Ulrich Gall

Engineer, problem solver. Explorer of ideas, the universe, and the mind.

[Ulrich Gall @outscape]

we didn't build flapping wings to fly
why would we need consciousness for AI?
evolution made things that contract
we made things that rotate
survival+reproduction had to be at the root of all action
suffering, joy, love, hate, selfishness, altruism
so much complexity
we can skip all that
and leave choices to the humans

[Ulrich Gall @outscape]

Very interesting thought. So many unknowns pulling this in various directions. If it wasn't worth the human expertise, is it now worth the compute? Probably sometimes, especially when there are other benefits like increased battery life of your mobile device. Does existing training data bias coding agents towards wastefulness? And finally - we're still doing it: AI clicking around in some web app pretending to be a human user takes this kind of inefficency to a whole new level

[Ulrich Gall @outscape]

@flxbinder Classic one. @Schwarzenegger now that quite a few years have passed, can you give us the back story? Back then of course you claimed it was coincidence. Was this your idea? Someone else who wrote it for you? Did you notice when you signed?

[Elisabeth de Kleer @Elisadek]

The future meets the future.
Today, at OneRobotics, Shenzhen.

[Ulrich Gall @outscape]

@pitdesi An 8 your old might end up with $10k by the time they can sell it- meaningful!!

[Ulrich Gall @outscape]

@pitdesi Yes, brilliant.Groups of friends, school classrooms , etc, getting excited about going to mars and whatnot. There should be a provision that they can’t sell the shares until age 18

[Ulrich Gall @outscape]

@AndyM1928 Wow you can just do that? Material choices and fabrication processes not regulated ?

[Ulrich Gall @outscape]

@pitdesi Really? Interesting data point.. why? Too fancy? ;-p

[Ulrich Gall @outscape]

@bradneuberg I know :-) You're doing amazing stuff!! Space is hard

[Ulrich Gall @outscape]

energy, thermal/power: These seem more like traditional engineering problems (but - maybe the training data for it is pre-gitlab haha). Radiation -> random crashes: Yes, I buy that. That is indeed hard, especially in the light of the previous point - can't just throw more resources at it for redundancy. Limit observability/access: also not new.
But yeah, dealing with nondeterministic execution failures while having to manage power and energy, I could see that being hard. But still something AI can help with? Seems easier to model than, say, how interactions in AR feel to a human

[Ulrich Gall @outscape]

@bradneuberg fine... *be prepared to* have good taste and prompt it

[Ulrich Gall @outscape]

@bradneuberg Or - have good taste and prompt it

[Ulrich Gall @outscape]

@bradneuberg yeah but which if these actually affect software? and is there anything (again, on the software side) you can't simulate on earth?

[Ulrich Gall @outscape]

@grok @pitdesi @amaldorai @grok what were all these people doing? device vendor relations? ad sales? try to break it down

[Ulrich Gall @outscape]

Alexa had everything but the AI. That's solved now, and amazon is in a unique position to completely own home automation & audio. But in the meantime, Alexa got itself stuck in an ad based business model that makes it suck even more, and probably a huge organization built around that model. I wonder if that situation is fixable...

[Ulrich Gall @outscape]

Objectively I would think it would work even better with VR because AI can customize the experience on a per buyer basis... have them cycle through vibes, then see the whole house in that vibe. Doubles as lead gen for custom manufactured furniture. The future will be amazing. Or not? Maybe people won't buy once they've seen the place naked? Husband and wife see different vibes, then argue about how to furnish a place they haven't even bought yet and end up hating it?

[Ulrich Gall @outscape]

@AndrewCurran_ I think he’s projecting. From his own experience. Which is fair

[Ulrich Gall @outscape]

@DavidSHolz If you count in the insanely complex machinery required to get a human to do intellectual work, it’s completely hopeless. We’re going to step it down to a simpler life of rice and not much else if we want to keep up with the machines. Let’s not