๐บ Mikhail Gorbachev
96. ๐ช๐ฌ Menes
97. ๐ฉ๐ช Charlemagne
98. ๐ฌ๐ท Homer
99. ๐ฒ๐ฐ Justinian I
100. ๐ฎ๐ณ Mahavira
๐ Note: Ranked by historical influence, not greatness. In the 1992 revised edition, Shakespeare is absent; Hart replaced him with Edward de Vere (#31), arguing he wrote the plays. Only two women appear: Queen Isabella I (#65) and Queen Elizabeth I (#94). Flags indicate modern-day countries of birth or origin.
๐ Source: Michael H. Hart, The 100 (1992 revised edition).
53
Jason retweeted
Whenever I feel demotivated i rewatch this...
Elon Musk in a living room in 2008, 12 days from bankruptcy, pitching a $109,000 car to people...
If money wasnโt a problem, Iโd probably spend my whole life studying.
A masterโs in philosophy, then astrophysics, ancient history, psychology, archaeologyโฆ whatever makes me understand this weird existence a little better.
Not for a career. Not for a degree.
Just because thereโs so much to know, and one lifetime already feels too short.
1
24
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 Agentโs multi-agent capabilities and our infrastructure powering these agent swarms.
Prime Agent now reaches usable input ~13ร faster and uses 83% less startup memory than before.
2
24
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 donโt work yet
They donโt have enough intelligence, theyโre not multimodal enough, they canโt use computers, and they donโt remember what you tell them.
Theyโre cognitively lacking. Itโll 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. Whatโs 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. Theyโre 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. Weโve probably recreated a cortical tissue, pattern-learning and general, but weโre 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.
Thatโs the problem: humans forget just enough to be forced to find patterns.
7. Anything truly new, code thatโs never been written before, ideas that have no template; they stumble
Theyโre still autocomplete engines with perfect recall and no understanding. Until we find that cognitive core, intelligence stripped of memory but full of reasoning, theyโll stay brilliant mimics, not minds.
1
28
No one's talking about how truly mindblowing opus5.5 is, no model i've used before felt that good.
4
123
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.
1
1
52
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?
3
153
Is there a need for an AI Operating system?
No desktop.
No settings.
No shell.
Just a chatbot interface.
1
3
147
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.
1
89
Jason retweeted
One of my favorite lessons Iโve learnt from working with smart people:
Action produces information. If youโre unsure of what to do, just do anything, even if itโs 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.










