[arrrg]

To me the hard problem isn’t building things, it’s knowing what to build (finding the things that provide value) and how to build it (e.g. finding novel approaches to doing something that makes something possible that wasn’t possible before).

I don’t see AI helping with knowing what to build at all and I also don’t see AI finding novel approaches to anything.

Sure, I do think there is some unrealized potential somewhere in terms of relatively low value things nobody built before because it just wasn’t worth the time investment – but those things are necessarily relatively low value (or else it would have been worth it to build it) and as such also relatively limited.

Software has amazing economies of scale. So I don’t think the builder/tool analogy works at all. The economics don’t map. Since you only have to build software once and then it doesn’t matter how often you use it (yeah, a simplification) even pretty low value things have always been worth building. In other words: there is tons of software out there. That’s not the issue. The issue is: what it the right software and can it solve my problems?

[pzo]

> To me the hard problem isn’t building things, it’s knowing what to build (finding the things that provide value) and how to build it (e.g. finding novel approaches to doing something that makes something possible that wasn’t possible before).

The problem with this that after doing this hard work someone can just copy easily your hard work and UI/UX taste. I think distribution will be very important in the future.

We might end up that in future that you have already in social media where influencers copy someones post/video and not giving credits to original author.

[kombookcha]

>The problem with this that after doing this hard work someone can just copy easily your hard work and UI/UX taste.

Or indeed, somebody might steal and launder your work by scooping them up into a training set for their model and letting it spit out sloppy versions of your thing.

[shinycode]

I agree. It’s really easier to build low-impact tools for personal use. I managed to produce tools I would never have had time to build and I use them everyday. But I will never sell them because it’s tailored to my needs and it makes no sense to open source anything nowadays. For work it’s different, product teams still need to decide what to build and what is helpful to the clients. Our bugs are not self-fixed by AI yet. I think Anthropic saying 100% of their code is AI generated is a marketing stunt. They have all reasons to say that to sell their tool that generates code. It sends a strong signal to the industry that if they can do it, it could be easier for smaller companies. We are not there yet from a client perspective asking a feature and the new feature is shipped 2 days later in prod without human interactions

[c048]

I wonder what happened to the old addage of "only 10% of the time you actually spend coding, the rest of the time is figuring out what is needed".

At the same time I see people claiming 100x increases and how they produce 15k lines of code each day thanks to AI, but all I can wonder is how these people managed to find 100x work that needed to be done.

[rtcoms]

For m, I'm demotivated to work on many ideas thinking that anyone can easily copy it or OpenClaw/Nanobot will easily replicate 90% of that unctionality.

So now need to think of different kind of ideas, something on line of games that may take multiple iteration to get perfected.

[pixl97]

>thinking that anyone can easily copy it

I mean this is how it's always been throughout history.

Creating something new is hard, copying something in terms of energy spent, is far easier. This is software or physical objects that don't require massive amounts of expensive technology to reproduce.

[meowface]

In a few decades, AIs will probably be better at those than most humans. Possibly even sooner.