The current HN title ("OpenAI researcher posts fake CCTV footage of a real person shoplifting") is outrageously editorialized and should be changed.

The contents of the tweet are:

"i have the most liked video on sora 2 right now, i will be enjoying this short moment while it lasts

cctv footage of sam stealing gpus at target for sora inference"

The video indeed depicts this; "Sam" refers to Sam Altman.

No reasonable person is going to look at this and conclude that it's anything other than an obvious joke. Altman himself does not seem to object to it. The editorialized title implies that an innocent person is being framed, which is just not the case.

Yes, the video's existence highlights potential malicious use cases for AI-generated video, and people will of course have opinions about that, but that doesn't make the title not misleading.

Yes, the title is editorialized and there's nothing inherently wrong with what was shared. But I do think this reflects poorly on OpenAI. They seem proud of how well Sora can create deepfakes, and share no real perspective or concern for the ethical dilemma and impact these models are already having. The marketing around Sora seems incredibly juvenile.

This reminds me of when a US judge believed evidence had to be free from AI manipulation: https://www.theverge.com/2021/11/10/22775580/kyle-rittenhous...

well Sam wouldn't just pick up a box 5 feet from a security guard and not even try to stuff it in his bag, and the doors wouldn't be 20 feet away from a shelf full of expensive stuff without any detectors or anything either. Obviously nobody would ever make a slightly more convincing scenario than this so there's no problem!

Imagine if the store actually set up the GPUs that close to the exit.