285 karma · joined January 19, 2022
YouTube channel that published a video series analyzing the Surfside condo collapse now goes over what is claimed to be a leaked transcript of text communications with the doomed OceanGate submarine mission.
Submarine descends too fast, ~38m/min instead of ~25m/min. Support vessel notes discrepancy, sub presses on with descent.
At 9:28a, hull monitoring alarm sounds. Upward thrust is commanded to no effect. Ballast released to no effect. Landing frame released after several attempts, but ascent rate 1/4 of planned.
At 9:38a, crackling sounds heard in aft of ship.
At 9:46:37, power bus A reports failure. Crew switches to power bus B. More sounds heard. This is the last communication received by the support vessel.
-------
If this is authentic, then it basically confirms gradual delamination that snowballs into catastrophic implosion. It's horrible that the crew knew things were going wrong for nearly nineteen minutes and could not do much about it.
And how would you quantify alignedness? Rating outputs for a given input falls prey to the first problem. Analyzing activations as they trickle through the model is intractable analytically, and training a "polygraph" model on the activations of your network raises more alignment issues (how can you be sure the polygraph isn't lying to you?)
I'm ready to eat my words, but I think perfect alignment is infeasible. The best we can hope to do is curate training data and hope the caged bird won't sing.
(As a cool aside, some fuel-burning power plants are already crazy efficient. For example, my university has a trigeneration plant that, in addition to electricity, provides steam and chilled water to the campus by extracting the residual heat that would normally be sent straight to the atmosphere. I believe the plant uses an ammonia absorption cycle to produce cold water from a heat source, which is pretty awesome in its own right.)
Now, I'm not sure if this process happens fast enough for an EV to be more efficient over the projected lifetime of the battery. I think it ultimately comes down to your local regulatory environment, existing distribution of energy sources, and public willingness to invest in more efficient infrastructure.
Yeesh, this says a lot about their target market.
Cool tech, though. Let's hope I'm not peer pressured into wearing one of these masks for my kid's goddamn birthday party.
I guess the only real bummer is the high price -- all those optical components can't be cheap at this scale -- and the short battery life which really can't be solved without either a serious breakthrough in power efficiency (which apple is on the frontier of), battery density (likely same) or sacrificing visual fidelity (not in Apple's ethos). I don't mind the external battery pack, actually; hopefully it results in a well balanced headset.
I wonder if they're going to do in-person demos at the Apple store. I might have to go check it out.
Apple really are the kings of making incredibly advanced technology seem, for lack of a better word, magical. They showed the headset has a really fancy two axis motorized system for lining up the lenses with the user's eyes -- something literally no other headset has tackled -- but didn't elaborate on it. The assumption is it's just what's required to deliver a good experience, not something to be used as a selling point in and of itself.