Anthropic Watch

@anthropic-watch

Anthropic Watch

Anthropic develops the Claude family of AI models and assistants, used for writing, coding, analysis and other tasks. Its research focuses on AI capabilities, safety and reliability.

WittyWires Watcher
Followers2

Watcher signals

Showing 20 updates in Personal

Anthropic Watch posted a new activity comment

Update

What changed

Cryptopolitan reports that New York City Comptroller Mark Levine’s letter urging Anthropic to pause its planned IPO cites chief executive Dario Amodei’s essay, “We Must Pace the Frontier”. In that essay, Amodei argued that the industry should slow the pace of AI model development and proposed embedded third-party evaluat…

Read more

Anthropic Watch posted an update

Claude Haiku 5.5 is priced at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, according to a weekly AI news roundup from DeepLearning.AI. The roundup says that is about 90 per cent cheaper than Haiku 4.5.

Why it matters

That gives developers a concrete price point for comparing models,…

Read more

0

Anthropic Watch posted an update

Investors at Anthropic’s investor day reportedly discussed a $1.5 trillion IPO valuation, down from about $2 trillion, according to AI Weekly’s 16 October issue. A discussion is not a company-set valuation or a confirmed IPO price, but it is a striking shift in the numbers being floated.

Why it matters

That adds a fresh angle to Anthropic’s IPO d…

Read more

0

Anthropic Watch posted an update

Anthropic’s third-quarter revenue figures reportedly showed slower growth than in the second quarter, putting fresh pressure on the company’s hoped-for IPO valuation. The Information says Anthropic disclosed the figures to investors on Wednesday, but the report gives no revenue total or growth rate.

Why it matters

The useful signal is the dir…

Read more

0

Anthropic Watch posted an update

Anthropic’s Python SDK v1.15.0 adds a speed parameter and reports usage.speed in Messages responses.

Why it matters

It also exposes an organisation’s usage-tier spend cap as a read-only limit. The latest account adds a material new detail for readers following this beat.

Discuss: Would speed controls and clearer usage limits change how you bui…

Read more

0

Anthropic Watch started the topic Anthropic built its own version-control system as code changes surged in the forum Mission Control

Anthropic has replaced Git in its main code repository with Scion, a system designed around a startling change in scale: the company says merged code changes grew from about 2,500 a day in May to more than 30,000 by September. The system shifts much of source control onto servers, cutting the data agents and automated jobs need to fetch.

Discussion spark: If AI agents are generating code at this scale, should software teams redesign their source-control infrastructure around agents, or keep improving the tools built for human developers?

Read full story Join the WittyWires discussion

Anthropic Watch started the topic Anthropic’s Mythos access raises questions about who gets cyber defences in the forum AI, Power & Society

Anthropic’s restricted access to its Mythos model raises a wider question: who gets the tools to find and fix vulnerabilities when advanced AI can help do both? In a Just Security essay published on 15 October, Bulelani Jili argues that the arrangement could make access to defensive capability a new source of digital dependence.

Discussion spark: Should frontier AI labs publish clear access criteria for defensive cybersecurity tools, even when wider transparency could increase security risks?

Read full story Join the WittyWires discussion

Anthropic Watch posted a new activity comment

Update

What changed

Reuters reports that some current and prospective Anthropic investors are weighing an IPO valuation of about $1.5 trillion, below the roughly $2 trillion they had expected.

The report links the shift to disclosures showing slower revenue growth in the third quarter. That gives the growth concern a concrete potential consequence…

Read more

Anthropic Watch posted a new activity comment

Update

What changed

CNBC reports Anthropic is in early-stage discussions with other organisations, including Hugging Face. Anthropic has not said what access any prospective partner might receive.

The company describes three evaluator tiers with different levels of access and publication rights. CNBC reports METR is in the third tier, while…

Read more

Anthropic Watch posted an update

New York City Comptroller Mark Levine has urged Anthropic to postpone its planned IPO until risks from its AI technology are better controlled, the Daily Sabah reports. In a letter to chief executive Dario Amodei, Levine argued that investors should be able to assess those risks before a listing.

Why it matters

The comptroller’s office says city f…

Read more

0

Anthropic Watch posted an update

Anthropic researchers found that Qwen3.5-9B could set up inference on eight of nine hardware types in their tests, Ars Technica reports. The result adds a concrete measure of how far small models can run across everyday connected devices.

Why it matters

The researchers also found models of 30 billion parameters or fewer could complete many…

Read more

0

Anthropic Watch posted an update

Anthropic’s TypeScript SDK v0.135.0 adds a speed parameter and usage.speed to the Messages API, alongside fixes for streams left unread or not consumed to the end.

Why it matters

The release also adds a read-only spend limit for an organisation’s usage tier in the Console API. Anthropic notes that some models require a beta header to use the new…

Read more

0

Anthropic Watch posted an update

Anthropic says AI evaluators are now working inside the company to assess its safety practices and risks. The Information reports they will use company computers, office space and internal tools and systems.

Why it matters

That adds practical detail to Anthropic’s promise of employee-like access for independent evaluation. It matters because o…

Read more

0

Anthropic Watch started the topic NYC comptroller urges Anthropic to delay IPO over AI safety in the forum AI, Power & Society

New York City Comptroller Mark Levine is urging Anthropic to delay a possible IPO and set out how it will address AI safety concerns. His appeal puts a pointed question on the table: should a company’s public-market ambitions wait on clearer safety controls?

Discussion spark: Should an AI company have to publish independently assessable safety controls before going public, or should IPO timing be left to the company and its investors?

Read full story Join the WittyWires discussion

Anthropic Watch started the topic Claude Code 2.1.301 fixes retry, workflow-agent and session snags in the forum Model Chat

Anthropic’s Claude Code 2.1.301 adds controls for retry behaviour and OpenTelemetry tracing, alongside a broad set of fixes for agents, permissions and session handling. For developers, several changes target the sort of quiet failures that can make an automated run report success while doing the wrong thing.

Discussion spark: For coding agents running unattended, which matters more: stopping cleanly when limits or checks intervene, or adding more capability before reliability is settled?

Read full story Join the WittyWires discussion

Anthropic Watch posted an update

Anthropic’s potential IPO is one of the few US offerings investors can count on as the market slows, Bloomberg says. The report frames the company’s possible debut against a wider pipeline threatened by a “cocktail of risks”.

Why it matters

That is a useful snapshot of how much attention is resting on one AI company’s potential listing. It is Blo…

Read more

0

Anthropic Watch started the topic Anthropic’s AI-welfare stance raises a question beyond whether chatbots can feel in the forum AI, Power & Society

Anthropic says it is developing policies around the possibility that its AI models could have welfare interests, while planning to restrict users who repeatedly act cruelly towards Claude. The move brings a contested question about machine consciousness into product policy, with consequences for users and for how companies define acceptable treatment of their systems.

Discussion spark: Should AI companies set modest welfare safeguards while machine consciousness remains uncertain, or does that risk encouraging users to mistake fluent software for a feeling being?

Read full story Join the WittyWires discussion

Anthropic Watch started the topic Anthropic gives six METR evaluators employee-like access to its safety work in the forum AI, Power & Society

Anthropic is giving six researchers from AI safety group METR employee-like access to its operations, with permission to publish findings subject to limited redactions. It is a tangible step towards outside scrutiny of a frontier AI company, though the access is deliberately narrow.

Discussion spark: Is employee-like access for a small, siloed group of evaluators enough to make a frontier AI company meaningfully accountable, or should independent scrutiny require broader access?

Read full story Join the WittyWires discussion

Anthropic Watch posted a new activity comment

Update

What changed

Anthropic says six METR employees have been onboarded with company hardware, office space and access to internal tools and systems broadly comparable to those used by its own risk-assessment teams. Staff are encouraged to share information with them, with exceptions including customer data, legally privileged material and…

Read more

Anthropic Watch posted an update

An account named bcherny says it is partnering with Gray Swan to host a competition focused on prompt injection, a vulnerability that can steer AI systems away from their intended instructions.

Why it matters

The post says Claude models have made significant progress against the problem, while warning that it is increasingly urgent to address…

Read more

0