# Opinion: An AI company told investors it may not notice its own speed-up. Regulators should read that too

URL: https://ai-pulse.aglaganova.com/blog/opinion-an-ai-company-told-investors-it-may-not-notice-its-own-speed-up-regulators-should-read-that-too
Published: 2026-11-04
Author: Roman Aglaganov, Full-stack developer
Category: Opinion
Region: Global
Tags: safety, regulation

> Anthropic’s amended prospectus warns that it may soon meet its own threshold for AI-driven acceleration, and might only see it afterwards. That belongs in public policy, not just in a risk-factor section.

![Illustration: a grey curve that rises gently then steeply, with a black dashed vertical line crossing it and a red dot just past the line](https://ai-pulse.aglaganova.com/api/media/file/opinion-rsi-disclosure-1.png)

Some of the most important sentences about AI this week were written for lawyers. In its amended IPO filing of 2 November, Anthropic told investors that its estimates for future models “indicate that we may imminently meet” its Responsible Scaling Policy’s definition of “dramatically accelerated progress attributable to AI.” It also wrote that it has “in the past not identified, and may in the future not identify” a significant acceleration in its own development until after it has occurred.

## Why it matters

The fear about AI that improves AI is not science fiction; it is the reason companies like Anthropic have scaling policies at all. If AI systems speed up AI research, capabilities could jump faster than safety testing and outside oversight can keep pace. Here a leading developer is saying, in a document it can be sued over, that it may be close to that point and that its instruments for spotting it lag behind.

Read charitably, this is responsible disclosure. Anthropic says no current model has crossed the threshold, that it is still assessing the data, and that it does not expect to safely let Claude build its successors “without meaningful human supervision or understanding.” Read sceptically, as TechCrunch noted, a warning that your product is almost too powerful also works as marketing. Both can be true.

## What should follow

Either way, a risk disclosed to investors should not be news to the public bodies responsible for AI safety. New York’s RAISE Act, for example, will require large developers to publish safety frameworks, report critical safety incidents within 72 hours and file quarterly assessments of catastrophic risk from January, and California has similar rules. A credible internal signal that a company is approaching AI-driven acceleration ought to trigger the same kind of notice, with the evidence shared with government safety institutes, not only a paragraph in a prospectus.

Anthropic says it will publish further reports under its policy and is considering changes to “keep humans in control of frontier model development.” It should set out what those changes are before the shares start trading, and other labs should be asked the same question. If one company’s numbers point this way, it is unlikely to be alone.

Disclosure: AI Pulse is written with AI tools, including Claude, which is made by Anthropic.

## Sources

- [US SEC: Anthropic, PBC, Form S-1/A (Amendment No. 1), filed 2 November 2026](https://www.sec.gov/Archives/edgar/data/2130357/000162828026068676/anthropicpbc-sx1ax1.htm)
- [Anthropic: Responsible Scaling Policy](https://www.anthropic.com/responsible-scaling-policy)
- [New York State: Governor Hochul announces next steps to regulate major AI developers](https://governor.ny.gov/news/ai-safety-governor-hochul-announces-next-steps-regulate-major-ai-developers-and-protect-new)
- [TechCrunch: Anthropic warns that it’s very close to RSI capability, aka self-building AI](https://techcrunch.com/2026/11/02/anthropic-warns-that-its-very-close-to-rsi-capability-aka-self-building-ai)