No Primary Source Confirms GPT-6.1 Astra Cancellation or FTC Probe Claims

A post on X dated October 4 claims OpenAI cancelled the rollout of a product called GPT-6.1 Astra and that the Federal Trade Commission opened a probe of both OpenAI and Anthropic. No FTC filing, press release, or docket number accompanies the claim. No OpenAI statement confirms a cancellation. The same thread cites a separate outlet, AI Breaking Wire, for four more claims: an OpenAI training pause, new Meta Muse Gadgets firmware, a figure that 60% of Airbnb's code is AI-authored, and a ChatGPT try-on feature. None of these carry a primary source either.

For workers inside these companies, the gap between claim and confirmation is the whole story. If a training pause or product cancellation were real and material, it would likely surface in a regulatory filing, an internal memo that leaks, or a status page update, not a single unlinked post. Managers deciding whether to brief teams on an FTC probe need something firmer than a screenshot. Policy watchers tracking antitrust activity against AI labs have a specific, checkable fact pattern to look for: a Federal Register notice, a formal FTC press statement, or a civil investigative demand referenced in a company's own disclosures.

The Airbnb figure is the kind of claim that sounds plausible precisely because it is unverifiable without the company's own disclosure. A 60% AI-authored code figure would normally come from an engineering blog post, a shareholder letter, or an earnings call transcript, none of which are cited here. Firmware claims about Meta Muse Gadgets face the same problem: firmware updates are typically confirmed through device changelogs or support pages, not social posts. A ChatGPT try-on feature, if real, would show up in OpenAI's own release notes or app store listings well before a secondhand aggregator names it.

The mechanism worth watching is attribution chains. AI Breaking Wire is described as the source for four of the five claims in the thread, with the original poster adding the Astra and FTC items separately. That makes two layers of unverified claims stacked on each other, with no primary document linking back to either OpenAI, Anthropic, Airbnb, Meta, or the FTC. When a claim cannot be traced to a filing or an official statement, the appropriate status is unconfirmed, not false. But unconfirmed claims about antitrust action carry real weight if repeated inside a company or cited in a policy brief without that caveat attached.

Anyone deciding what to do with this information has a narrow set of next steps. Check the FTC's public enforcement actions page for any new matter naming OpenAI or Anthropic. Check OpenAI's and Anthropic's own newsrooms for statements on a training pause or product cancellation. Check Airbnb's investor relations materials for any disclosure of AI-generated code share. None of those checks have been done as of October 14. Until one of them returns a primary document, the five claims in the October 4 thread remain exactly that: claims, routed through an aggregator, with no filing or statement behind them yet.