AI: The Rise of the Machines... Or Just a Lot of Overhyped Chatbots?

Sanrith Descartes said:

Seems tangentially appropriate to this thread so posting it here. I have no idea what app this guy built and don't know that I am going to look but he is correct. We are at an inflection point we haven't been at in a long fucking time. There is opportunity for everyone with half a brain to enter the app market and they no longer have to understand how to code. They just need an idea and to find a space that needs that idea and that isn't already crowded with entrants.

@Amod this thread has encompassed AI in general now that ChatGPT (OpenAI) is no longer the only available LLM. Want to rename the thread to just the AI thread?

Click to expand...

Well, kind of. If not being able to code (or write, or draw, or whatever someone is using AI for) was the only thing stopping someone from being a millionaire, why didn't they learn to code? They had better stuff to do than learning something that was apparently worth millions to them? And why aren't all coders currently running million dollar businesses? Coding was never the important part.

I'm sure some people will make a ton of money with their pop-up AI businesses, but people have been writing similar books for years: make money with etsy, amazon, ebay, seo, dropshipping, etc. None of them required coding. Just like with the others, a few will make a killing and then copycats will flood their niche. The difficulty in doing something is a big part of what makes it valuable. If someone can set up a business in an hour with no special skill or experience, so can everyone else.

I think the magic really happens when people who already have skills and experience are able to leverage those to do something that wouldn't be possible (or at least obvious) to others. I mean, this was always true, but AI acceleration might make it truer.

Leadsalad said:

Click to expand...

I made a similar post here somewhere after looking at the prices of new hardware versus the rental prices, but I don't think the infrastructure is as short lived as they're saying. The rental places are still renting 2018 quadro cards, and the new models getting released are getting more efficient, so you can do more with (relative) potato hardware than before.

Control said:

I made a similar post here somewhere after looking at the prices of new hardware versus the rental prices, but I don't think the infrastructure is as short lived as they're saying. The rental places are still renting 2018 quadro cards, and the new models getting released are getting more efficient, so you can do more with (relative) potato hardware than before.

Click to expand...

Thats one aspect of it, but if you remember the movie "transcendence" with Johnny Depp - they've built entire computer installations that are pretty similar to whats in that movie. Billions and billions of dollars in infrastructure in these really low value locations. I dont think those installations will ever pay off in the long run. But, similar to what they were saying about Fiber internet - They'll prove useful for at least a decade with just current gen nvidia architecture installed in them.

Daidraco said:

Thats one aspect of it, but if you remember the movie "transcendence" with Johnny Depp - they've built entire computer installations that are pretty similar to whats in that movie. Billions and billions of dollars in infrastructure in these really low value locations. I dont think those installations will ever pay off in the long run. But, similar to what they were saying about Fiber internet - They'll prove useful for at least a decade with just current gen nvidia architecture installed in them.

Click to expand...

No idea about that part, but it does look like big money is betting on them being useful somehow. Maybe it's just hype, but I have to wonder about the assumptions they're making. How different would the ROI calcs look if you assume that 50% of current employees will be AI in 10 years (for example)?

On the individual level, I look at all the models being released and try to imagine how to make use of them or how to put them to work for me, and I struggle a bit. Personally, I don't think I really need much that it does, but it's probably like my grandparents looking at an iPhone. They got along fine without it for decades, so why do they need it now?

Professionally, maybe. If I can use it to speed up things I can already do, that's good, but it's not exactly paradigm changing. If I can use it to do something I couldn't otherwise do, maybe if it's doing something that far outside of my abilities, I probably wouldn't have the knowledge/skill to put it to use. I know there are "exceptions" with kids vibecoding apps that make millions of dollars, but the coding was never the hard part about that, it was the deciding what to make and how to market/monetize. Ultra fast iteration to see if anything catches on? Maybe, but throwing spaghetti at the wall makes me a little nauseous.