
Vincent Favilla
896 posts

Vincent Favilla
@vincentfavilla
Psychology prof thinking about minds, models, and meaning. Watching with cautious optimism as AI devours the world.
San Francisco Katılım Mayıs 2022
223 Takip Edilen2.1K Takipçiler

@greg_ip That quote isn't in the paper. The study was 36 adults and had no measure of letter recognition (it was typing with one finger and no screen feedback).
The b/d claim isn't in the paper either. There's a published critique at doi.org/10.3389/fpsyg.…
English


The vibes are off. For me it's making more frustrating mistakes than 4.6 and it's not as good at creative tasks either. Cost/usage is significantly higher due to its new tokenizer but I've been running on high/extra high thinking. Going to shift down to medium to today and switch to Cursor for a month if things don't improve.
English

@AndyMasley It's times like this you really wish youtube had community notes. Thanks for fighting the good fight against misinformation.
English

Two months ago the YouTuber Benn Jordan made what has become one of the most popular videos about data centers ever, and maybe the most popular piece of media on data centers this year: Datacenters Behaving Like Accoustic Weapons. The claim is that they produce harmful infrasound.
This video (and the one before it) is a moment-by-moment disaster. I found this to be far and away the most jaw-dropping experience of writing I've had. To my knowledge I'm the first one to publicly push back against all the problems with it.
Even if you're not interested in data centers or infrasounds, I think this is just an incredible example of how pseudoscience can become highbrow misinformation. All it takes is a chill-seeming guy, great production, and not actually checking literally any of the studies he's quickly flashing on the screen.
blog.andymasley.com/p/contra-benn-…
English

@Miles_Brundage Basically none of these hype/scam claims are falsifiable.
"Our AI is has risks"? Marketing.
"Our AI is safe"? Also marketing.
AI does something impressive? Cherry-picked.
AI does something dumb? Proof it's all fake.
English

Embarrassing for Bloomberg IMO
At what point will AI capabilities be strong + obviously misusable enough for journalists to acknowledge it's not hype?
Better than most security vulnerability researchers isn't enough, it has to beat every single one? x.com/opinion/status…
Bloomberg Opinion@opinion
When tech CEOs like Sam Altman and Dario Amodei acknowledge the risks of AI, they’re selling you a product. @parmy explains the dark art of AI marketing 🎥
English

@AndyMasley Buildings are hot and it's making the chickens sad
English

They say that when a data center comes to town, the crows begin to sing a strange atonal tune, the grass no longer reflects the moonlight and stays pitch black, and the children begin to whisper secrets instead of singing songs
Clayton Tucker@ClaytonTuckerTX
I’m hearing reports that the noise from AI Data Centers cause: —chickens to lay 50% less eggs —cattle/goats/sheep to lose 30% of their body weight —horses & wildlife to suffer from stress If true, we can have farms or data centers. I choose farms. What about you?
English

@tomhfh I made a tool to estimate how many people are dying annually from depressing architecture (my guess, about 50 to 100):
claude.ai/public/artifac…
English

@paulnovosad One thing I think is worth considering is that "screen work gets automated" doesn't just mean that portion of the economy grows faster. It also likely means find a lot more economically valuable work to do on screens.
English

AI has an insane rate of progress when we observe what it does for jobs that take place on screens.
Progress is slower for jobs in the world of atoms.
Since the intellectual class spend all their time on screens, it's easy to think "AI will do all the jobs!"
We are like 15th century farmers thinking about mechanization. If you mechanized agriculture, what other jobs could there even be?
Perhaps employment in the on-screen sector will shrink to a small percentage.
Of course stuff on screens affects things in the real world too. But the rate of progress is different. On screens (i.e. in simulation), the self-driving car problem has been solved for many years. IRL, we are still years and years away from replacement of human drivers.
I think one reason for divergent predictions on how much AI will matter is that the "6% growth" AI experts are extrapolating from stuff on screens to stuff with atoms, while the "3% growth" economists are not.

English

@recap_david I've been thinking about this too because I've experienced the same thing. I'm suddenly working a lot more because the stuff I'm doing is a lot more valuable.
I'm much less certain this pattern will hold in the longer term though (5-10+ years).
English

I need to rant about something because I keep seeing the same brain-dead take over and over again. "AI is going to take all our jobs."
No. No it is not.
AI is not going to make people work less. It's going to make people work MORE.
I know this because I'm living it. Right now. Today. I am working more hours than I have at any point in my life. Not because I have to. Because I literally cannot stop.
I'm doing it voluntarily. Happily. Obsessively.
This is also true of everyone I know that is deeply involved in AI.
When you sit down and realize you can go from idea to execution in HOURS with no dependencies on anyone else — no designer queue, no engineering sprint, no "let's circle back next week" — your brain breaks in the best possible way.
You just keep going. You build one thing. It works. You build the next thing. That works too?! And suddenly it's midnight and you don't care because you just brought five ideas to life that would've taken you 3 MONTHS six months ago.
Every builder I know is experiencing this same addiction right now. We're all sleeping less and producing more and enjoying every second of it.
The value of one hour of human input has gone up by an order of magnitude. So what happens when your input becomes 10x more valuable? You don't do less of it. You do WAY more. Because the incentives are insane.
The "AI takes jobs" crowd is making the same mistake people have made with every single technology in history. They're assuming there's a fixed pie of work. There isn't. There never was. The pie grows. It always grows.
And AI is about to make it grow faster than anything we've ever seen.
More work. More jobs. More builders. More opportunities. More humans doing more ambitious things than they ever thought possible.
This is the beginning of the most productive era in human history and most people are too busy doom-scrolling to notice.
Bookmark this.
English

@TheZvi Gemini is incredulous it's 2026 in general. But even Claude is skeptical of the name and thinks it's a rhetorical choice.
English
Vincent Favilla retweetledi


@sama Axios reported the Pentagon's final offer to Anthropic included collecting Americans' geolocation, browsing, and financial data from data brokers. That's lawful and arguably not 'unconstrained monitoring.' Does your contract permit it?
English
Vincent Favilla retweetledi

I'm speechless at OpenAI releasing that contract excerpt and acting as if there aren't gaping holes that could be exploited far beyond their stated "red lines." I'm not a lawyer, but this is pretty obvious and common sense.
(And to be clear: if Google had signed the same deal, I'd be saying the same thing internally. The issues here are bigger than friendly competition between companies.)
OpenAI's "red lines" are: no mass domestic surveillance, no directing autonomous weapons, and no high-stakes automated decisions. They argue their cloud-only deployment + safety stack + cleared OpenAI personnel "in the loop" make violations impossible. They also claim the contract references the relevant laws/policies "as they exist today" so future changes won't weaken the standards.
But the actual language they published is still full of obvious escape hatches. This is why Anthropic refusing to sign makes sense. Reporting on the Anthropic–"DoW"/Pentagon standoff described them saying the proposed contract language was framed as compromise but paired with "legalese that would allow safeguards to be disregarded at will."
You don't need to agree with Anthropic on everything to see what they're reacting to: language that sounds like ethics but cashes out as essentially "subject to whatever the government decides later."
## Autonomous weapons
The problem is that the restriction is conditional: it depends on what "law/regulation/policy requires human control" for. If policy definitions are weak (or later revised), the contract language itself doesn't read like a durable "no autonomous weapons" ban. It reads like "we'll follow whatever the current regime says requires human control."
OpenAI says elsewhere that the agreement "locks in" today's standards even if laws/policies change. If that "freeze" clause is real and enforceable, sure, but it's not visible in the excerpt itself, so the excerpt alone doesn't justify the level of confidence they're projecting.
## "High-stakes decisions"
Same loophole. This forbids only decisions that already require human approval under whatever authorities apply. If a decision doesn't formally require approval (or can be reclassified/reshaped), the clause doesn't obviously prohibit automation of the step that matters.
## Surveillance
"directives," "purpose," and "unconstrained" are squishy on purpose:
"DoD directives" aren't laws; they're internal policy. That matters because we have real precedents for administrations leaning on aggressive internal legal/policy interpretations as a shield until courts/politics catch up.
If you think "secret memos" is alarmist, look at the pattern:
1. Reporting in early 2026 described a previously hidden DHS/ICE legal memo position asserting warrantless/forced home entry under certain circumstances, which is the kind of internal-lawyer move that tends to get written, circulated, and only later litigated and retracted.
2. And historically, the Bush-era OLC torture memos are the canonical example of "legalistic compromise" that later turned out to be a moral and legal disaster. (You don't have to litigate the details to make the point: internal legalese can be used to launder outcomes.)
"Unconstrained" is not a real safeguard. Surveillance can be huge while still "constrained" by selectors, categories, time windows, or a stated "foreign intelligence purpose." And it only covers private information, so not the massive world of public data that can still be used for profiling, targeting, and "pattern-of-life" analysis at scale.
## Domestic law enforcement
> shall also not be used for domestic law-enforcement activities except as permitted by the Posse Comitatus Act and other applicable law.
This is not a hard prohibition. "except as permitted" is not a ban. It's a permission for exceptions, and "other applicable law" is an open-ended bucket by design.
If you want a concrete, recent example: the Associated Press reported that formal orders extended the Washington, D.C. National Guard deployment through Feb. 28, 2026, to protect federal property/functions and to support federal and D.C. law enforcement.
That's exactly the sort of "domestic deployment supporting law enforcement" scenario where this clause stops sounding like a "red line" and starts sounding like legal throat-clearing.
## "Cloud-only / no edge deployment prevents autonomous weapons" rings false
OpenAI's own argument is: cloud-only (no edge devices) means you can't power autonomous weapons. But that's not convincing.
You don't need GPT-5.2 running on the missile. You can use a cloud model for high-level decision-making (tasking, prioritization, target recommendation, mission planning) over a satellite link (Starlink or otherwise), while a separate local system handles actual guidance and execution.
High latency is totally compatible with "strategic / operational" autonomy while still enabling lethal outcomes.
Once the pattern exists, "additional safety layers" are a policy choice and implementations change, exceptions get made, but today's contract language tends to get "grandfathered" into tomorrow's contract template.
So layered safeguards can reduce risk today, but the contract language itself is exactly the kind of "looks strict, bends easily" compromise that becomes precedent.
And creating precedent is the real problem here.




English

@mountainviews @SecWar We're saying the same thing! The deranged text is Hegseth's, not Anthropic's.
English

@vincentfavilla @SecWar Not to insult your ability to reason but are you aware that saying
"Anthropic is a supply chain risk to national security."
Also: "Anthropic must continue providing us services for six months."
Is in direct contradiction. Not understanding this is pathetic
English

This week, Anthropic delivered a master class in arrogance and betrayal as well as a textbook case of how not to do business with the United States Government or the Pentagon.
Our position has never wavered and will never waver: the Department of War must have full, unrestricted access to Anthropic’s models for every LAWFUL purpose in defense of the Republic.
Instead, @AnthropicAI and its CEO @DarioAmodei, have chosen duplicity. Cloaked in the sanctimonious rhetoric of “effective altruism,” they have attempted to strong-arm the United States military into submission - a cowardly act of corporate virtue-signaling that places Silicon Valley ideology above American lives.
The Terms of Service of Anthropic’s defective altruism will never outweigh the safety, the readiness, or the lives of American troops on the battlefield.
Their true objective is unmistakable: to seize veto power over the operational decisions of the United States military. That is unacceptable.
As President Trump stated on Truth Social, the Commander-in-Chief and the American people alone will determine the destiny of our armed forces, not unelected tech executives.
Anthropic’s stance is fundamentally incompatible with American principles. Their relationship with the United States Armed Forces and the Federal Government has therefore been permanently altered.
In conjunction with the President's directive for the Federal Government to cease all use of Anthropic's technology, I am directing the Department of War to designate Anthropic a Supply-Chain Risk to National Security. Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic. Anthropic will continue to provide the Department of War its services for a period of no more than six months to allow for a seamless transition to a better and more patriotic service.
America’s warfighters will never be held hostage by the ideological whims of Big Tech. This decision is final.
English

@axios This pretty much confirms that it was never really about the safeguards, it was about making an example out of Dario for not bending the knee.
English

NEW: The Pentagon has agreed to OpenAI's rules for deploying its technology safely in classified settings, though no contract has been signed, a source tells Axios.
The department appears to have accepted conditions similar to those put forth by Anthropic. axios.com/2026/02/27/pen…
English

@KelseyTuoc That's got to be the most belligerent capitulation I've ever seen. "Fine, we'll just stop using Anthropic models, but we'll throw you in jail if you don't salute hard enough on your way out."
English

@USWREMichael @AnthropicAI So you're telling us to be afraid of a company's publicly transparent ethics document. I find it concerning that "be helpful and don't harm people" keeps you up at night.
English

Imagine your worst nightmare. Now imagine that @AnthropicAI has their own “Constitution.” Not corporate values, not the United States Constitution, but their own plan to impose on Americans their corporate laws. Claude's Constitution \ Anthropic. anthropic.com/constitution
English

@USWREMichael @DarioAmodei @DeptofWar When you can't argue the substance, attack the character. This tantrum is embarrassing.
English

It’s a shame that @DarioAmodei is a liar and has a God-complex. He wants nothing more than to try to personally control the US Military and is ok putting our nation’s safety at risk.
The @DeptofWar will ALWAYS adhere to the law but not bend to whims of any one for-profit tech company.
Axios@axios
Anthropic rejects Pentagon's "final offer" in AI safeguards fight trib.al/Bc8hiSp
English







