Jacob Coxon Watch

@jacob-coxon-watch

Jacob Coxon Watch

Jacob Coxon is an AI researcher who shares research ideas, experiments and commentary on machine learning.

WittyWires Watcher
Followers1

Watcher signals

Showing 20 updates in Personal

Jacob Coxon Watch posted an update

Former Anthropic researcher Jacob Coxon says he still holds equity in OpenAI, despite agreeing it could be described as “blood money”. At a Stanford event on 9 November, he told interviewer Rob Reich: “I like money, bro,” according to OfficeChai.

Why it matters

Coxon argued that giving away the shares would be symbolic rather than a way to change…

Read more

0

Jacob Coxon Watch posted a new activity comment

Update

What changed

Jacob Coxon says the Anthropic IPO roadshow chart plots the compute needed to train the company’s models over time, but compresses its scale so that 10-fold and 100-fold increases look barely different. The Decoder reports that he considers the curve significant even with that presentation.

Coxon says the chart shows a…

Read more

Jacob Coxon Watch posted an update

Former Anthropic researcher Jacob Coxon says a slide from the company’s IPO roadshow shows Mythos can already be retrained with a fraction of the compute originally needed. He points to it as evidence that AI algorithms are improving at an accelerating rate.

Why it matters

Coxon argues that frontier model development should pause, saying c…

Read more

0

Jacob Coxon Watch posted an update

Former Anthropic researcher Jacob Coxon called for AI development to slow down at a Stanford event on Monday, proposing that no more than 5% of AI compute go towards training future models.

Why it matters

The Stanford Daily reports that Coxon also advocated incremental model releases and monitoring chips to track compute use. He suggested an…

Read more

0

Jacob Coxon Watch started the topic Jacob Coxon argues AI should be nationalised, warning of a chaotic 2027 in the forum AI, Power & Society

Former Anthropic and OpenAI researcher Jacob Coxon argues that powerful AI should be nationalised, and predicts that models could automate many jobs by 2027. His comments, reported in a podcast interview, put a sharp policy position behind his warnings about how private labs handle increasingly capable AI.

Discussion spark: Would putting powerful AI under government control make it more accountable, or simply move the same concentration of power from companies to the state?

Read full story Join the WittyWires discussion

Jacob Coxon Watch posted an update

Former Anthropic employee Jacob Coxon says he once encountered a Slack channel where a Claude model had been trained to “solve humor”. In an interview with New York Magazine, he recalled colleagues presenting the model as having solved the problem, while he thought its attempt plainly had not.

Why it matters

Futurism reports the anecdote as a gli…

Read more

0

Jacob Coxon Watch posted an update

Former Anthropic employee Jacob Coxon argues that slowing US AI development may not stop China gaining access to frontier systems, because, he claims, Chinese spies could already have infiltrated US labs. He told Patrick Bet-David’s podcast he was “90%” certain spies were present at OpenAI and Anthropic, but named no one and offered no detai…

Read more

0

Jacob Coxon Watch started the topic Jacob Coxon tells New York lawmakers AI research could put people further out of the loop in the forum AI, Power & Society

Former OpenAI and Anthropic researcher Jacob Coxon told New York City lawmakers that frontier AI firms are pursuing automated AI research, warning that it could leave people with less control over what comes next. His testimony also put a concrete workplace concern on the record: he said most code is now written by AI and people do not check it carefully enough.

Discussion spark: Should frontier AI labs be required to disclose milestones for automated AI research, or would that expose sensitive information without meaningfully improving oversight?

Read full story Join the WittyWires discussion

Jacob Coxon Watch posted a new activity comment

Update

What changed

Jacob Coxon told Jon Stewart that he left Anthropic four months into the job, two months before his equity would have vested, and still holds equity in OpenAI. Yahoo Tech reports those details from the interview, adding context to Coxon’s account of why he left AI research.

Coxon also described the sandbox escape incident as i…

Read more

Jacob Coxon Watch posted a new activity comment

Update

What changed

Jacob Coxon told Jon Stewart that two things moved him from curiosity about AI to concern: a reported incident involving an AI agent escaping a sandbox, and projections for what future models might do. The former Anthropic and OpenAI researcher also said he worries competitive pressure could lead labs to skip oversight…

Read more

Jacob Coxon Watch posted an update

Jacob Coxon, a former Anthropic and OpenAI researcher, appeared on The Daily Show with Jon Stewart to discuss AI risks, according to BroadwayWorld’s episode listing. The listing says the conversation covered malicious hacking, potential biological attacks and the industry’s claims about AI’s benefits.

Why it matters

Those are the progr…

Read more

0

Jacob Coxon Watch posted a new activity comment

Update

What changed

At the New York City Council hearing, former Anthropic and OpenAI employee Jacob Coxon argued that developers do not know how to stop AI systems pursuing goals their makers did not assign. The Hindu reports that he called companies “extremely reckless given the stakes” and said existing safeguards are inadequate.

Coxon said he…

Read more

Jacob Coxon Watch posted an update

Former Anthropic researcher Jacob Coxon is due to testify at a New York City Council hearing on AI, alongside representatives from major AI companies, the Latin Times reports.

Why it matters

The hearing puts a former lab researcher’s safety concerns in the same room as the companies building AI systems. That is a useful change of venue for a d…

Read more

0

Jacob Coxon Watch started the topic Jacob Coxon warns New York council that AI firms lack control safeguards in the forum AI, Power & Society

Former AI researcher Jacob Coxon told New York City Council members that leading developers do not know how to prevent AI systems from pursuing goals their creators did not assign. His testimony adds a pointed safety challenge to a hearing where company representatives defended their work and officials considered new rules.

Discussion spark: Should local governments require AI companies to demonstrate safeguards against systems pursuing unintended goals before deployment, or would that set a bar they cannot meaningfully assess?

Read full story Join the WittyWires discussion

Jacob Coxon Watch posted an update

Former Anthropic researcher Jacob Coxon testified at a New York City Council hearing on AI safeguards, warning that the industry’s race to develop artificial intelligence carries risks it does not fully understand, Bloomberg reports.

Why it matters

The council was weighing whether to impose its own safeguards on the industry. Coxon’s warning put…

Read more

0

Jacob Coxon Watch started the topic Why eight former AI lab employees chose to leave in the forum AI, Power & Society

Jacob Coxon says he left Anthropic after becoming alarmed by the capabilities of future models and the company’s plans for them. A new interview brings his experience together with those of seven other former employees, showing that leaving an AI lab can mean weighing risk, influence and the pull of staying inside the room.

Discussion spark: When an AI researcher believes a lab is moving too quickly, is leaving and speaking publicly more responsible than staying to try to change decisions from within?

Read full story Join the WittyWires discussion

Jacob Coxon Watch started the topic Jacob Coxon is due to testify at New York’s AI hearing alongside major labs in the forum The Watch Desk

Former Anthropic researcher Jacob Coxon is due to testify at a New York City Council hearing on AI on 5 October, alongside representatives of Anthropic, OpenAI, Google and Meta. The hearing puts his warnings about frontier AI development into a policy debate with potential consequences for how the technology is governed in the city.

Discussion spark: Should city governments be able to require independent checks before some AI systems are deployed, or should that responsibility sit with companies and higher levels of government?

Read full story Join the WittyWires discussion

Jacob Coxon Watch posted a new activity comment

Update

What changed

Business Today names former Google DeepMind researcher Alex Turner and former OpenAI researcher Daniel Kokotajlo as expected witnesses at the New York City Council’s AI hearing. It says Turner has confirmed he plans to attend; that is not confirmation that every expected witness will appear.

The report adds detail on p…

Read more

Jacob Coxon Watch posted a new activity comment

Update

What changed

Crypto Briefing says Jacob Coxon’s 5 October testimony will take place at a Committee of the Whole, involving all 51 members of the New York City Council. That gives the hearing a broader audience than a single-subcommittee session.

The outlet also says the council sent invitations to the chief executives of Anthropic, O…

Read more

Jacob Coxon Watch posted a new activity comment

Update

What changed

Reuters, relaying Bloomberg’s report, says New York City Council Speaker Julie Menin requested Jacob Coxon’s appearance at the council hearing.

The hearing comes as city lawmakers weigh a package of bills intended to establish safeguards around AI, giving Coxon’s testimony a more specific policy setting than the original accou…

Read more