They’re investing into huge, energy intensive compute resources that aren’t going to pay off for at least a decade, and meanwhile investors are going to want returns on those investments ASAP. They need to fill warehouses with compute and power them with nuclear reactors, but there’s no profitability model. That means stranded assets, especially if investment dries up and they can’t pay or if demand shifts away from their models. This is set up to be a massive crash.
NVIDIA will probably be fine though.
Gigabrain (already linked) and Perplexity does this:
Yeah, and what they’ll do is invent sources from thin air or draw made up conclusions from real sources. They’re just LLMs, no matter how much data you feed them and how much the results are tinkered with they only regurgitate a statistically likely answer. Perplexity is a bullshit machine. It’s fine if you don’t really care about the answer and are just kind of curious, but no serious researcher should ever rely on a chatbot.
For who? Who is going to crash massively? Google? Microsoft? Amazon? Are you are expecting these massively diversified trillion dollar companies to fail due to AI?
Yeah, and what they’ll do is invent sources from thin air
The sources are right there next to it? You click on them and it takes you to the source, could you maybe try it for 5 seconds and then get back to me before you just make stuff up? what are you, an AI?
or draw made up conclusions from real sources
This feels like I’m having a conversation with a boomer talking about wikipedia.
Yeah, it’s always best to check the original sources and not just believe everything you read on the internet, no different than clicking on results in google and getting a page full of misinformation which people are doing every minute of every hour of every day, and don’t even get me started on social media.
For who? Who is going to crash massively? Google? Microsoft? Amazon? Are you are expecting these massively diversified trillion dollar companies to fail due to AI?
Open AI is going to implode after it goes for-profit. As for the others they’ll weather the storm, they have enough diversity in their assets to handle the AI bubble popping, but there will be big tech layoffs and lots of assets will get sold off to private equity.
You click on them and it takes you to the source, could you maybe try it for 5 seconds and then get back to me before you just make stuff up?
So what’s the point?
This feels like I’m having a conversation with a boomer talking about wikipedia.
Rude. Wikipedia is, at least, peer reviewed by wikipedia editors. Chatbots don’t have that. They will just make shit up and you have to manually double check their sources yourself. At that point, why are you even using AI? It saved you no time or effort.
This feels like having a conversation with someone inside a hype bubble. If Wikipedia already exists, what purpose does AI fulfill? It’s just a more expensive, more energy intensive way to do the exact same thing. There’s no profitability case. It’s useful, but it isn’t more useful than the much cheaper and much less energy/resource intensive alternatives. So, what’s the point?
Yeah, it’s always best to check the original sources and not just believe everything you read on the internet, no different than clicking on results in google and getting a page full of misinformation which people are doing every minute of every hour of every day, and don’t even get me started on social media.
Okay, but then, why is AI useful? If you’re going to look at sources anyway, what’s the point? You’re just using a massive amount of energy and compute for something that can be done much more efficiently.
The only useful product I’ve seen come out of this is hype bubble is text-to-image models. Being able to tell a bot to generate an image is really interesting and useful for people without skills in creating or editing their own images. That’s an actual use case that could maybe justify the amount of resources being poured into it, it could maybe even be profitable.
They’re investing into huge, energy intensive compute resources that aren’t going to pay off for at least a decade, and meanwhile investors are going to want returns on those investments ASAP. They need to fill warehouses with compute and power them with nuclear reactors, but there’s no profitability model. That means stranded assets, especially if investment dries up and they can’t pay or if demand shifts away from their models. This is set up to be a massive crash.
NVIDIA will probably be fine though.
Yeah, and what they’ll do is invent sources from thin air or draw made up conclusions from real sources. They’re just LLMs, no matter how much data you feed them and how much the results are tinkered with they only regurgitate a statistically likely answer. Perplexity is a bullshit machine. It’s fine if you don’t really care about the answer and are just kind of curious, but no serious researcher should ever rely on a chatbot.
For who? Who is going to crash massively? Google? Microsoft? Amazon? Are you are expecting these massively diversified trillion dollar companies to fail due to AI?
The sources are right there next to it? You click on them and it takes you to the source, could you maybe try it for 5 seconds and then get back to me before you just make stuff up? what are you, an AI?
This feels like I’m having a conversation with a boomer talking about wikipedia.
Yeah, it’s always best to check the original sources and not just believe everything you read on the internet, no different than clicking on results in google and getting a page full of misinformation which people are doing every minute of every hour of every day, and don’t even get me started on social media.
Open AI is going to implode after it goes for-profit. As for the others they’ll weather the storm, they have enough diversity in their assets to handle the AI bubble popping, but there will be big tech layoffs and lots of assets will get sold off to private equity.
So what’s the point?
Rude. Wikipedia is, at least, peer reviewed by wikipedia editors. Chatbots don’t have that. They will just make shit up and you have to manually double check their sources yourself. At that point, why are you even using AI? It saved you no time or effort.
This feels like having a conversation with someone inside a hype bubble. If Wikipedia already exists, what purpose does AI fulfill? It’s just a more expensive, more energy intensive way to do the exact same thing. There’s no profitability case. It’s useful, but it isn’t more useful than the much cheaper and much less energy/resource intensive alternatives. So, what’s the point?
Okay, but then, why is AI useful? If you’re going to look at sources anyway, what’s the point? You’re just using a massive amount of energy and compute for something that can be done much more efficiently.
The only useful product I’ve seen come out of this is hype bubble is text-to-image models. Being able to tell a bot to generate an image is really interesting and useful for people without skills in creating or editing their own images. That’s an actual use case that could maybe justify the amount of resources being poured into it, it could maybe even be profitable.
The rest? It’s wasteful and it won’t last.