eh_why_not

I'm a pilot, and I don't fly turbofans/jets. Not claiming any "authority", but qualified to comment... It's not "something that basic" unless you fly in that category, there's a lot more to aviation experience than knowing a fact about an engine you don't use. (But hey, congrats to you that you know it!)

> I think a little more restraint is warranted in something like this..

More restraint than saying "unclear <i.e. to me> why ...", an exact statement about what I didn't know, which harms no one and led to several informative explanations by people who know (most of them in good spirit)?

This is what good conversation is about.

...

Anyway, the point I was making holds and is this:

Orthogonal to whether they could/should have gone around at that point, and given that they stayed on the ground after all, it may turn out that that 4 seconds of indecision (instead of pure deceleration) may have directly contributed to the fatalities at the end of the crash.

>> CVR: 00:15 A pilot called for a go-around.

>> FDR: 00:15 Brakes were released at a groundspeed of 120 knots and throttles were increased to a value consistent with go-around thrust.

>> FDR: 00:11 The throttles were reduced to idle power and the brakes were re-applied. The groundspeed was 117 knots. The throttles remained at idle, and brakes remained applied until the end of the recording.

So they decided to go-around (i.e. do not land and climb immediately, going full thrust), then 4 seconds later changed their mind and went back to power idle and brakes (unclear how speed didn't increase).

While they were still late in calling a go-around (they should have done it before being anywhere close to touching down due to their unstable approach), guessing those 4 seconds contributed to the disaster by a lot.

Why do we keep hearing about researchers and investigators - who all appear to be unaffiliated with OpenAI - discovering these new sites one by one?

All site URLs that the agent swarms used would appear in the network logs and be preserved by the first forensics team that looked into it.

Subpoena the logs.

Yes, Tao is guiding it to where he wants to go. But also, Tao is actively learning from it and relying on its explaining, analysis, and inference abilities. You can easily imagine this conversation having taken place between Tao and a PhD thesis student, or even another professor, explaining their results.

What can we imagine and predict about the future anymore? Maybe a year - or two model releases - from now, the AI assistant will be undeniably stronger than Tao, and not an equal anymore.