RMRob Miles@robertskmiles10:29 AM · Jul 23, 2026

I think it's like "drug that company plans to administer to everyone whether they want it or not has deadly side effects". Awkward metaphor. Point is if boeing makes bad planes, you can choose not to fly in them. Perhaps if the plan was to somehow load everyone onto one giant plane? Or to make hundreds of millions of planes and fly them over every populated area. Maybe you think Boeing can't achieve that, but they think they can, and clearly they should not be allowed to attempt it

BLblueblimp@blueblimpms10:33 AM · Jul 23, 2026

@binarybits I think the incident should be taken seriously in the sense of determining how it happened and putting in layered mitigations to prevent recurrence. But it also shows how low-risk AI is that, unlike the other examples you mentioned, nobody died or even suffered injury.

BLblueblimp@blueblimpms10:35 AM · Jul 23, 2026

@binarybits It seems like a point in favor of the view that "AI is powerful, but it's okay to address risks as they arise (rather than proving it risk-free ahead of time)". Of course, an important part of that is to address the risks when they do show up, which they now have.

TBTimothy B. Lee@binarybits10:23 AM · Jul 23, 2026

1868TECH

To be clear the OpenAI hack didn't kill anyone so in that sense it's not nearly as serious as an airplane crashing. I just mean something can be a major screw-up for a particular company without having civilizational implications.