مولد الصوت AI Elon Musk بواسطة Fish Audio

It's basically mind-blowing how much nonsense accumulates in these large organizations. You have to look at the physics of the problem and just delete the parts that aren't working. If a commercial company operated this way, they’d be bankrupt in a week. We really need to get back to first principles.

When you look at it from first principles, most of the complexity is actually unnecessary. We need to delete parts and simplify the process. It's insane how much waste exists in these legacy systems. If we don't fix the fundamental architecture, we're simply not going to reach the next level.

It's really quite remarkable when you look at the first principles of how these legacy systems operate. I mean, the level of inefficiency is just, it's mind-blowing. If any commercial company tried this, they would be bankrupt immediately. We have to fundamentally rethink the entire structure for the future.

Look, when we're developing new technology, we have to sort of think from first principles, basically. It's not just about making things work, it's about understanding the fundamental physics, the core engineering challenges, and then iterating based on feedback from the community.

Look, the fundamental problem is, you know, we need to make life multiplanetary because, like, Earth has a finite time window. Mars is challenging, obviously, but we've got to solve these engineering problems and make humans an interplanetary species. That's the only logical path forward.

So like we're working on this crazy AI thing, and you know, most people think it's impossible, but that's what they said about landing rockets. We've got maybe a 50% chance of success, but if it works, it's going to be absolutely mind-blowing. We're basically trying to solve intelligence.

The sheer scale of the waste is just incredible, really. If you look at it from first principles, it is like a Rube Goldberg machine of bureaucracy. We need to streamline everything, cut the nonsense, and focus on actual output. Otherwise, the system eventually collapses under its own weight.

So, if we’re talking about AGI, it’s really about the scaling laws. It's not just about more compute, but higher quality data. Most of the information we have is trapped in video, right? We need systems that can actually understand the physical world, not just predict the next word.