[Sharlin]

Honestly, what are the positive viewpoints of generative AI in the end? Are there others major ones than the following?

* My vibe coding machine goes brrrrt and that's all I care about

* My college essay cheating machine goes brrrt and that's all I care about

* My custom waifu/porn-generating machine goes brrrt and that's all I care about

* The concept of AI is drawing all the investor money and that's all I care about

The common factor being self-centeredness and/or being part of a small ingroup that benefits, possibly at the expense of others.

[stetrain]

The positive viewpoint is basically like the Industrial Revolution or the post-WWII consumer/convenience boom.

If productivity can increase significantly per worker, the result will be major overall economic growth.

It might be sold to consumers the way vacuums and washing machines were. With these automated modern conveniences you'll spend less time working and have more time for leisure.

Of course the reality for the actual workers on the line is that their job and industry may be disrupted and the overall benefits of that economic growth may not reach them during their lifetime. The Industrial Revolution was followed by a century of major and sometimes violent disputes over the relationship between corporations and labor and the rights of workers.

The post-WWII promises of convenience and leisure were replaced by the reality of the baseline adjusting and households needing to work the same or even more combined hours to make ends meet.

Even if the optimistic levels of economic growth occur, the benefits are unlikely to be evenly distributed.

[watwut]

Industrial revolution was pretty much disaster for average workers. It took a lot of literal fights till things got better.

[kranke155]

Decades. It took decades. My understanding is a lot of historians see the first decades as major economic growth while most British people actually became smaller and shorter due to malnutrition.

[Macha]

Yeah it’s easy for people now to say “Look how great the Industrial Revolution was for us, the descendants of the people who didn’t die of cholera in an overcrowded tenement” and assume the Luddites were wrong on the impact on them, but that’s an over simplification

[tonyedgecombe]

The funny thing is most of the evangelists aren’t really in the in group and will be just as exposed to the results as the rest of us.

[rileyphone]

Here's one: AI democratizes the ability to produce software, which has mostly been an arcane craft wielded by a priestly class. Now anyone, if they know what they want and it isn't too complex, can talk to AI and get working (if not also janky) software in a very short amount of time. Hopefully this breaks the grip that platforms/large corporations have on personal software and the internet.

[Sharlin]

I would counter that:

a) it seems likely to me that in the end, few normal people know what to do with the ability to create their own software for their private use,

b) getting bespoke software working on the platforms that the majority of people actually use (Android and iOS) is somewhere between hard and impossible, and

c) large corporations have a de facto grip on AI as well, local models require you to have the knowhow and beefy hardware to run them, and they’re not magic software machines like Claude.

All in all, it seems rathet optimistic that AIs could do much if anything to help consumers against corporations. But I concede that it is a viewpoint that’s at least less selfish than most.

[TheOtherHobbes]

It's an incredible search, research, and learning tool, and far better than a search engine. You can get almost anything explained at any level up to undergrad, with the option to ask questions if you don't understand, and with links to references, so you can check that what you're learning is correct.

The low-quality content machine angle is one of the least interesting things about it.

[duped]

But the "research" you're getting from the AI is also low-quality content. And particularly susceptible to the X-Y problem because unless you're already learned in a subject you won't even know how to craft the prompt to get the answer you're looking for.

[TheOtherHobbes]

It really isn't. I'm not talking PhD grade R&D, I'm talking about everyday queries that take a long time manually and are easily automated with AI.

Without going into details, I have used AI to find genuine, provably effective solutions to multiple real world problems that would either have been impossible without AI or would have taken a very, very long time.

And it would have been a boon if it had been around while I was getting my degree, because it's been excellent at clarifying foundational concepts.

It's not perfectly reliable, but neither are human professionals.

[duped]

A pattern I have seen repeatedly for the last year is people who assert this, then have a problem and come to me/people like me for help, only to find that the basic research they did was complete garbage.

AI is very, very good at spitting out language. It's very, very bad at ascertaining the quality or correctness of the language it spits out.

And the problem with plausibly correct output is that sometimes it's good enough until it isn't, which is the pattern of behavior I've seen with heavy AI driven research.

I'm not going to discount that it's better than google searches. But it's hardly good enough to equate to undergraduate understanding. The act of reading and writing from primary/secondary sources crafted by humans is in itself the way to acquire and retain knowledge. Having it hallucinated at you in a plausibly correct way is dangerous to equate to that.