Speculation: OpenAI and Anthropic may use 'safety' to mask slowed model progress before IPOs

StewartalsopIII · x · 2026-08-23

The author speculates that the narrative pushed by OpenAI and Anthropic—having better internal models too dangerous to release with US government approval—might be pure marketing. This could be a tactic to hide a slowdown in model progress ahead of IPOs. The author urges skepticism, drawing parallels to the 'COVID delusion' and noting that few attention-holding groups act in good faith.

Original post →

More from Companies & People

Companies & People channel →