Remix.run Logo
▲ tombert 36 minutes ago

I'm not sure I agree with this argument, but I think the argument is that the amount of work to do will not scale with the productivity gains. If one software engineer can now do the work of what previously took four engineers, that's great, but that doesn't necessarily imply that there will be four times as much total work to do.

▲pegasus 31 minutes ago | parent | next [-]

My feeling is that there's an unlimited amount of work left to do. Software today sucks in so many ways. It's buggy and inflexible and limited in so many ways. My bet is on Jevon's paradox kicking in big-time.

▲tgsovlerkhgsel 29 minutes ago | parent | next [-]

Software tends to be made only good enough to do the job. There is an unlimited amount of work that could be done but not an unlimited amount of work companies are willing to fund, and the bar for "good enough" keeps dropping.

▲macintux 29 minutes ago | parent | prev [-]

Sure, there's an unlimited amount of work out there. But is there an unlimited amount of money to fund it?

▲tombert 3 minutes ago | parent [-]

I don't think that that logic follows. For most jobs, even before generative AI, we weren't paying "per work item", we were paying "per person per time". We could conceivably build 4x as much for the same cost, and it would require the same amount of money.

▲unixfreak0037 21 minutes ago | parent | prev | next [-]

Just sharing what I'm experiencing, we're using AI heavily at work, and I am more busy, with more things to do, than ever before, and I'm getting more done than ever before. And now we have these "software factories" coming. With AI it seems like so many new possibilities have opened up, because we can create anything we need now.

▲dist-epoch 18 minutes ago | parent [-]

> we can create anything we need now.

in aggregate there is not as much need as we can create now

▲brightball 32 minutes ago | parent | prev | next [-]

We’ve seen the 1 to 4 multiplier for years in more productive languages.

It was trivial for 1 Rails dev in 2010 to do 10-20x more than a Java dev back then.

Today’s world is an interesting dynamic but I’m pretty excited about it.

▲tombert 28 minutes ago | parent [-]

Sure, I wasn't really around for a time before high level languages, but I suspect the transition from machine code to FORTRAN was probably like a 10x productivity gain.

I tend to be more on the "Jevons" side of things, where I do think that demand will more or less expand. There are lots of projects I've been putting off for years because there would be an aspect about it that I hate doing and don't have the money to hire people to do for me (e.g. frontend web development). Those are projects that otherwise wouldn't exist that do now exist and potentially might make money or something (though with my business acumen that's debatable).

▲sunrunner 17 minutes ago | parent | prev | next [-]

If it's anything like much of recent history, the amount of work to do will simply increase while the nature of the work itself will continue to get more abstracted, less meaningful, and more disconnected from anything tangible. More abstracted, ever more distant.

▲idontwantthis 31 minutes ago | parent | prev [-]

In my experience there has been. I've been a part of two major rewrites that just wouldn't have happened without AI. Personally, I'm making 40% more than I was 3 years ago. I'm seeing ambitious work being done that just would have been unthinkable 3 years ago. We no longer have to trudge through legacy code that is too expensive to touch.

▲gib444 27 minutes ago | parent [-]

That's great. But do you worry about providing huge amount of training data to the AI companies who for example eventually/soon offer /rewrite in the CLI? That there'll be a point soon when there's nothing left to take from us?