Comment by caseyy | Hacker News Reader

Comment by caseyy | Hacker News Reader

There is a small, hopeful flipside to this. While people using AI to produce art (such as concept art) have flooded the market, real skills now command a higher price than before.

To pull this out of the games industry for just a moment, imagine this: you are a business and need a logo produced. Would you hire someone at the market price who uses AI to generate something... sort of on-brand they most definitely cannot provide indemnity cover for (considering how many of these dubiously owned works they produce), or would you pay above the market price to have an artist make a logo for you that is guaranteed to be their own work? The answer is clear - you'd cough up the premium. This is now happening on platforms like UpWork and Fiverr. The prices for real human work have not decreased; they have shot up significantly.

It's also happening slowly in games. The concept artists who are skilled command a higher salary than those who rely on AI. If you depend on image-generating AI to do your work, I don't think many game industry companies would hire you. Only the start-ups that lack experience in game production, perhaps. But that part of the industry has always existed - the one made of dreamy projects with no prospect of being produced. It's not worth paying much attention to, except if you're an investor. In which case, obviously it's a bad investment.

Besides, just as machine-translated game localization isn't accepted by any serious publisher (because it is awful and can cause real reputational damage), I doubt any evident AI art would be allowed into the final game. Every single piece of that will need to be produced by humans for the foreseeable future.

If AI truly can produce games or many of their components, these games will form the baseline quality of cheap game groups on the marketplaces, just like in the logo example above. The buyer will pay a premium for a quality, human product. Well, at least until AI can meaningfully surpass humans in creativity - the models we have now can only mimic and there isn't a clear way to make them surpass.

> real skills now command a higher price than before.

Only if companies value/recognize those real skills over that of the alternative, and even if they do, companies are pretty notorious for choosing whatever is cheapest/easiest (or perceived to be).

> There is a small, hopeful flipside to this. While people using AI to produce art (such as concept art) have flooded the market, real skills now command a higher price than before.

It's "hopeful" that the future of all culture will resemble food, where the majority have access to McDonalds type slop while the rich enjoy artisan culture?

It's hopeful because AI has not devalued creative human labor but increased its worth. Similar to how if one were a skilled chef, they didn't start working for McDonald's when it came to be, but for a restaurant that pays significantly above McDonald's.

Most people's purchasing power being reduced is a separate matter, more related to the eroding middle class and greedflation. Many things can be said about it, but they are less related to the trend I highlighted. Even if, supposing the middle class erosion continues, the scenario you suggest may very well play out.

It doesn't make sense to suggest that AI has made human effort more valuable. Before, to do X, Y, or Z you needed human effort. Now, you can do X with AI. You just need human effort to do Y or Z. There is less demand for human effort. Why would that result in an increase in the price of human effort?

>Most people's purchasing power being reduced is a separate matter, more related to the eroding middle class and greedflation.

Greedflation, is that where companies suddenly remember to be greedy again after years of forgetting they're allowed to be greedy, which happens by random chance to coincide exactly with periods of expansionary monetary and fiscal policy?

> It doesn't make sense to suggest that AI has made human effort more valuable.

In that case, I welcome an alternative explanation for the human labor price increase on UpWork and Fiverr while AI work replaced work at the previous price level. The same is seen in the hiring of affected disciplines.

If you have a distribution of work where most is easy and cheap, some is moderate and moderate, and a little is difficult and expensive, and you take out all the cheap and easy work, the moderate and difficult work could drop in price but the average of the remaining work will still be higher than before.

e.g.

You have tasks advertised in the distribution $1, $1, $1, $1, $1, $1, $2, $2, $3, $3, $5, $5, $10. Median price is $2, and average is $2.76.

All the $1 and $2 tasks are replaced with AI. Old tasks get $1 cheaper each as there are more people that can do them. Now the distribution is $2, $2, $4, $4, $9. Median is $4, average is $4.2.

So you have made labour less valuable but the prices advertised go up because only the more expensive work now gets advertised.

The situation with food is that everyone today has access to good quality food if they choose to actually put their money towards it, but large numbers of people enjoy McDonalds and KFC and such slop, so they choose to spend far more on it than they'd spend cooking for themselves.

It is still much better than when large numbers of people starved if it rained a bit in the wrong week.

    He spoke of the grass and flowers and trees
    Of the singing birds and the humming bees;
    Then talked of the haying, and wondered whether
    The cloud in the west would bring foul weather.

The weather and its effect on the food supply was the preoccupation of 90% of the population 90% of the time for all of agricultural man's history (and pre-history) and hunting and gathering was even worse for quality of life.

Reply on news.ycombinator.com