I read the YouTube transcript. This old, bald guy dribbling on is important to the future of my family how?
Sorry wasnt expecting a family crowd here fixed*
But seriously dem kids be tiktoking the yolo ai's get with the program or be left in the dust
despite my lack of activity, and I've noticed a recent uptick in posts discussing AI hype and how it's surely a bubble ready to pop. This is obviously due in large part to the recent NVDA explosion, but there's more to consider.While you guys spend most of your time focused on the money and investments side of AI, I spend my time focused on the technology itself - and the beliefs that seem to emanate from it.Let's look at NVDA. Right now, NVDA is making its mark on the world almost entirely due to its ability to produce and supply cutting edge GPUs for companies looking to fashion larger and larger clusters and have more compute than any reasonable person could possibly fathom. Surely, this ends at some point, right? Surely, the people in charge realize that they've built large enough GPU clusters (never mind the hundreds of billions recently committed to building the largest datacenters on the planet by the likes of Google and Microsoft) and Jensen stops getting orders. Surely, this ends.A lot of people have been comparing AI hype to the dot com bubble of the early 2000s, which seems to be a fair comparison on a surface level. However, this misses something so fundamentally different about AI today that it almost entirely invalidates it.Whether you think AI is real or a scam or a stochastic parrot or whatever, the people in charge (Sam Altman, Elon Musk, Bill Gates, Sundar Pichai, Tim Cook, etc) believe 100% that this is one of the final pieces of technology mankind will ever create. They believe that the first company that creates AGI (artificial general intelligence - typically defined as an AI that can do any task as well as the average person) will be the first company to create ASI (artificial superintelligence - typically defined as an AI that is smarter than every single living human being put together) and will end up deciding the future. And they believe this is going to happen in ten to fifteen years.If this seems silly to you, I don't blame you. Technology typically doesn't move especially fast or change too much over short amounts of time. It's not like a refrigerator purchased today is a wildly different beast than one purchased fifty years ago. Sure, it might be more efficient and have more room, but fundamentally it's the same technology.This is how life has been ever since the industrial revolution: a new technology is invented, distributed, and then refined over the next 50 years until it basically asymptotes.That is not what is happening with AI, and it isn't what is going to happen in the future. Progress is being made at a blistering rate. People who aren't constantly paying attention to every single development and published paper have no idea what's going on. They just wake up one day and go "Oh wow look at that, AI can just make videos and music out of nowhere. This must be something that was developed overnight and is a magical thing that can only happen once or twice in my life. Any discussion about further development is silly and just hype!"Whether you like it, believe it, or even choose to acknowledge it, AI isn't in a hype bubble this time. It's easy enough to look back at the last half-century and see AI's previous bubble bursts, but those were due in large part to the technology itself being substandard. It's hard to make ASI when your supercomputer has less computing power than the first iPhone in 2007. This is why today's companies are pouring what seems like an unreasonable and unjustifiable amount of money into buying GPUs - and why they will continue to. Everyone has realized that compute is all you need.Whether or not this ends up making "god in a box" isn't especially important. What is important is that the companies trying to do it aren't going to stop anytime soon. The risk/reward on trying to build the final thing mankind will ever build is so astronomically skewed that it's hardly even a consideration. The reward for success is so vast that it makes the cost of failure (a few hundred billion dollars) seem insignificant.I'm not trying to sell you on NVDA or MSFT or anything like that, I'm just trying to let you know what's actually going on. You can't sit there and say "I looked on the internet and saw a tech bubble from before I was born, that must mean this is the same thing!" because that makes you wrong. You fundamentally misunderstand what's happening in tech right now if you think this is going to follow conventional patterns. This is a once in a lifetime event, and it might very well succeed.
