AI Good and Bad
- Thread starter bigredfish
- Start date
They're re not even trying to hide it now. Target the weakest link, those 30% that literally need an instruction manual on how to walk.
All we need now is for Trump to tell the faithful that he’ll send them $10,000 if they convert to Trump coins
whoami ™
Getting comfortable
Maybe I'm miss reading the last few, and only few posts I've read in this thread, but I got the impression AI topped and that they've over promised and underdelivered? Or is this just a thread where any info a poster finds interesting regarding AI gets posted? I'm one of the 20M people (0.02%) that see's first-hand that large, complex projects that used to take weeks/months can now be completed by frontier agents with a prompt. The latest models that came out in the last 30days are crazy impressive. I don't see things slowing down anytime soon. I can build out whole projects as fast than I can come up with ideas and my skill level in coding was elementary. Seems like I need a new tablet or bigger computer/server every week for some thing Im working on, and Im not even a professional. Now, if open source software doesn't do what i want it to I just fork it write a pull request and run my own patched version. Closed source is the same, I just haven't hooked my harness up to Ghidra yet and got in to decompiling. As good and as cheap as models are right now... copying, translating, decompiling software from binaries is the going to be next big thing.
I literally just stopped by here just to see how far the average user has advanced with AI since this was one of the few forums I use to drop by and saw this post and it seemed to be the opposite of what I was expecting to find so i bit.
I literally just stopped by here just to see how far the average user has advanced with AI since this was one of the few forums I use to drop by and saw this post and it seemed to be the opposite of what I was expecting to find so i bit.
Its a general thread.
I dont see anything saying AI isnt a breakthrough technology.
I dont personally use it.
The problem is the CAPEX numbers $$, and timeframe to profitability, and datacenter costs, and electricity and water, and Chinese open source, and tokens... I'll stop
Maybe ask JP Morgan and Morgan Stanley to explain it?
I dont see anything saying AI isnt a breakthrough technology.
I dont personally use it.
The problem is the CAPEX numbers $$, and timeframe to profitability, and datacenter costs, and electricity and water, and Chinese open source, and tokens... I'll stop
Maybe ask JP Morgan and Morgan Stanley to explain it?
Last edited:
whoami ™
Getting comfortable
Fair enough, I think we were talking past each other. I was arguing that it works, you're arguing whether it pays for itself, and those aren't the same question.
Also my number was off. It's 0.2%, not 0.02%, and I should have said where it came from. I was paraphrasing a post that breaks AI users into tiers. Roughly 75% of people have barely touched it, maybe 20% use the free chatbots like a better Google, around 0.2% use frontier models for code and math every day, and a few thousand people inside the labs see what's coming next. You and I are looking at it from opposite ends of that list, so it makes sense we see different things.
The money question is real though, and honestly my own usage makes your point. I pay [$100/mo for Claude Max] and if a cheaper model or an open source one I could run locally did the same job tomorrow I'd switch without thinking about it. The 0.2% alone aren't going to pay for all those datacenters. For the capex to make sense, the 20% have to start paying or agents have to replace enough paid work to cover it, and I have no idea how long that takes. I don't think anyone does, JP Morgan included.
Also my number was off. It's 0.2%, not 0.02%, and I should have said where it came from. I was paraphrasing a post that breaks AI users into tiers. Roughly 75% of people have barely touched it, maybe 20% use the free chatbots like a better Google, around 0.2% use frontier models for code and math every day, and a few thousand people inside the labs see what's coming next. You and I are looking at it from opposite ends of that list, so it makes sense we see different things.
The money question is real though, and honestly my own usage makes your point. I pay [$100/mo for Claude Max] and if a cheaper model or an open source one I could run locally did the same job tomorrow I'd switch without thinking about it. The 0.2% alone aren't going to pay for all those datacenters. For the capex to make sense, the 20% have to start paying or agents have to replace enough paid work to cover it, and I have no idea how long that takes. I don't think anyone does, JP Morgan included.
mat200
IPCT Contributor
- Jan 17, 2017
- 19,697
- 32,849
Recall the Rail Road bubbles .. Rail Roads worked, yet mania to invest in them did create a bubble, AI imho is a crazy hyper version of that ..
( and yes AI in it's various forms is very impressive .. yet really does require significant attention to get it precise )
The railroad bubble refers to periods of intense speculative frenzy and subsequent crashes in the railroad industry, primarily in the United Kingdom during the 1840s (Railway Mania) and in the United States during the late 19th century (notably the panics of 1873 and 1893). These bubbles were driven by overinvestment in infrastructure, excessive leverage, and fraudulent schemes, leading to massive bankruptcies and economic depressions.
( and yes AI in it's various forms is very impressive .. yet really does require significant attention to get it precise )
The railroad bubble refers to periods of intense speculative frenzy and subsequent crashes in the railroad industry, primarily in the United Kingdom during the 1840s (Railway Mania) and in the United States during the late 19th century (notably the panics of 1873 and 1893). These bubbles were driven by overinvestment in infrastructure, excessive leverage, and fraudulent schemes, leading to massive bankruptcies and economic depressions.
United Kingdom: Railway Mania (1840s)
- Peak and Causes: In 1845–1846, speculative investment surged to 7% of UK GDP, fueled by low interest rates, aggressive advertising, and parliamentary approval of thousands of railway bills.
- The Crash: Rising interest rates and recognition of unprofitable routes caused share prices to plummet by roughly 50% between 1846 and 1850.
- Outcome: Despite the financial ruin of many investors, the bubble resulted in a net expansion of the British railway system, with many failed lines later becoming viable under larger, consolidated companies.
United States: Late 19th Century Bubbles
- Panic of 1873: Triggered by the bankruptcy of Jay Cooke & Co. (financier of the Northern Pacific Railway), this crisis wiped out 121 railroads and 18,000 businesses, causing unemployment to reach 14% by 1876.
- Panic of 1893: Caused by railroad overbuilding and the bankruptcy of the Philadelphia and Reading Railroad, this panic led to the failure of over 15,000 companies and 500 banks, with unemployment peaking at 25% in some states.
- Long-term Impact: While the busts caused severe short-term misery, the surviving railroad infrastructure became the backbone of American industrial growth, enabling the rise of national brands and efficient logistics.
Comparison to Modern Bubbles
Historians and economists frequently compare these railroad bubbles to the AI bubble and the 1990s Dot-com crash. Key similarities include excessive infrastructure spending, investor euphoria around transformative technology, and speculative excess. However, unlike the railroad era, which was heavily debt-financed, the current AI buildout is largely funded by corporate free cash flow.
Honest answer. Appreciated.
I’m only on coffee #2 and didn’t finish college so it’ll take me until later today to type up my broader perspective, but between our two posts, I think we’re hitting on 2 of the 3 big questions.
Question #3 is a lot more complicated and I’m not smart enough to figure it out.
Preview:
Spent a week in Hatteras on the OBX a couple of months ago.
Met a very interesting local old timer (85ish) who lives in the house his grandfather built on the island. Saw me pecking away on the laptop one morning over coffee at the docks and after exchanging pleasantries, sat down to chat.
This became a routine each day over the course of the week and I was surprised at his knowledge and understanding of larger worldly issues like geo-politics, finance, and most surprising, technology. Fascinating man, made me feel dumb.
Leaving out for now the impression he made on me and his stories, (owning the premier charter business on Hatteras and handed down through generations), one of our chats ended up at AI
Again for brevity, he posed a great question. (Two really)
If AI is going to replace:
Teachers
Engineers
Doctors
Call center employees
Lawyers
Programmers
Anyone using/analyzing/working with data
Anyone in finance
Etc etc etc….
1) What will we all do?
Example: If you previously aspired to be a great engineer and design the worlds greatest bridge/building/car and there’s no longer a need, what happens to your aspirations?
What is it that drives you to get out of bed?
2) Money.
If we are all replaced and become dependent on UBI, who pays the $5 Trillion in taxes to keep it all afloat and hand out the money?
Again this is a simplistic description of the issues, but it stuck with me