← Back to context

Comment by rileymat2

14 hours ago

I disagree, understanding systems and problem solving, understand technical capabilities, requirements and delivering a product is what always paid for by consumers. Paying programming labor was just a means to an end, removing that expensive and time consuming part creates a world with so many more possibilities.

A few years ago the owner of a local bar and friend of mine, comes up and asks about building a loyalty program and bar operating software. It sounded well thought out and really good for his business, but I told him, at that scale you would need 20 or 30 other bars to split the cost of development, it just was not economically feasible. But with what I have seen out of codex, I started talking to him about it again a few days ago.

Even with art, I have a lot of ideas and concepts but can't draw or paint. Now I can create images that I personally enjoy. I don't post this slop anywhere, but I really enjoy it. One of my favorites recently is a humanoid robot 3d printing a baby robot to give birth, I am sure no one else would like it, but it makes me happy.

It's a great future.

Cool, I don't care about any of the stuff you just said you shipped. I've got enough trash software in my life, and I'd pay to have it reduced. The things you listed are more stuff that I don't want to interact with.

  • I did not say I shipped anything.

    With open source you can also have it improved, and have less trash software in your life.

    Codex created a custom build of firefox with a feature I wanted (essentially the ability to override ssl dns like you get with curl --resolve option). Also, it was able to take apple-music-for-linux that I like for streaming apple music and update it so it would work again. There are incredible possibilities here.

    • And the downsides far outweigh the benefits. Your future still fucking sucks; it's just not good for humans to live in, no matter how many SSL DNS entries you override.

      5 replies →

>Even with art, I have a lot of ideas and concepts but can't draw or paint.

99.8% of artists all start out at a similar place. They have good "taste", but they suck at making art. Only through practice and feedback from other artists were they able to improve their skills and creative sight, and these things in turn further developed their "taste" in art.

The art pieces that people enjoy are the culmination of an artist spending countless hours practicing, failing, retrying, and experimenting with different ideas. You don't get a Zdzislaw Beksinski without endlessly iterating and honing your craft.

This is one of by biggest gripes with AI: it creates an incentive for people to skip all the hard work that results in great art. Why spend thousands of hours developing your art skills and your unique artistic fingerprint when you can press the instant gratification button right now? I know its not something you can blame a single individual for, but when society as a whole starts taking shortcuts that's when we'll really see a degradation in art.

  • But maybe you can get good at the more important "taste" part first and to a higher degree, and not the mechanical part. Old time artists (think Rembrandt) used to have make paint by hand, a consuming process, and partially poisonous. Modern artists, it's my understanding largely save all this part by buying it in tubes, giving them time or more important things.

    • Taste is essentially a set of heuristics for what has worked for you and the people you learned from, and how the details of your current situation feed into it. You could argue that AI has almost nothing but taste.

      For humans, taste is learned through friction against the details of your problem, and finding what ends up sliding smoothly through the details of the medium you're working in. For AI, it's the same, but it comes from the details of the training data.

      We've automated taste well enough for most production needs, and its only going to get better over time. If you don't think it'll be good enough for artistic needs too, you're placing a pretty hard bet against AI.

      "Taste is all that's left" is pure cope. Taste is one of the first things where humans become useless.

      4 replies →

    • It's debatable depending on who you're talking to whether taste is more "important" than execution/the mechanical part (I think calling it the mechanical part is a little misleading because that suggests one is brainlessly implementing a fixed algorithm). Maybe for the manager/CEO types who have a grand pie in the sky vision that they need an army to implement "taste" is more important. But I think for the artist themselves being able to spawn something into existence with their own labor is a large part of the experience.

      Yes, we'll have lots more art for people to enjoy, but fewer artists.

    • The point is that these are not separate. They cannot be separate.

      Deep understanding, taste, aesthetic, sensitivity to nuance, and so on...none of that can be developed without intimate involvement with the process, with the thing itself.

      Your example of poisonous paint vs buying it in tubes is a fallacy, because that isn't what we're talking about. We're talking about _actually painting_ vs just...telling someone else to paint.

    • >But maybe you can get good at the more important "taste" part first

      I keep putting quotes around "taste" because its hard to properly describe. Whatever it is, it seems like a mixture of observing what has worked for others, personal preference, worldview, etc. "Getting good at" taste is therefor hard to quantify.

      Furthermore, having good "taste" in the art you consume doesn't translate to having good "taste" in the art you create. There exists a gap between the appreciation and application of said "taste", and you need to practice creating a lot of stuff to shorten that gap. And again, these two sides of "taste" are deeply intertwined. The kind of art you consume tends to change as you hone your creative skills and begin to see the technique behind it all, and that in turn influences the kind of art you may wish to create.

      4 replies →

    • No, you don't understand what art is. Art emerges from a process. If you have no process, you can't make art. An idea, or a prompt, is not a process.

  • yeah there's no point developing skills anymore.

    • Understanding the underlying implementation can have value even if rarely used. I think learning assembly has had a profound effect on my understanding of higher level code, but there is no /real/ point to learning it with great compilers, pre-ai.