‘Planes are going faster and faster and higher and higher. And a decade ago they were impossible. What if we accidentally leave the earth’s atmosphere?’pic.twitter.com/wR4akCq0CN
-
-
Prikaži ovu nit
-
(Also, please read the second sentence. Yes, we went into orbit. Not, not with anything connected to those aircraft. The existence of aircraft had no predictive value for spaceflight.)
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
Disagree to the central point. Planes engines power has been growing sub linearly in the last 100 years. Computational power is on an exponential growth path. Yes, we should worry, AI will go on orbit at one point.
-
AGI is not just a function of computation power. We do not have any theoretical model of what it is and we have no idea how to build it.
- Još 3 druga odgovora
Novi razgovor -
-
-
No, it’s like looking at the V2 and the Manhattan project and worrying someone’s going to invent ICBM’s.
-
The V2 WAS an ballistic missile. Very very few people working on ML think they have any roadmap at all to get to AGI. It's not just the same thing but with more range/power/computer. So, no.
- Još 2 druga odgovora
Novi razgovor -
-
-
Take your analogy the other way. Sure, the tech was mostly unrelated (not completely e.g. fluid dyn. and thrust) but if there was no airplane first, would engineers focus on space? Once the easy problem is solved the focus becomes the harder one. It’s a matter of collective will.
-
AGI is not a matter of 'collective will'
- Još 5 drugih odgovora
Novi razgovor -
-
-
You're metaphor is getting slammed. The part of this tweet that I focused on is whether worrying about something far away but inevitable is justified. The expected value of not worrying about it seems too negative given the cost of getting it wrong when our time comes.
-
A few people didn’t read it properly. Yes, we did go into orbit, but not with anything connected to those aircraft. The existence of the aircraft had no predictive value for thinking about space flight. Equally, it is not self-evident that ML tells us anything about AGI
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.