[Tom'sHardware] AI bubble is worse than the dot-com crash that erased trillions, economist warns
- Thread starter SomeNorseGuy
- Start date
- Business Analysis
poppabk
Cheeks Spread for Digital Only Future
Yeah the crash happened because everyone was trying to get in on the ground floor and you just had every stupid idea getting buckets of money thrown at it and they were all burning through it trying to expand as fast as possible.
I still remember this ad as being indicative of the largesse at that time.
bitbydeath
Member
Even the corner cases are heavily reliant on people keeping it steered in the right direction.
It would be a huge disaster if they ever tried AI without the constant human intervention.
Many companies will find this out the hard way in the near future.
Codeblew
Member
That is not true at all. it makes me at least 2x more productive when doing coding sessions and it has pretty much completely replaced google/stackoverflow.
This is true though.
Last edited:
You bring up WebVan but have you heard of Builder.AI? Far worse a scam and collapse. Even had MS funding it.
StreetsofBeige
Member
I have no experience with AI as my job and company arent exactly AI driven.
But with AI being the rave and being jammed into every company (supposedly) and people are scared shitless about losing their job to AI bots, in your opinion what's the downlow on that topic?
Which jobs are the winners and losers? Sounds like youre on the winning end holding a tech job and doing better with AI. But then you got all the scare stories about AI taking over their job and they get fired.
StreetsofBeige
Member
Last edited:
If you want a quick history of the company this is a better page:
From $1.3B to Bust: The Rise and Collapse of Builder.ai
The Leading Platform AI Agents and Agentic Process Automation. Build AI, Integrate, Scale, all on One Platform.
beam.ai
Ironically AI
Last edited:
Codeblew
Member
That is a very tough question. I think for software development, AI just helps you be a lot more productive and potentially lead you in the direction of writing better code and teaching you things you may not have known along the way. You still have to know how to code to produce a good result when using AI. It doesn't always give you the best advice or result and you have to be able to recognize when that happens.
Not sure who the big losers will be since I don't really know a lot how AI is used outside of my sphere of expertise. I guess it could be kind of like robotics. It replaced a lot of jobs but also created new ones at the same time.
StreetsofBeige
Member
I must had missed all the news about this scamming company. Or perhaps it all flew under the radar on business sites and I never saw it if it wasnt front page news or related to one of my stocks.If you want a quick history of the company this is a better page:
From $1.3B to Bust: The Rise and Collapse of Builder.ai
The Leading Platform AI Agents and Agentic Process Automation. Build AI, Integrate, Scale, all on One Platform.
Ironically AI
StreetsofBeige
Member
There must be some jobs that can be crushed by AI, hence all the people fired or shitting bricks. A lot of creatives claim they dont like or use AI when doing art and they are afraid AI bot art will replace them.
In finance, ERP systems with IT guys doing all the coding for databases and such only goes so far (implementation and fixing errors). You still need people to analyze the data meticulously since the data can be garbage or wrong. You need experienced people to be able to just eyeball data and know it's wrong. People are always amazed when I can just glance a spreadsheet and know whats wrong if a formula looks off or do calculations in my head to a decimal place. And no AI tool will ever completely replace anyone analyzing or making decisions off it because no matter how many slick pie charts and automated AI recos there are, it will never know all intrinsic data, knowledge or business strategies which can be counter intuitive to a normal answer since there's so many variables.
Hot5pur
Member
I only know the free LLMs, and they are OK for some things, like getting bits of code or looking something up, but I would not trust them with any real work as they hallucinate too often and don't "understand" what they are doing. It's like expecting a linear regression to understand it's a linear regression. LLMs are basically higher dimensional curve fitting. They are damn impressive though.
My bigger concern is WTF is the tech industry working on? We had mining Bitcoin, NFTs, self driving cars (how about fixing rush hour traffic first), social media apps that brainwash you into spending more time on them, connecting your microwave to the internet, massive data surveillance so you can get advertised to better ..... And now LLMs so they can replace lawyers, accountants programmers and people in call centers...
Like wtf are we training PhDs for in IT/Comp Sci? So they can make a bunch of cunts more money and put the rest of us out of work? The spirit of innovation for human progress has never felt more depressing (at least in IT).
My bigger concern is WTF is the tech industry working on? We had mining Bitcoin, NFTs, self driving cars (how about fixing rush hour traffic first), social media apps that brainwash you into spending more time on them, connecting your microwave to the internet, massive data surveillance so you can get advertised to better ..... And now LLMs so they can replace lawyers, accountants programmers and people in call centers...
Like wtf are we training PhDs for in IT/Comp Sci? So they can make a bunch of cunts more money and put the rest of us out of work? The spirit of innovation for human progress has never felt more depressing (at least in IT).
nemiroff
Member
Nah, I disagree. I think "AI" is great when you're smart enough to use it to your benefit. The "awful" usually come from using it wrong, or expecting it to be magic when it's just a tool.
F.ex., if people saw the results of an advanced research prompt in comparison to a simple "tell me about..."-prompt I'm pretty sure many would have parts of their brains blown.
Last edited:
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
I'm not saying it's the truth. Printing money like that helps when recession would normally hit.
Capitalism baby, if it isn't sought-after by the rich to save or make more money it's not that important. Innovation that reduces cost or increases sales wins every time.
StreetsofBeige
Member
Pretty much.
No different than farming. Long time ago, it was a family of farmers busting their backs growing crops in their shitty field where every family member joined in (even 10 year old Bobby). Then they had hired hands. Then animal power. Then machines. Then GMO seeds and whatever chemicals they spray on the stuff.
Add it up and now that same number of people can probably get 10x the yield across more acres with the same human manpower after 50 years of progress. I made up the 10x number for sake of argument, but you get the idea.
TKOFromTokyo
Member
AI is real. It's not great, but I work for a large public company that is more conservative with spending and even we are spending a lot on it and have some big future plans with it. I imagine most large companies are going to spend a lot on AI.
SomeNorseGuy
Gold Member
I did read a MIT article about the energy footprint and water usage of AI from earlier this year.
Explained: Generative AI’s environmental impact
MIT News explores the environmental and sustainability implications of generative AI technologies and applications.
news.mit.edu
More focused on energy then water, but still an interesting read.
nowhat
Member
Study finds AI tools made open source software developers 19 percent slower
Coders spent more time prompting and reviewing AI generations than they saved on coding.
arstechnica.com
poppabk
Cheeks Spread for Digital Only Future
Ironically they were mainly using actual people and pretending it was AI doing it.If you want a quick history of the company this is a better page:
From $1.3B to Bust: The Rise and Collapse of Builder.ai
The Leading Platform AI Agents and Agentic Process Automation. Build AI, Integrate, Scale, all on One Platform.
Ironically AI
Codeblew
Member
Interesting but definitely not my experience. I am not sure how sound that study is. For instance, I wouldn't use AI to refactor my code except for maybe asking it to split a large function into smaller functions to reduce the cognitive complexity of my code. If I am just renaming something, why would I use AI for that. My IDE is great at doing that sort of thing.Study finds AI tools made open source software developers 19 percent slower
Coders spent more time prompting and reviewing AI generations than they saved on coding.
One of the most productive things I use an LLM for is writing test cases for my classes or packages. I can write a bunch of code and ask the AI to write test cases for them. It spits out 500 lines of test code in a few seconds. They aren't always 100% correct the first time. Easy, just paste the error into chat and it fixes its mistake. I would write a lot less test code if AI didn't make it so easy.
Another thing that I like to do is once that I have finished up a section of code, I ask the AI something like "As a senior software engineer, do you see any issues with my code or do you have any recommendations for improvement?" This usually helps me find issues that I had not thought of or I get some great suggestions to improve my code. I don't accept all of the suggestions of course but it is still very helpful.
Last edited:
Durin
Member
What sucks is I never had enough disposable income to invest in companies, because it was easy money for a while, and I'd probably be pulling out now because I think by next year we'll see some bleeding happen.
There is more utility to AI than crypto or NFTs that's already proven itself, like AI search is regularly better than the trashfire google has become, and over time we'll see better integrations in software to add real value or lower the barrier to entry further...but tons of companies are gonna fail.
The snake-oil is thinking AI can wholesale replace tons of jobs, when using AI it's pretty shit on its own unless the tasks are very simple, and for everything else you need a human to guide it or set the foundational structure before you let it help.
There is more utility to AI than crypto or NFTs that's already proven itself, like AI search is regularly better than the trashfire google has become, and over time we'll see better integrations in software to add real value or lower the barrier to entry further...but tons of companies are gonna fail.
The snake-oil is thinking AI can wholesale replace tons of jobs, when using AI it's pretty shit on its own unless the tasks are very simple, and for everything else you need a human to guide it or set the foundational structure before you let it help.
Last edited:
FoxMcChief
Gold Member
TerribleAtThisGame
Member
I so wanted to pretend to be smart and actually learn something from some of these replies but alas...
OuterLimits
Member
I would be more concerned about the rapidly rising energy demand AI will require. Meanwhile many Governments want to primarily power their grid with renewable energy. Plus want to transition to mainly electric vehicles.
Not going to work. Fossil fuels or nuclear power will need to be a major player for the short and medium term. Pick one.
Not going to work. Fossil fuels or nuclear power will need to be a major player for the short and medium term. Pick one.
Last edited:
Haint
Member
This shit isn't even true, the "tech stock market" (i.e. the Nasdaq) at the dot.com peak had a P/E over 100 with some estimates placing it as high as 200. It's currently somewhere in the 30's, forward P/E is mid-20's.
Most of the tech companies during the "dot.com" crash had no actual businesses and LOST mountains of money every quarter, buried in debt up to their ears. Today most of the core AI companies MAKE mountains of money every quarter and have $100's of billions cash on hand.
These historical bar charts only make stocks seem inflated cause the dollars/currencies that buy the shares have depreciated massively in an incredibly short time frame cause world governments printed 10's of trillions of dollars out of thin air during Covid. Homes literally cost double (2x) what they did 4 or 5 year ago, everything else is up 50-75%. Share prices have naturally followed suite. These aren't really 3 and 4 trillion dollar companies, they're 1.5 - 2 trillion dollar companies. The reality is your $1's buying power has depreciated to 50 cents in like 2 or 3 years.
Most of the tech companies during the "dot.com" crash had no actual businesses and LOST mountains of money every quarter, buried in debt up to their ears. Today most of the core AI companies MAKE mountains of money every quarter and have $100's of billions cash on hand.
These historical bar charts only make stocks seem inflated cause the dollars/currencies that buy the shares have depreciated massively in an incredibly short time frame cause world governments printed 10's of trillions of dollars out of thin air during Covid. Homes literally cost double (2x) what they did 4 or 5 year ago, everything else is up 50-75%. Share prices have naturally followed suite. These aren't really 3 and 4 trillion dollar companies, they're 1.5 - 2 trillion dollar companies. The reality is your $1's buying power has depreciated to 50 cents in like 2 or 3 years.
Last edited:
seething left
Member
As much as I dislike the AI taking over so many aspects of our lives (from personal as well as from enterprise perspective), it's here to stay. Some things cannot be reverted. And it will only get worse from now on. And since the most powerful and capable AIs will always require huge datacenters to run, datacenters that only big tech can afford to build and operate, it's a life long lock-in for the majority of companies (and individuals). The race to the AI market dominance will swallow many more billions in the upcoming years. And all the losses will make it into the product and service margins and be passed through to the customers. Terrible times we're living in.