Joined March 2023
"it's only predicting the next token" has done tremendous damage to the discourse. Next token OF WHAT SEQUENCE, MOTHERFUCKER? The source of the training signal can be literally whatever. And it's largely not any directly human-written text anymore.
LLMs cannot evolve to superintelligence via recursive self-improvement because that would require it to invent a new technology unknown to man But that’s impossible because the entire LLM technology is based on predicting what man will think/say/do Skynet isn’t happening w/LLMs
11
12
368
12,766
worrying about recursive self-improvement on australian TV (lateline) in 1996.
142
561
94
4,031
373,572
All of that seems correct?
Essentially none of Darios "time-lined warnings" have played out yet..
5
5
336
19,324
(((Palee))) retweeted
I feel like I'm taking crazy pills with how many smart people I read and respect are saying stuff like this. The idea that the AI labs are suddenly, only now, just in September of 2026, advocating for AI safety, and that they just came around to this position bc of sudden financial precarity, is just completely, utterly, conclusively, 100% wrong. Here's Dario Amodei telling Ross Douthat in February that he agrees with the case for slowing down; that he's in favor of "collaborating" internationally to organize a slowdown; that "I would be all for" a global slowdown. This was 7 months ago, when Anthropic's annualized recurring revenue was rising faster than any company in modern history. He's been saying stuff like this for years, when Anthropic was worth millions of dollars and when Anthropic was projected to IPO for trillions.
fwiw I think the actual reason leaders of AI frontier labs suddenly seem to agree to pace AI development is (a) safety measures are currently so crap that they're likely to end up in court if not jail, and they all agree that no one wants that (b) they see no major new model advancement coming up soon anyway and need an excuse (c) shift in public opinion the talk of existential threat is mostly there to keep you distracted and the stock market happy.
244
252
29
2,276
413,316
It's hard to take these people seriously when they've been predicting a plateau or a "burst" that never happened for the past TWO YEARS
Hi, I’m here to piss you off. The fact that bunch of SF morons drank their own koolaid does not mean it’s true. These guys also genuinely believe that wealth “trickles down” and that an AI deity will torture a digital simulation of their soul for all eternity if they misbehave
9
29
5
651
22,257
(((Palee))) retweeted
this talk of “preventing human extinction” is obviously a front for the ai companies to increase their customer base by ensuring humanity continues to exist and grow
8
36
3
619
12,039
(((Palee))) retweeted
we're cooked
86
390
44
14,464
460,957
(((Palee))) retweeted
If you are not a e/acc at GPT-4, you have no heart. If you are not a doomer at GPT-6 you have no brain.
55
50
12
1,009
30,155
(((Palee))) retweeted
Look at him. That’s my quant
I gave the fly brain $100 to trade bitcoin. Dopamine neurons are stimulated when the fly makes profit. Neuron activity controls buy/sell decisions and makes trades on coinbase. Will the fly get rich?
17
214
3
6,545
383,598
im on my knees bloodied an Anthropic killbot pushes a gun into my temple "any last words?" "coxon was right" "coxon?" i grin through broken teeth "coxon your face" *BANG*
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
61
310
19
7,552
398,711
(((Palee))) retweeted
whoever prompted this needs a Grammy the official AI BUBBLE anthem 💀
most people still have NO clue how good AI music has gotten. we are so cooked 😭
7
21
3
254
9,667
(((Palee))) retweeted
Claude-Pop - I'm Upping My P(Doom)
134
233
156
2,647
749,977
(((Palee))) retweeted
It's great to have Jacob's "AI is >10% to kill us and I'm quitting" at the same time as Evan's "AI is >10% to kill us and I'm staying", so we get to see all of the "If you really believed that, you'd stay" cope and the "If you really believed that, you'd quit" cope simultaneously
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
30
98
7
1,827
53,840
(((Palee))) retweeted
Replying to @DouglasYaoDY
Babe are you okay you barely touched the psychiatric medication I synthesized in the garage for you
21
161
3
5,581
86,478
(((Palee))) retweeted
That's right! Audio spectrography goes in the transformer hole.
So Astra is able to identify sounds from mel spectrograms zero-shot. I don't think we've scratched the surface of what this model can do (and this is light reasoning btw)
21
147
11
3,496
111,568
(((Palee))) retweeted
Thank you for your family's contributions to literature
1
1
12
7,897
(((Palee))) retweeted
tapping the sign
GPT-6 Astra scored 95% on a robot control task, up from Fable 5.1's 40%, with 6.2x fewer output tokens at 2.3x lower cost. 🧵
14
138
10
2,933
132,584
. . . This appears to be the result of a single person who MAY OR MAY NOT BE JEWISH being cast in Heated Rivalry. Just a person with a Jewish last name. He doesn’t even have a Wikipedia page. Absolutely floored.
unavailable
38
52
83
2,093
422,394
Russia is basically everything that people accuse Israel of being but like actually.
56
190
29
2,474
41,321
The play button already knew where this was going.
112
351
26
3,222
52,204