Hacker News Reader

Comment by dentemple | Hacker News Reader

There's a potential future where these discussions (about whether to work with AI or work against it) will be happening across ALL industries, not just in the x-rated one.

We're seeing it here and in the graphic design industry right now because that's what the large majority of people are using AI for. But just like with photography, and just like with the internet, it's not going to stop there.

In my personal opinion, this wave is coming EVEN IF AI ends up not being as perfect as people are currently envisioning it as. Because even if it's not perfect, it's convenient and that's already been proven thanks to the likes of ChatGPT and Midjourney.

The AI Revolution is here, and, unfortunately, everyone will need to figure out their place in this new world.

View on HN

> There's a potential future where these discussions (about whether to work with AI or work against it) will be happening across ALL industries, not just in the x-rated one.

Uh, that’s not a potential future, its today. It's happening in tech, today (as mentioned in TFA, that’s part of what got Reid interested), it is a big part of the issues behind the recent SAG-AFTRA and WGA strikes in mainstream film/TV. It's happening across a wide range of industries today, not just porn.

I read another article on HN where an illustrator was saying that if the tools evolve to be more designed for artists, they could be great for cutting down time-consuming aspects of illustration. I’m not an illustrator but I’m guessing things like coloring and shading are areas where AI could make quick work.

We saw technological innovations like Disney’s Xerox machine technique that was used for 101 Dalmatians, and of course we saw computer animation basically decimate hand-drawn animation almost entirely. Even in areas where animation is hand drawn, it’s usually done so in a digital workflow.

The people out there that are scared of AI and refuse to use it are making a mistake. It’s here. It’s not going to get less capable, it’s going to keep getting more capable. I think the average person should try to be an early adopter to have an advantage in our new world.

Hell, it can give you a finished image

What does it mean to "work with AI" in the context of the future or even near future? A symbolic job of checking it's outputs?

Even if AI doesn't take your job, it almost certainly will take most of the value it has. I don't need a $175k dev when I can use a model and a $60k dev to do the same thing. Or on the graphics side, artists are a dime a dozen, I don't need the cream of the crop if they are just gonna be fixing up mangled AI hands.

> Even if AI doesn't take your job, it almost certainly will take most of the value it has. I don't need a $175k dev when I can use a model and a $60k dev to do the same thing.

Current trends in writing & editing appear to be that you need a rare and expensive person to get the kind of whole-number multiplier out of AI tools that people initially expected. And it's still not a large whole number.

Give the tools to someone who's not got exception taste and skill, and you're lucky if the multiplier's greater than 1 at all.

[EDIT] Exception if your output is expected to be crap and you don't care about accuracy. It's a huge boon to spam of most forms, and astroturfing.

I think it's a mistake to use current AI in current times as the status quo in ~5 years from now.

AI will likely improve dramatically, as will the general persons ability to work with it. I wouldn't be one bit surprised if typewriter operators outpaced nascent word processor users in the beginning.

> AI will likely improve dramatically, as will the general persons ability to work with it.

I'm not so sure about ability to work with it. Middling Excel skills still make you a wizard able to do things nobody else can, in most offices (though not most tech offices). Hell, I've seen knowing how to use the Windows file explorer at a very low level of competence put a person in the top five percent by computer ability in an office. Within the last couple years.

Most folks can barely use the tools we have, and aren't going to be able to get good output from anything like the current set of AI tools. The ones who can, won't be cheap, and may be hard to replace like-for-like—I think a lot of companies are going to settle on more, cheaper, more-interchangeable workers than replacing them with a few AI tool experts with hard-to-measure-or-quantify skills.

The tools themselves need an added ability to reflect and distinguish reality from imagination and suited-to-task from unsuited-to-task, to get as-much better as they need to be to deliver what some are supposing they surely will, and AFAIK that's not the kind of thing we're going to get with "LLMs, but more and bigger" and related techniques, but will require some new breakthroughs.

> Hell, I've seen knowing how to use the Windows file explorer at a very low level of competence put a person in the top five percent by computer ability in an office. Within the last couple years.

This is a really interesting point, phrased in a way that I wouldn't have thought to vocalize but that resonates with me. Tool-based and infrastructure-based changes that vastly improve efficiency and output based on competency with the tools/infrastructure are common; there's a lot of low-hanging fruit that is nevertheless uncommon for people to grab and use.

Assuming current AI generation improves in exactly the way its proponents are expecting (which is it's own separate conversation), it is still probably worth spending a little time trying to figure out whether it's a tool that will actually see mass adoption or whether it'll end up in the same category as many of the other excellent tools that exist today that have remained inaccessible or unattractive to most people.

Some tools do get mass adoption and people become competent with them at a mass-level. But only some, and it doesn't seem to strictly map to the quality of the tool. I'm not sure I can easily tell/predict the difference and where a theoretical future AI would fall; I would need to think about it more. But I just wanted to chime in and say that it's a good insight.

> Middling Excel skills still make you a wizard able to do things nobody else can, in most offices (though not most tech offices)

Do those skills translate directly to a job/pay, or are they simply nice-to-have competencies inside other jobs? I've always worked as an ops/analyst/spreadsheet guy but I've found over the years that those formula manipulation skills never actually translate to "advancement". They enable me and my company to do things easier than they would have otherwise, but appear as nice-to-have rather than valued in it's own right.

Put another way - Is there actually a career where the primary hard skill is "I'm good at excel"?

At the current state of the art, it looks more like keeping the $175K dev and laying off four $60K devs.

Which may be worse, because that $175k dev probably didn't walk out of college at that rate. Who needs a next generation when you can profit today.

Productivity-increasing tools increase the value of human labor (they don’t take it), but, in a capitalist society, also facilitate the capitalist capturing more of the value of that labor.

Reply on news.ycombinator.com