What's cool (and tough to keep up with) with this wave of tech is just how quickly it moves.
On the plus side there's a lot of interesting things and it is generally easy to follow/figure out what they did.
On the minus side it's a little exhausting, and there's so much money in it feels like the vast majority of it is grifting. To add to that, the people who are trying to catalog it (the AI guy on LI) are the griftiest of them all.
I've found the best way to keep up is find one topic you want to learn about, deep-dive, and read all the related papers, then explore breadth-first from there until you find another topic...
How do you keep your knowledge after deep dive? Do you try to use it somehow? I found that reading a lot usually does not contribute to long term proficiency in a given topic, unless followed by non trivial amount of practice.
Yeah so I’m lucky that I work in/adjacent to the space so it doesn’t get buried. Otherwise I think it’s near impossible to retain learning without practice.
It's still in the early phase where the bubble is building up. This is necessary if we want a prosperous market. Hopefully, after the bubble bursts, some good companies will remain (which is very likely).
> On the minus side it's a little exhausting, and there's so much money in it feels like the vast majority of it is grifting.
It is all grifting. The moment someone creates something that can improve upon itself there will be an intelligence explosion and it won’t need press releases or debates about its intelligence. The current path of research will not lead to that, if there was something to be discovered here it would have been discovered already. It’s just new to the general consumer and there’s a wow factor associated with it like crypto and NFTs before. The truth is tech has lost its momentum and is desperate to find a new trick.
The rate of improving is important.
If it's as intelligent as the average human, the rate of improving will be slow, very slow compared to what the researcher can do currently.
I think the opposite. There is value in intelligent software, but IMO we’re a long way from AGI. So lots of grifting but some gold along the way. And it’s intellectually interesting/nuanced (cool math, interesting infra), unlike crypto which was more of a massive energy burning waste than anyone likes to admit.
If our standard is cool math, crypto is also full of cool math. (Have you read Vitalik's explanation of Quadratic Arithmetic Programs?) Our standard can't be that low.
Fair enough. More reacting to the concept of a blockchain which is like old news and an extremely inefficient way to do 90% of what people are (we’re) trying to do with it.
On the plus side there's a lot of interesting things and it is generally easy to follow/figure out what they did.
On the minus side it's a little exhausting, and there's so much money in it feels like the vast majority of it is grifting. To add to that, the people who are trying to catalog it (the AI guy on LI) are the griftiest of them all.
I've found the best way to keep up is find one topic you want to learn about, deep-dive, and read all the related papers, then explore breadth-first from there until you find another topic...