Hacker News Reader

Comment by floundy | Hacker News Reader

I had a (definitely non-billionaire) coworker crowing to me about the scientific discoveries AI was going to make the other day. I was poking him about how that would happen, how the AI could do anything other than recite and recombine based on tokenized probability of what it had ingested, and he kept using words like "think" to refer to the AI.

Then he's going on about how "AI can write anything you want, in the style of any author that's ever existed" so I asked him "what value is added to the world by having text that looks like it might have been written by Shakespeare but really wasn't" and he didn't have an answer.

Blind worship of AI is rapidly becoming one of my biggest tells of low cognition humans. I'm just so tired of hearing about AI and I'm ready to move onto the next "big thing".

View on HN

> so I asked him "what value is added to the world by having text that looks like it might have been written by Shakespeare but really wasn't" and he didn't have an answer.

Satire. For instance, "You are Alex Jones during the French revolution. Give a news broadcast in old English covering the execution of King Louis XVI at the hands of radical leftists."

But other than humor and getting inspiration for your own writing, I'm not sure what value it really holds.

> Blind worship of AI is rapidly becoming one of my biggest tells of low cognition humans. I'm just so tired of hearing about AI and I'm ready to move onto the next "big thing".

100% agree. It's like the anti-nuclear movement: the people most invested in it either have a financial stake in it (like coal and oil companies astroturfing) or favor vibes over engineering principles.

I've said it before and I'll say it again: "We're going all in on AI" is the modern equivalent to adding .com to your company name to juice the stock price. It's utterly meaningless, yet it works shockingly well.

> "what value is added to the world by having text that looks like it might have been written by Shakespeare but really wasn't"

This is an easy question to answer for anyone like myself who actually enjoys reading Shakespeare. Once you've read all that is available, it would be amazing to have more. Whether it was written by Shakespeare himself or AI, is irrelevant once the AI gets to the point where it can do the same thing. It's obviously not there yet, and maybe it never can be, but if it is, that would be a huge value add to the world.

Think about it in other ways. Would you like to have a 4th season of Star Trek the original series, or more seasons of TNG or DS9 or Voyager or Enterprise? I sure would! Maybe it doesn't add any value to the world in your opinion, but part of the value I get out of the world is entertainment and enjoying great stories. It definitely adds value to me, and I'm technically in the world, so even if it's only a trivial amount in the grand scheme of things, it is still adding value.

On the subject of scientific advancement, there are surely mountains of discoveries just waiting for us in the data we've already collected, but as humans we're not capable of fully recognizing Long difficult patterns, especially when there are many variables involved. However, AI is getting very good at that. We Don't necessarily need AI to conduct original research and studies, but we can absolutely use it to comb through data and look for things we may have missed. The reasoning models are also getting pretty good and will surely get much better with time, and will be able to crunch through long chains of data looking for patterns. Working in tandem with a human scientist, there is potential even now. As AI gets more capable of being agentic, I think that will only increase.

The people who worship AI are in my opinion misguided at best, but that doesn't mean that AI is a nothing burger

>Whether it was written by Shakespeare himself or AI, is irrelevant

I couldn't disagree more, the products of AI lack all value of the original works, namely that it was created through the mental effort of a real human, is known to this day due to being impactful or influential in some way, and that the output of the originals is limited in scope to what the author produced during their lifetime (scarcity).

>On the subject of scientific advancement, there are surely mountains of discoveries just waiting for us in the data we've already collected, but as humans we're not capable of fully recognizing Long difficult patterns, especially when there are many variables involved. However, AI is getting very good at that.

This is doing nothing other than adding a chatbot GUI to existing statistical software analysis tools. It might make it "easier to use" but lowering the bar to entry, will (IMO) allow people who lack the skills to verify the AI's claims and hypotheses to proliferate. At some point in the future after deployment of this, the humans' skills at verifying and validating will atrophy and the AI will be leaned on more heavily, when at the end of the day it's really just code. We're already seeing this with the kids that just graduated college, they had ChatGPT since their sophomore year. Most of them can't write, can't reason, can't code, can't do anything without their little AI.

I'm hearing kids in interviews respond with "I'd ask AI to do it." I ask them, "the AI says it's not familiar with what you're asking, what would you do then, describe the process for approaching the problem." Most of them can't. Young engineers are being completely stumped when I ask them about verification and validation of what the AI has produced.

I can see that we have a fundamental different understanding of what "AI" is, and also of what "value" is:

> This is doing nothing other than adding a chatbot GUI to existing statistical software analysis tools.

If that's all you think AI is and all that people are doing with it, I would respectfully suggest that you don't know what you're talking about. The chat interfaces are just UIs. Most researchers are not just typing prompts into a chat UI. If that's all it was, then I would agree with you, but it's not

> > Whether it was written by Shakespeare himself or AI, is irrelevant

> I couldn't disagree more, the products of AI lack all value of the original works, namely that it was created through the mental effort of a real human, is known to this day due to being impactful or influential in some way, and that the output of the originals is limited in scope to what the author produced during their lifetime (scarcity).

Firstly, you did clip a sentence fragment and take it out of context (the rest of that sentence adds an important qualifier) so I encourage others to actually read what I wrote and not just this clip. I think you did it in good faith but I just wanted to state that this clip alone is not representative of my view.

Why does it being created through the mental effort of a real human give it more value? If you were looking at a human created painting and an AI one and you couldn't tell the difference, then what does it matter for whether you enjoy it or not? What if there were other intelligent species capable of creating art? Would their production not be valuable because they aren't human? IMHO art is valuable because of the enjoyment it provides to people, not because of who created it.

> Why does it being created through the mental effort of a real human give it more value?

There's no perfect answer to this, but we do know it's the case.

It's entirely feasible to get a very, very good reproduction of the Mona Lisa, but that reproduction won't be worth anything near what the real one is.

When you say "worth" are you talking strictly monetary value? If so then you're clearly right.

But to me the "worth" of the Mona Lisa isn't it's monetary value. It's the feelings I get when I look at it. There is some clear historical value in that it captured images of the time, but that doesn't diminish with a reproduction in my mind either.

Unless we can all get on the same page about what we mean when we say "worth" or "value", we're not likely going to get anywhere in this discussion

> It's the feelings I get when I look at it.

I'd agree. But those and the monetary worth are often related. The Mona Lisa's original is worth a lot, in part, because of human feelings about it being The Real One.

My kids made some art in kindergarten I treasure to this day. I would probably not treasure a mass-produced perfect copy of it; if I saw one in Target it wouldn't be worth a dollar to me.

I had another thought on this that might be where the disconnect is. When I say it adds value, I'm not saying equivalent value. AI generated Shakespeare or Star Trek etc wouldn't (likely) be the same value as the originals, but they add some value, even if that value is just a couple hours of entertainment for a single individual. Part of the beauty of AI is the empowerment it will give to the individual that previously required huge amounts of investment and work by professionals, which necessitated that films/books/etc be widely marketable and economical. Even consider the difference between 1930s films and films today where originally it was "mass appeal to all audiences" but now there is a ton of variety for all sorts of different tastes. When AI gets to the point it can produce stuff people actually want to watch, the potential for hyper customization is huge. I don't think that will eliminate broadly appealing productions as we humans still love to share and talk about what we watched, and jump on bandwagons for the shared culture (Severance being a good example haha), but it will add more arrows to our quivers.

> Once you've read all that is available, it would be amazing to have more.

There isn't more. There never will be more. Shakespeare's dead. He died 409 years ago.

Instead of ghoulishly synthesizing a fake, confected Shakespeare, re-read the actual Shakespeare. Read deeper.

Reply on news.ycombinator.com