Shibl Mourad (@shibl) profile picture

DeepMind Canada Engineering Lead. Interested in philosophy, cocktails, ramen, machine learning, manga and computer science, not necessarily in this order.

Montreal
Joined August 2008
Shibl Mourad retweeted
This is so good.
127
1,116
79
8,385
245,031
Replying to @PAHoyeck
This isn't accurate enough, they both must end their lives after the mere mention of Hegel.
1
1
19
653
What an interview question. This is remarkable TV. ‘What do you think happens when you die?’
483
879
356
10,233
2,967,664
PhD Proposal -> PhD Thesis
5
54
9
575
20,972
Finally I have a good way of checking whether numbers are integers. This will be a life-changer.
hello there ℤ is Diophantine in ℚ, thanx to my close friends Astra and Claude The polynomial characterizing ℤ isn’t quite Tweet-sized; it has m=292,890,231,905 variables, is of degree d < 2^^7, and all its coefficients have absolute value < 2^^8.
8
24
524
11,142
Heidegger had already warned us about technology.
23
194
20
1,047
57,075
Dev life before agents Dev life after agents
166
134
67
1,928
163,774
Shibl Mourad retweeted
Mathematics has had peer review for centuries. In 1732, Euler demolished Fermat's conjecture with a counterexample.
15
55
4
1,012
59,836
Shibl Mourad retweeted
Replying to @paulg
and then we trained llms on those essays
1
2
65
4,625
Shibl Mourad retweeted
2
17
1
263
5,454
100% agree, we focus on group 1 and 2 but the real exciting work in happening in group 3.
AI will create 3 types of PhD students: Those who refuse to use it. Those who use it to publish more papers. Those who use it to discover things that were previously impossible. The first group falls behind in productivity. The second gets more publications, citations, and academic recognition. The third might actually push the boundaries of human knowledge. The irony? Academia is still designed to reward the second group.
1
5
148
Shibl Mourad retweeted
can’t believe they built the pyramids without MCPs
55
90
9
1,035
30,937
It finally struck me why we are all so negative about AI: loss aversion. Even if AI can help us with many things we care abiut, the fact that we might lose one thing makes very anxious. More generally: all change, however positive is painful.
2
4
241
Protect the real data centers 📚
110
3,067
147
14,253
192,372
Shibl Mourad retweeted
the hardest part about building a startup isn't product or sales its cringe-posting on linkedin
103
70
12
1,521
50,828
71
1,676
56
10,662
214,834
Shibl Mourad retweeted
I can’t believe we’re two months away from 2027 and the easiest way to save something as a PDF is still to lie to your computer and tell it you’re about to print it.
402
8,456
282
137,427
3,075,367
I once tried to explain Bayesian reasoning at a dinner party. By dessert, everyone had updated their beliefs about inviting me again.
33
248
15
3,477
66,366
Great work by @mhutter42 and others to untangle the cluster of things that pop up when we talk of consciousness.
"Assessing AI Consciousness" is now available as a book: hutter1.net/ai/bayesconsc.ht… It develops a Bayesian framework separating theoretical credence from empirical evidence, providing a precise, scalable method for calibrated assessments (with online tool and code): ai-cognition.org/cacophony-t… Happy Halloween.
2
70
Shibl Mourad retweeted
The problem with AI is that it’s causing this massive brain drain where our brightest kids are working on curing cancer and creating abundant energy instead of building derivatives exchanges
83
319
27
6,326
172,371