> It also appears to be true of organizations too. How many startups were just handcrafted answers to a specific problems? The big labs are a general solution with loads of compute.
If he continues this worship of the "bitter lesson" to its logical conclusion, why stop at startups? All economic, or even human, activity could be seen as "handcrafted solutions".
So is the idea that the big labs should take over everything in the end? Why is that a good idea? Why would you want to live in such a world?
I think some people have come to that conclusion. Many have joined the big labs.
Whether they’re right or not, we might not get to choose. On some level we’re all just witnesses to this moment in history. It will be whatever it is and if the models keep improving that’s exactly what will happen.
If the s curve flattens out. Things might be different
Well, I guess, better join the people protesting the data centers then. This stuff does not (yet) build itself. Someone has to decide and bring up the capital to do it.
Opposing regulation for selfish reasons is something I can at least understand on a human level. But people who submit to market dynamics as the One Truth, even if they acknowledge themselves that those dynamic are harmful are entering cult territory IMO.
The kind of work we were "geared up to do" was what the industry was paying for at the time. For a very long time the industry has paid for trivia experts. If you were gluing together libraries and microservices, you resonate with the author. Not all "code" is the same, nor "programming."
I don't agree. For example, I think they're very good at writing physics simulators, they're great at writing shaders, they seem to have done a good job formally verifying the signal protocol[1].
That's not all code, to be sure but it's not what I've ever heard talked about as pipe fitting.
But aside from that, I think there are lots of things that made people's lives better and are now irrelevant, and people enjoyed that time. It's gone now, and we'll adapt. But it felt worth capturing the feeling.
There is a fifteen hundred and thirty sixth dimension beyond that which is known to man.
It is a dimension as vast as space and as timeless as infinity. It is the middle ground between light and shadow, between science and superstition, and it lies between the pit of man's fears and the summit of his knowledge.
This is the dimension of imagination. It is an area which we call the the embedding space.
> It also appears to be true of organizations too. How many startups were just handcrafted answers to a specific problems? The big labs are a general solution with loads of compute.
If he continues this worship of the "bitter lesson" to its logical conclusion, why stop at startups? All economic, or even human, activity could be seen as "handcrafted solutions".
So is the idea that the big labs should take over everything in the end? Why is that a good idea? Why would you want to live in such a world?
I think some people have come to that conclusion. Many have joined the big labs.
Whether they’re right or not, we might not get to choose. On some level we’re all just witnesses to this moment in history. It will be whatever it is and if the models keep improving that’s exactly what will happen.
If the s curve flattens out. Things might be different
Well, I guess, better join the people protesting the data centers then. This stuff does not (yet) build itself. Someone has to decide and bring up the capital to do it.
Opposing regulation for selfish reasons is something I can at least understand on a human level. But people who submit to market dynamics as the One Truth, even if they acknowledge themselves that those dynamic are harmful are entering cult territory IMO.
The kind of work we were "geared up to do" was what the industry was paying for at the time. For a very long time the industry has paid for trivia experts. If you were gluing together libraries and microservices, you resonate with the author. Not all "code" is the same, nor "programming."
I don't agree. For example, I think they're very good at writing physics simulators, they're great at writing shaders, they seem to have done a good job formally verifying the signal protocol[1].
That's not all code, to be sure but it's not what I've ever heard talked about as pipe fitting.
But aside from that, I think there are lots of things that made people's lives better and are now irrelevant, and people enjoyed that time. It's gone now, and we'll adapt. But it felt worth capturing the feeling.
[1]https://deepmind.google/blog/introducing-alphaproof-code/
I have a blog post on this drafted but not published. But the gist of it is, what gets more valuable when code gets cheaper?
That's what I'd focus on for a career or business, because code is getting darn cheap.
More general: what remains / becomes valuable as "thinking power" (+creativity) becomes cheaper & cheaper?
Shared human experience :)
Yeah, this. 100%. When the time cost of coding, creating, $anything approaches zero then taste becomes art. [0]
[0]: https://matthewsinclair.com/blog/0195-taste-as-art
It's not just code, it's "intelligence." Everybody has it now.
That was a good episode. I miss these spooky anthologies (Twilight Zone, The Night Gallery, etc...)
Before AI, chop wood, carry water. After AI, chop wood, carry water.
There is a fifteen hundred and thirty sixth dimension beyond that which is known to man.
It is a dimension as vast as space and as timeless as infinity. It is the middle ground between light and shadow, between science and superstition, and it lies between the pit of man's fears and the summit of his knowledge.
This is the dimension of imagination. It is an area which we call the the embedding space.