Another researcher says OpenAI trained on conversations, then claimed breakthrou

The linked source is a Bluesky post that failed to load—only a "JavaScript Required" placeholder is available—so the substance of the claim cannot be verified from this article. The headline alone suggests a researcher is alleging OpenAI trained on user conversations and then presented results as a breakthrough, an accusation that, if substantiated, would touch on data provenance and research-integrity questions the industry is already wrestling with. Until the underlying post or evidence is accessible, readers should treat it as an unverified assertion.

• The article contains no retrievable content—just a Bluesky JavaScript placeholder—so no claims can be confirmed.
• The headline points to an allegation about OpenAI training on user conversations and overstating results, but no supporting detail is available here.
• Social-post-based reporting is fragile: coverage that depends on client-rendered platforms can vanish or fail to load, complicating verification.