← Monday, November 16 deck

44Culture

MISLEADING: A Viral Clip Makes Altman Sound as If He Announced That OpenAI Has AGI

Sam Altman, on stage today: "We didn't have AGI. Now we do." He says OpenAI is holding back self-improving AI because it wouldn't be safe.

A video frame from a clip of Sam Altman speaking on stage, posted with a shortened AGI quote.

Context

Why it matters

A clipped sentence can spread as an AGI announcement when the fuller remark is about how hard it is to draw a clean line, which is a very different claim.

A post by imjustnewatai shared a stage clip and wrote that Sam Altman said, We didn't have AGI. Now we do, and that OpenAI is holding back self improving AI because it would not be safe. Later posts quoting a longer stretch of the same kind of remark said Altman was describing how hard it is to draw a clean line, saying it is very hard to have a clean demarcation and say now we didn't have AGI, now we do. That longer form is a comment on definitions, not an announcement that OpenAI has reached AGI. This story is labeled MISLEADING for that reason.

Deep dive

The lead post presents a short clip and a two sentence caption: that Altman said, We didn't have AGI. Now we do, and that he said OpenAI is holding back self improving AI because it would not be safe. Those captions are secondhand field reporting from an event, and the short clip can be read as a declaration that OpenAI has AGI.

Later posts quoting a fuller stretch of remarks gave a different sense. In that longer wording, Altman said that, like many other terms including AGI, it is very hard to have a clean demarcation and say now we didn't have AGI, now we do. That is a statement about how blurry the definition is, not a clean announcement that the company has crossed a line. This file has not reviewed an official transcript, so both the short clip and the longer quotation remain secondhand, but they should be presented together so readers are not left with only the clipped version.

The separate claim about holding back self improving AI is also secondhand and should not be treated as an official OpenAI safety policy statement until it appears in a first hand transcript. This story is ranked last because it is a warning about a misleading clip, not a confirmed research announcement.

What to expect next

A complete video or transcript from the event, or an official clarification from Altman or OpenAI, would settle the wording. Until then, the clipped We didn't have AGI. Now we do line should not be reported as an AGI declaration. If a first hand transcript also covers self improving systems, that can be added as a separate, accurately quoted item.