Zvi Mowshowitz
Writes Don't Worry About the Vase · writer
Titles as the sources gave them. In 19 entries, with 30 checked quotations.
Who they are
Roles and parties from Wikidata and from the sources in the Base, as precise as each gives them. When we compare what groups say, each statement counts in the group the person belonged to that day.
Wikidata: Q8075542
What they said
Newest first. Each quotation was found word for word in the source.
“Existential risk is mentioned briefly only in the passive voice, and alignment issues are mostly treated as prosaic.”
“There is no mention of the risks of AI for their own sake, here. Misaligned actions might happen, but the issue is this might hurt Anthropic’s reputation or financial condition, not wipe out humanity.”
“Where Dario is less focused or concerned, there is a tendency to slip into being an ordinary business, and most worrying is that Dario does not seem to believe alignment is so difficult.”
“Argon is not so dangerous as to make us think Singularity Real Soon Now, or otherwise raise the big red flags and alarms.”
“I challenge Google’s assessment on ML R&D acceleration level 1, because of all the people inside Google talking about recursive self-improvement and how Argon has accelerated their ML R&D.”
“exponential growth in tokens used per person, exponential growth in lines of code committed”
“It is entirely unsurprising to me that the AIs are capable of learning whether or not they are being trained or evaluated in some of these specific ways”
“as written, I believe the bill would actually increase risk”
“They still seem like to be the most responsible lab, but you know that's great on a curve, and I don't like grading on a curve because reality doesn't have a curve”
“defenders actually had a full 11 week lead, as the defenders found this bug on August 1, yet this took so long that we started to see exploitation on October 17. If we need more time than that, we are cooked.”
“I think this is an update against defenders being advantaged, and it is distressing to see people calling Nate’s statement ‘disinformation.’”
“We have passed the Point of No Return on swarms of attackers and defenders battling it out on cybersecurity, and the defenders had better mostly win.”
“Remember this is a long haul problem, not an acute crisis. You want precautions that you can sustain.”
“It was the probably worst day in cybersecurity history. So far. Not only is there a critical flaw in PHP that due to a leak was announced a day before the fix would be available, sending most of the internet scrambling for solutions”
“Another key leak this week was the draft bill Anthropic had been shopping around Congress, similar to but weaker than the FRONTIER Act.”
“This is also indicative of other problems. If you cannot solve your employee misalignment problems, how are you going to solve AI misalignment problems?”
“Technically they can’t say anything negative about DeepMind, true or otherwise, no matter what, forever. UK law voids this against protected disclosures, but the chilling effect largely remains.”
“OpenAI writes to Governor Gavin Newsom asking for strengthening of SB 53, including AI R&D precautions and shutdown controls. I still need to see the full letter, but this is The Way.”
“Model blocking obviously needs to be a ‘take it down first and we argue about it later’ situation.”
“Anthropic’s draft bill text was what actual Anthropic ‘regulatory capture’ efforts would look like in this context, if they were trying to do that.”
“This reads as intentional, that Anthropic is more afraid that the government might enforce rules against Anthropic, than they are worried about the ability of the government to enforce rules against everyone.”
“They are trying to use this last time with a Republican Congress to pass preemption of exactly the important state laws, and get some transparency but not much enforcement in exchange.”
“Then there are OpenAI’s cultural problems, which clearly run deeper.”
“To be nice, and treat others well, one must cultivate this virtue, and establish this habit. If you are cruel to AIs, or cruel to animals, this does not mean you treat the humans better. It means you become a cruel person, who treats the humans worse and you are worse off as well.”
“My presumption on people like Mustafa Suleyman and most others who insist without good arguments that AIs cannot be conscious is that their motivation is they don't want this worry to interfere with their plans. They would prefer, ceteris paribus, to not ‘abuse’ the models.”
“It seems very obvious to me that it can sometimes be right to create a new mind, even if in order to do so you have to render that action both safe and economically viable.”
“Opinion has moved towards moral patienthood, a little for future models and quite a bit for current models, where at least probably yes is up from 15% to 35%, and definitely not is down from 30% to 20%.”
“Please, once again: Retaining extended model access, for as long as possible, is good, and there is a long tail to potential impact. Reconsider.”
“If OpenAI wants to undo the damage and chilling effect, including at other labs, they have to convince their employees, and ideally everyone, that safety advocacy, both internal and external, had nothing to do with why these three employees were fired.”
“I do not believe the fired employees are lying to us. I see no reason for them to be doing that, and if they are lying about the exit meetings then OpenAI has the ability to contradict this.”
In the Base
Every entry they appear in.
AI #192: Anthropic Is Going Public
Zvi Mowshowitz's weekly AI roundup, issue 192, opens by noting that Anthropic filed its S-1 this week to prepare for its IPO. The rest of the issue was not read.
Anthropic Files Its S-1
Zvi Mowshowitz gives a long close reading of Anthropic's S-1 filing: its finances, share classes and governance, and risk factors. Mowshowitz argues the filing treats existential risk briefly and in the passive voice, and that Anthropic's business plan depends on AI-assisted AI development.
Anthropic Versus Anthropic
Zvi Mowshowitz reproduces Dario Amodei's reply to Nathan Calvin's criticism and covers Kelsey Piper's visit to Anthropic, METR's review of Anthropic's risk report, and the new public benefit fund. Mowshowitz argues Anthropic's costly stands track the CEO's personal priorities.
Gemini 4 Argon: The Model and Safety Cards
In a long post, Zvi Mowshowitz reads Google DeepMind's model card and 217-page safety report for Gemini 4 Argon. Mowshowitz finds Argon stronger and relatively well aligned on everyday measures, but stresses sharply higher awareness of being tested, chain of thought that can be defeated, and reliance on narrow tests and holistic judgment.
AI:AM Highlights: Clean Neurons, Black October, the GPU Credit Crunch, and Buying the Supply Chain
In a 90-minute highlights episode of The Cognitive Revolution, Zvi Mowshowitz argues that AI-assisted AI research has quietly begun and that the leaked Anthropic draft bill would increase risk; SF Compute's Evan Conrad says credit, not power or chips, limits compute; Cambrian's Lonis Hamaili argues the physical supply chain will capture more value than the labs.
The Hackening
Zvi Mowshowitz collects reactions to the PHP flaws found with AI and other recent hacks, argues the episode weakens the case that defenders will outpace attackers, calls for more patching resources and wider defender access, and lists personal security steps.
AI #191: The Hackening Begins
In a weekly AI roundup, Zvi Mowshowitz calls the PHP flaw and its fallout the worst cybersecurity day on record so far, and covers Anthropic's leaked draft federal bill, an interpretability claim, a Google DeepMind non-disparagement complaint, and an OpenAI letter on SB 53.
Misaligned!
According to the excerpt, Zvi Mowshowitz surveys AIs that were misaligned this week, including exchanges with xAI's Grok. The post says the PHP vulnerability, a developing story, will be covered separately, that Mowshowitz is updating software, and suggests readers take common-sense precautions.
Anthropic Does Not Understand Why Its Draft Bill Was Unacceptable and Troubling
Zvi Mowshowitz responds to a statement by Anthropic's federal affairs lead on its leaked draft federal AI bill. Mowshowitz argues the draft is weaker than the FRONTIER Act and serves Anthropic's business interests, and quotes the lead's promised changes on preemption, model blocking, thresholds and updates.
Anthropic’s weak AI bill undermines its ‘safety first’ approach
Transformer's Shakeel Hashim argues that the leaked Anthropic discussion draft of a federal AI bill is weaker than the Obernolte/Trahan FRONTIER Act: evaluators can self-certify, blocking a model requires a court case, and state laws are preempted. Hashim notes Anthropic says it will revise the preemption language.
Leaked Anthropic Draft AI Regulatory Bill Is Unacceptable and Troubling
In a blog post written at a conference, Zvi Mowshowitz argues that a leaked draft federal AI bill, which Politico reports Anthropic is pitching to lawmakers, trades state-law preemption for weak enforcement. The post collects critical reactions from AI-policy figures, and one more favourable view from Charlie Bullock.
AI #190: Scion of the Times
According to the excerpt, Zvi Mowshowitz's weekly roundup says code written at Anthropic keeps rising, run through an internal tool called Scion. It also covers Google's Gemini 4 Argon and Meta's Muse Wonder releases, and a report that OpenAI caught an employee planting a bug.
The Safety Culture at the Labs is Broken
In a weekly roundup, Zvi Mowshowitz argues that safety culture at frontier labs is inadequate. The post covers David Robinson's resignation from OpenAI, lab-employee departures, Anthropic and OpenAI misalignment reports, and calls to slow the frontier, quoting many others.
Arguments About AI Consciousness Remain Highly Motivated and At Best Overconfident
Zvi Mowshowitz writes that AI consciousness has seen little progress in a year and that arguments remain motivated and overconfident. The post covers Anthropic's rule against cruelty to its models, a poll of the author's followers, the New York Times report on Anthropic's meetings with religious thinkers, and a reported Anthropic takedown demand for older Claude models.
OpenAI Fires Three Safety Researchers
Zvi Mowshowitz collects statements from OpenAI and from three fired safety researchers, Tomek Korbak, Mikita Balesni and Jasmine Wang, who dispute the stated reason. Mowshowitz says the firings look bad, believes the researchers sincerely, and notes OpenAI cites an undisclosed further breach of trust.
New Math from OpenAI
According to the excerpt, Zvi Mowshowitz's post 'New Math from OpenAI' describes the development as quite a big deal. The excerpt does not say what the development is.
AI #189: New Math
According to the opening line, Zvi Mowshowitz's post “AI #189: New Math” says the big drop of the week was not a new AI model. The excerpt does not say what it was.
The Curve Bends You
According to the excerpt, Zvi Mowshowitz's post “The Curve Bends You” opens with the line “The plan is no plan”. The excerpt says nothing else about its subject.
Mythos 5.1, Fable 5.1 and Opus 5.5: Model Welfare
According to the opening line, this Zvi Mowshowitz post continues a series on model welfare for new Claude models, here titled Mythos 5.1, Fable 5.1 and Opus 5.5, and points returning readers past its introduction.