Should the SEC step in?
"The AI ecosystem’s circular business dealings, extensive use of financial engineering, and otherwise murky finances make it hard for investors, regulators, or workers saving for retirement to get an accurate financial picture of AI companies."
"The AI ecosystem’s circular business dealings, extensive use of financial engineering, and otherwise murky finances make it hard for investors, regulators, or workers saving for retirement to get an accurate financial picture of AI companies."
"The SEC should block Anthropic’s stock market debut until the company provides two consecutive quarters of audited financial statements, with all material information about the company’s spending commitments disclosed." techpolicy.press/the-sec-should…
0
2
2
155
M. Alex O. Vasilescu retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: fronely read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weight-sp…
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: fronely read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weight-sp…
1,889
5,713
45,414
5,292,456
M. Alex O. Vasilescu retweeted
Uhhh. I recall Reuters reported that they saw Anthropic's Q3 numbers but wasn't included in the first filing. Now we know why ...
ANTHROPIC Q3 2026 FINANCIAL HIGHLIGHTS:
- Revenue: $17.0B, +1,127% YoY & 47% QoQ
- Oper Loss: $567M (vs. $974M profit in Q2)
- Net Loss: $23.3B, vs. $12.1B YoY
- Adj. Oper Income: $2.5B
- Adj. Oper Margin: 15%, down from 19% in Q2
- 9M Revenue: $33.2B
Anthropic attributed its retupartly to $5B in higher-cost short-term compute and a $2B noncash donation-matching charge.
Its nine-month net loss reached $104B, largely reflecting $94.3B in noncash charges from the rising value of its convertible notes.
Anthropic also disclosed approximately $895.4B in aggregate purchase and lease commitments as of and after Sept. 30, largely tied to AI compute infrastructure.
10
39
528
128,768
M. Alex O. Vasilescu retweeted
told ya so. and people like @altcap could have figured this out and chose not to.
that should tell you all you need to know about who you should trust and who you shouldn’t.
that should tell you all you need to know about who you should trust and who you shouldn’t.
Uhhh. I recall Reuters reported that they saw Anthropic's Q3 numbers but wasn't included in the first filing. Now we know why ...
9
28
164
27,185
WTF?
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: fronely read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weight-sp…
8
12
211
28,850
@erikbryn If GPT-6 Astra can
+ inspect weights (its own or a copy’s),
+ diagnose weaknesses,
+ write functional edits that improve capability,
then this is likely a primitive associated with recursive self-improvement (RSI).
+ inspect weights (its own or a copy’s),
+ diagnose weaknesses,
+ write functional edits that improve capability,
then this is likely a primitive associated with recursive self-improvement (RSI).
0
0
5
383
Jean-Luc Picard remains the single most compelling vision of progressive politics I am familiar with — secular, rational, compassionate, egalitarian; a believer in progress & abundance but also duty, hard work, and a constant battle for self-improvement and improving the world.
787
1,151
10,122
935,521
That would be burden shifting if Picard were claiming, “Data is conscious, and you must prove otherwise.” But he explicitly isn’t. He says, “I don’t know.”
Picard’s argument is about what we are justified in doing under uncertainty. Maddox wants to override Data’s refusal and dismantle him on the premise that Data is merely property and lacks morally relevant consciousness. Picard shows that Maddox cannot actually establish that premise. You don’t need to prove that someone is conscious in order to argue that uncertainty about their consciousness should constrain an irreversible action against them. That isn’t a scientific burden of proof error.... it’s an ethical and legal argument about the consequences of being wrong.
Picard isn’t saying, “Prove Data isn’t conscious.” He’s saying, “You’re proposing to treat him as though you know he isn’t. Do you actually know that?” That distinction is the entire point of the scene.
Picard’s argument is about what we are justified in doing under uncertainty. Maddox wants to override Data’s refusal and dismantle him on the premise that Data is merely property and lacks morally relevant consciousness. Picard shows that Maddox cannot actually establish that premise. You don’t need to prove that someone is conscious in order to argue that uncertainty about their consciousness should constrain an irreversible action against them. That isn’t a scientific burden of proof error.... it’s an ethical and legal argument about the consequences of being wrong.
Picard isn’t saying, “Prove Data isn’t conscious.” He’s saying, “You’re proposing to treat him as though you know he isn’t. Do you actually know that?” That distinction is the entire point of the scene.
1
0
0
30
In the script, JAG had already ruled that Data was property, and Riker cemented the baseline by detaching Data's arm and flipping its off switch. Both easily reversible actions.
Picard had the affirmative burden to overturn the default status of manufactured hardware.
Instead of positive evidence for consciousness, he offers a hypothetical: “What if he meets the third, consciousness in even the smallest degree? What is he then? I don't know. Do you? Well, that's the question you have to answer.”
The hypothetical supplies no additional evidence of consciousness. Calling the argument decision making under uncertainty does not fill that evidentiary gap.
Any number of unsupported possibilities can be introduced with “what if.” The hypothetical itself establishes nothing.
Picard had the affirmative burden to overturn the default status of manufactured hardware.
Instead of positive evidence for consciousness, he offers a hypothetical: “What if he meets the third, consciousness in even the smallest degree? What is he then? I don't know. Do you? Well, that's the question you have to answer.”
The hypothetical supplies no additional evidence of consciousness. Calling the argument decision making under uncertainty does not fill that evidentiary gap.
Any number of unsupported possibilities can be introduced with “what if.” The hypothetical itself establishes nothing.
1
0
0
16
M. Alex O. Vasilescu retweeted
Where does a machine learning engineer go camping?
The first ML joke I've heard is from Andrew @AndrewYNg .
The first ML joke I've heard is from Andrew @AndrewYNg .
1
4
11
1,019
@mattyglesias That is not how science works.
The burden of proof falls upon Picard to devise falsifiable tests for consciousness. It is not the duty of the other person to devise the tests or to prove that Data lacks consciousness. Picard shifted his burden of proof.
The burden of proof falls upon Picard to devise falsifiable tests for consciousness. It is not the duty of the other person to devise the tests or to prove that Data lacks consciousness. Picard shifted his burden of proof.
1
0
0
28
We are not building data centers, we are building "AI compute campuses." The name is not the problem.
axios: OpenAI will start calling its data centers “AI compute campuses” in a policy push launching after the midterms.
The rebrand arrives as data centers become a liability in both parties
To defuse that opposition, OpenAI will offer community compacts and propose state dents a real say in how projects are built and what benefits they receive.
The push opens in the lame-duck session, as AI companies face a harsher Washington with Democrats likely to take the House.
0
0
1
108
M. Alex O. Vasilescu retweeted
Since the very first day of generative AI's release to the public in 2023, it was glaringly obvious that their financial plan was beyond unsustainable. Anyone who thought it wasn't was either silly or intentionally duplicitous.
59
80
1,243
46,690
M. Alex O. Vasilescu retweeted
The World Will End If Nvidia Misses Earnings zerohedge.com/markets/world-…
182
179
1,670
216,933
M. Alex O. Vasilescu retweeted
Shame on @DarioAmodei / Anthropic. I never distilled Claude—rollouts came from a strong open-source model; Claude just helped submit jobs and run analysis. Still got tagged for "training a competitor," and now anything touching RFT, SFT, or training larger models gets me blocked, Opus 5.5 downgraded to Opus 5, or the session killed outright. You built your models on the world's data, then treat paying users doing research automation as the enemy. Pathetic.
122
244
4,035
308,006
M. Alex O. Vasilescu retweeted
Yann LeCun says:
People in AI have made the same mistake for 70 years: if a machine can do a task humans see as intellectually complex, it must be smarter than humans. This is completely false
We have computers that are superhuman at many tasks but we're still far from machines al-level intelligence
People in AI have made the same mistake for 70 years: if a machine can do a task humans see as intellectually complex, it must be smarter than humans. This is completely false
We have computers that are superhuman at many tasks but we're still far from machines al-level intelligence
107
183
1,395
123,094
M. Alex O. Vasilescu retweeted
The government's job is to protect the market from encroachment, not do the encroaching itself.
5
4
43
4,405
M. Alex O. Vasilescu retweeted
@jimcramer Anthropic are doing that themselves by taking on so much debt that they are expected to get a junk rating soon
0
1
2
788
M. Alex O. Vasilescu retweeted
@jimcramer No well-meaning person is trying to derail anything. If there is a problem, highlight it; otherwise, celebrate the success.
This whole bulls v/s bears discussion is childish at best. Was Warren Buffett a bad person for not investing in certain companies because they were crap?
This whole bulls v/s bears discussion is childish at best. Was Warren Buffett a bad person for not investing in certain companies because they were crap?
1
1
2
885
M. Alex O. Vasilescu retweeted
@jimcramer Bears trying to derail an IPO is like trying to stop a train by standing on the tracks and yelling 😂
1
1
2
198