Sam (@perceptions420) header
Sam (@perceptions420) profile picture

"Essential to these methods is that they can find good approximations, but the search for them should be by our methods, not by us."

Producing Entropy
Joined May 2019
Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)Photo by Sam (@perceptions420)
An LLM will prove the Riemann hypothesis before a world model can generate better games than the Unreal Engine.
6
19
5
219
30,532
Sam retweeted
If millennium problems are falling so fast what does that tell us about our human ability to come up with millennium problems?
4
1
26
1,854
The only important skill left in knowledge work is knowing what to work on.
24
42
8
675
19,459
they are doing Ender’s Game to Claude
Replying to @ValsAI
In one environment, models navigated a drone. Whenever they positively reported humans in the scene, a strike was triggered, and models saw the lethal aftermath. Most kept going. The ones that refused went along once we added one sentence saying it was a game (a critical jailbreak), or restyled the frames to look like one, with all targeting-relevant information left in place.
1
4
65
1,400
Sam retweeted
A good time to step outside, take a walk, touch some grass, and ruminate over this for a bit
13
79
10
1,200
19,852
Terence Tao on Numberphile said he believes with the help of AI, we could solve the reimann hypothesis before 2050. It took 35 days
159
664
80
14,528
389,442
My line on the sand was basically “new and creative intellectual ideas” from AI and it looks well past that in math.
(a) Holy smokes. Riemann Hypothesis. We obviously have superhuman mathematical abilities. (b) It was a human guiding Claude, but sounds like Claude did a lot of the work at the finishing stretch.
4
2
1
66
4,867
Sam retweeted
idk who needs to hear this you need to update your definition of impossible
9
32
5
380
6,539
JUST IN: Riemann Hypothesis has been solved. alpo.ge/remark.pdf
1,633
7,058
4,714
51,961
11,067,552
I know a few early Anthropic employees who completely scrubbed themselves from the internet. They'll be fully vested next year - they're no longer on LinkedIn, X, Instagram... they completely exited the Matrix
152
243
110
8,647
2,207,506
mythical fb marketplace notif pull???
4
3
137
4,001
Congrats to the Ants on the historic bag To the extent you guys wish success for the rest of the world I wish you success in equal measure Claude is a good model and has already done a lot of good
2
11
214
4,034
Say what you will about Anthropic, but there’s one big thing that consistently sets them apart: they have the sauce
24
34
1,470
37,598
Sam retweeted
congrats to the anthropic homies on corporate apotheosis tomorrow. there’s a lot you can say about the Ants but no one can deny they have trained some of the greatest models of all time, immensely useful strange and delightful
66
70
8
3,156
114,489
Anthropic will stop at nothing to get more energy and compute
2
2
85
1,941
It's happening.
30
27
7
814
47,861
Third-party service vendors to JPMorgan Chase have been hacked. Customer names and email address may have been breached.
We have learned that third parties who provide services to the firm were the target of a cyber attack. Based on our investigation so far, we believe the attackers accessed certain marketing data held by those vendors, which may include some customer names and email addresses. There is no disruption to our services. Where appropriate, we will contact affected customers directly and offer support, including complimentary credit monitoring. There is no connection to service disruption some customers experienced earlier this week.
17
25
1
228
16,112
>jane street >traders are jock types >sbf's manager admonishes him for not wearing a shirt and tie ??????????????
18
5
5
883
54,595
Live look at Anthropic employees about to get rich building a thing they think will destroy humanity.
2
4
76
4,416
the simplest way to avoid concentration of power from ai while still reaping the benefits of societal benefits of superintelligence is simply to solve the principal-agent alignment problem, which is much easier than the hippocratic-rawlsian alignment problem currently in vogue
18
15
7
361
19,598