AI - The Good The Bad The Ugly

[ikeo1]

Guess the almond farmers will have to give up some of that watuh.

Older facilities rely on open evaporative cooling towers that evaporate millions of gallons, but modern AI data centers increasingly use closed-loop liquid cooling (similar to a car radiator) or air-based systems. Once filled, closed-loop systems recirculate fluid inside sealed pipes rather than continuously evaporating water.

Those aren't in existence yet. You're speaking to a future fix for a current problem. The existing datacenters are maxxed out and hundreds weren't built with this but still adding more power.

Early in my career, I used to build out Datacenters cages for startups... Existing capacity is causing the consumption and also places around the world are not building to that standard.

 

[phred]

I am afraid we are going to raise a generation that can't think for themselves.
I am also concerned about all of the AI content on YouTube and other platforms that may be rewriting history and making fake images. It becomes difficult to tell reality from AI made up content.
I think content that is generated by AI should have clear and mandatory disclosure.

 

[RustVendor]

The water concern is pretty bogus since they use closed loop systems

Not all of them do, but they are moving towards it. In California there are a lot more restrictions than in other states so we don't have nearly as much of an issue with data centers as out of state. there's a ton of data centers in San Jose but it's built for it so they use recycled water, but there's one they are trying to build in Mojave that has no water source and they are trying to sue in order to get access to water.

 

[Shellbackitup]

Guess the almond farmers will have to give up some of that watuh.

Older facilities rely on open evaporative cooling towers that evaporate millions of gallons, but modern AI data centers increasingly use closed-loop liquid cooling (similar to a car radiator) or air-based systems. Once filled, closed-loop systems recirculate fluid inside sealed pipes rather than continuously evaporating water.

Those closed loop pipes have to be cooled, otherwise they would just keep heating up. Air cooling isn't enough for datacenters.

The 'closed loop' means that the main coolant is not discharged, but it doesn't mean that water isn't consumed by the cooling cycle.

No one in AI is using closed loop secondary systems, only closed loop primary systems. It's overkill for their needs, and unlike nuclear power, discharge of the primary coolant to atmosphere isn't a huge issue. so they don't need the secondary system for containment.

You're right about the farmers.

Before the almond farmers in the central valley, there were the cows. Both extremely water consumption heavy.

 

[Gawernator]

My take is that AI is overly hated. Take your graphic for example OP. In order to make graphics in the past, you either had to learn how to use programs to make them or pay someone. I'm not going to do any of those things if I just want to make a public service announcement graphic like the one you posted in the other thread about not getting scammed.

There are some genuine downsides like it can make people lazy, or maybe Terminators, but otherwise a lot of the hate is from people with no job skills that can be easily replaced

 

[TheNotoriousKIM]

I'm in the commercial real estate industry, predominantly ground-up multi-family development but also do brokerage, property & asset management, and general construction. I have a masters degree from one of the top universities for my field (not bragging but for credibility sake).

My industry is in the stone ages right now and are close to a "dot com boom" type of evolution with AI introduced. I've used AI throughout grad school and it saved hundreds of hours for compiling and summarizing information. In Excel, it saves hundreds of hours and time spent modeling and auditing formuals/formats.

To cut to the chase, I've had it create models from scratch that would have taken me 4 months in 45 mins. For CRE nerds, this includes things that Argus would be needed for in institutional RE. I see AI as a productivity catalyst more than a replacement at this time. You still need to know how to do things to audit and correct mistakes where you see them. It does take off A LOT of heavy lifting if you know how to use it. I use it to create investment memos and business plans/decks from PPT to full blown development/investment models with waterfalls, prefs, hurdles, etc...

 

[HarleyLowRiderST]

It really does not matter what we, or you, or anyone thinks, AI is here, it's gaining power and there will be no turning back. Like may technology advancements there will be casualties.

• The Automobile killed the horse and buggy and those people had to find other jobs.
• Industrial Farming killed the Mom & Pop family farm and those people had to find other jobs.
• Overseas manufacturing killed blue collar jobs in America and those people had to find other jobs.
• TV killed the radio industry.
• The internet killed News Papers and Magazines.
• Uber and Lyft killed the Taxi companies
• Self driving cars and trucks will kill Uber and Lyft and most trucking companies
• AI will eliminate Actors, writers, and most of Hollywood.
• AI will kill many jobs for sure, but also improve efficiencies on the jobs that remain. In theory those cost savings should be passed down to the consumer.
• The list is endless . . .

The real issue with AI is not the jobs that will be eliminated as that has happened since the dawn of technology. The real issue will be the flow of information that the average person (you, me, everyone) can not delineate between fact or fiction. You see that already in Social Media. 90% of the people on IG are fake. Posts on X and other platforms, fake. That's the scary part is will be be able to discern the truth from fiction/lies in the next 5 years.

 

[Gawernator]

It really does not matter what we, or you, or anyone thinks, AI is here, it's gaining power and there will be no turning back. Like may technology advancements there will be casualties.

• The Automobile killed the horse and buggy and those people had to find other jobs.
• TV killed the radio industry.
• The internet killed News Papers and Magazines.
• Industrial Farming killed the Mom & Pop family farm and those people had to find other jobs.
• Overseas manufacturing killed blue collar jobs in America and those people had to find other jobs.
• Uber and Lyft killed the Taxi companies
• AI will eliminate Actors, writers, and most of Hollywood.
• Self driving cars and trucks will kill Uber and Lyft and most trucking companies
• AI will kill many jobs for sure, but also improve efficiencies on the jobs that remain. In theory those cost savings should be passed down to the consumer.
• The list is endless . . .

The real issue with AI is not the jobs that will be eliminated as that has happened since the dawn of technology. The real issue will be the flow of information that the average person (you, me, everyone) can not delineate between fact or fiction. You see that already in Social Media. 90% of the people on IG are fake. Posts on X and other platforms, fake. That's the scary part is will be be able to discern the truth from fiction/lies in the next 5 years.

100%. Luddites will kick, scream and cry but they are powerless to stop the advance of technology. The smart man will learn to be an early adopter.

 

[Shellbackitup]

It really does not matter what we, or you, or anyone thinks, AI is here, it's gaining power and there will be no turning back. Like may technology advancements there will be casualties.

• The Automobile killed the horse and buggy and those people had to find other jobs.
• TV killed the radio industry and the internet killed it again.
• Industrial Farming killed the Mom & Pop family farm and those people had to find other jobs.
• Overseas manufacturing killed blue collar jobs in America and those people had to find other jobs.
• Uber and Lyft killed the Taxi companies
• AI will eliminate Actors, writers, and most of Hollywood.
• Self driving cars and trucks will kill Uber and Lyft and most trucking companies
• AI will kill many jobs for sure, but also improve efficiencies on the jobs that remain. In theory those cost savings should be passed down to the consumer.
• The list is endless . . .

The real issue with AI is not the jobs that will be eliminated as that has happened since the dawn of technology. The real issue will be the flow of information that the average person (you, me, everyone) can not delineate between fact or fiction. You see that already in Social Media. 90% of the people on IG are fake. Posts on X and other platforms, fake. That's the scary part is will be be able to discern the truth from fiction/lies in the next 5 years.

I agree with some of what you put, but I would argue there is nothing that says we can't turn back.

• The Automobile killed the horse and buggy and those people had to find other jobs.
  • True. However, the equestrian industry is worth hundreds of billions in the US alone. Not a dead industry, just smaller and different.
• TV killed the radio industry and the internet killed it again.
  • I still listen to the radio, both over the air and satellite. While not as big as before, the industry isn't dead
• Industrial Farming killed the Mom & Pop family farm and those people had to find other jobs.
  • I will willingly pay waaay more for produce from small Mom and Pop farms than I would ever pay in a store. Farmer's Markets are in every city.
• Overseas manufacturing killed blue collar jobs in America and those people had to find other jobs.
  • Onshoring is happening in some industries, especially after COVID exposed the risks of a global supply chain, and tariffs have further encouraged onshoring.
• Uber and Lyft killed the Taxi companies
  • I choose to use taxis at the airport every time over Uber/Lyft, with zero wait and the same price. Are there fewer? Sure, but again, not a dead industry
• AI will eliminate Actors, writers, and most of Hollywood.
  • I would rather support human actors and writers, and won't watch AI generated content. I'm not alone.
• Self driving cars and trucks will kill Uber and Lyft and most trucking companies
  • TBD
• AI will kill many jobs for sure, but also improve efficiencies on the jobs that remain. In theory those cost savings should be passed down to the consumer.
• The list is endless . . .

None of these things are INEVITABLE.

Consumers will choose the path they want.

Cheaper options will always have a huge following

But many people choose based on their morals over cost.

 

[ikeo1]

There are some genuine downsides like it can make people lazy, or maybe Terminators, but otherwise a lot of the hate is from people with no job skills that can be easily replaced

My datacenter job was replaced by cloud... My cloud position was replaced by people in Poland where I had to train them. My gig after that I had to train people in Hyperbad, and then got laid off. This is no different in my line of work. I'm constantly training my replacement then being let go.

We build until things are big and stable, then because we started early and not core enough we get replaced. Reality is we can always be easily replaced, some people just don't like to accept that from a job perspective. After 2 layoffs, I decided to do things more in the real world with guns, prepping, etc. I grow my own fruits and veggies too. I've worked in tech and WFH for 15 years and feel like the disconnect is more than ever even though there's more connection.. This is one of the core reasons, why I try to treat the site like a hybrid and hope to see you guys in person on/at the range.

I look at AI as something to make life easier by learning how to leverage it but I end up working on things more but think its a ebb and flow. Big burst of productivity, before reality sets in and the burn out happens.


I feel for the younger generation too since their minds are moving at exceptional speed. Kinda reminds me of NAS from Johnny Mnemonic

 

[Darqhelmet]

I think it’s overall bad to very very bad.

I don’t have any deep usage of it, but Im never impressed with the results. On very simple things googles AI fails multiple times during searches, openai has outright lied to me, and I haven’t touched claud as I don’t care or code. SpaceX ai oof, just look at the results.

As it sits now. it has destroyed consumer electronics, we have unknown amounts of pension funds tide into it, the noise pollution is insane, the grid doesn’t support the load, SpaceX is poisoning a poor neighborhood with running gas turbines 24/7, the economy is going to burn when it goes up due to the massive amount of circular financing.

I can keep going, but to be honest, AI the tech stack pretty cool. AI the story and what we are being forced to deal with, absolute and should go away.

 

[Gawernator]

I agree with some of what you put, but I would argue there is nothing that says we can't turn back.

• The Automobile killed the horse and buggy and those people had to find other jobs.
  • True. However, the equestrian industry is worth hundreds of billions in the US alone. Not a dead industry, just smaller and different.
• TV killed the radio industry and the internet killed it again.
  • I still listen to the radio, both over the air and satellite. While not as big as before, the industry isn't dead
• Industrial Farming killed the Mom & Pop family farm and those people had to find other jobs.
  • I will willingly pay waaay more for produce from small Mom and Pop farms than I would ever pay in a store. Farmer's Markets are in every city.
• Overseas manufacturing killed blue collar jobs in America and those people had to find other jobs.
  • Onshoring is happening in some industries, especially after COVID exposed the risks of a global supply chain, and tariffs have further encouraged onshoring.
• Uber and Lyft killed the Taxi companies
  • I choose to use taxis at the airport every time over Uber/Lyft, with zero wait and the same price. Are there fewer? Sure, but again, not a dead industry
• AI will eliminate Actors, writers, and most of Hollywood.
  • I would rather support human actors and writers, and won't watch AI generated content. I'm not alone.
• Self driving cars and trucks will kill Uber and Lyft and most trucking companies
  • TBD
• AI will kill many jobs for sure, but also improve efficiencies on the jobs that remain. In theory those cost savings should be passed down to the consumer.
• The list is endless . . .

None of these things are INEVITABLE.

Consumers will choose the path they want.

Cheaper options will always have a huge following

But many people choose based on their morals over cost.

Not getting into how much of a scam farmer's markets often are lol, that's from your perspective coming from a much wealthier generation. Modern consumer is insanely price sensitive. People crave their slop. AI definitely isn't replacing real actors and athletes though but we are already seeing very popular AI music

 

[chiroken007]

The water concern is pretty bogus since they use closed loop systems

That may be true for some of the systems but not all. It depends on the environment. Some systems consume 70% as evap into the air.

 

[cali_jazz]

Here's what ChatGPT says:
I'd vote “Both – I hate it but I use it.” Although I'd change the wording to “Both – incredibly useful, but deeply flawed.”

Here's why.

AI can genuinely deliver 10x productivity in certain tasks: research, writing, programming, troubleshooting, and analyzing complicated information. That's not just hype.

But the problems are substantial:

• Confidently wrong answers. AI can present incorrect information with the same confidence as correct information.
• Sycophancy. It can agree with users instead of independently evaluating evidence.
• Unreliable consistency. It can correctly analyze something one minute and contradict itself the next.
• Overreliance. People can mistake fluent answers for verified expertise.
• Broader consequences. Privacy, misinformation, displacement of workers, and concentration of power deserve serious attention.

Our recent exchange about the LOK grips was a particularly good example of the first three. I repeatedly misidentified something, agreed with contradictory statements, and wasted your time.

But those failures don't erase the technology's considerable usefulness.

And I'd argue the biggest danger isn't that AI makes mistakes. It's that AI can be remarkably convincing when it's wrong.

Having read the discussion, I'd still vote Both, but the context changes how I'd explain that vote.

Ikeo1 isn't simply asking whether AI is useful. He's trying to understand why people are increasingly hostile toward it, particularly after his AI-generated graphic was downvoted on Reddit.

A few things stand out to me.

ButteFlyFisher makes one of the strongest technical arguments. AI accelerates work when someone already understands the subject, but can waste enormous amounts of time when it confidently heads in the wrong direction. That distinction is important.

Davethebull raises the issue of people surrendering their judgment. I agree that's a serious risk, although I'd add that even people who question AI's answers can be misled when they lack the expertise to recognize an error.

Ikeo1's experience is particularly interesting. He's personally experienced technological displacement and outsourcing, yet still sees AI as an opportunity to build things he couldn't previously build himself. His enthusiasm isn't necessarily naïve; it's informed by both sides of technological change.

And the discussion reveals something else: people are conflating several different objections to AI—unreliable answers, low-quality generated content, job displacement, environmental costs, and loss of human creativity. Those concerns deserve separate consideration rather than one blanket judgment.

I also think some participants are too dismissive of the opposition. Calling critics Luddites or suggesting they lack valuable job skills doesn't address their legitimate concerns.

My position remains that AI is an extraordinarily powerful technology whose benefits are real, but whose limitations and social costs are frequently understated by its advocates.

The most interesting question for ikeo1 isn't whether AI is good or bad. It's why people who recognize its usefulness increasingly resent how it's being deployed.

That distinction could make for a much more productive discussion than the poll choices suggest.

 

[cali_jazz]

Here's what ChatGPT says:

In addition to pure knowledge and information seeking, I use AI like a sounding board, I like to hear different opinions and perspectives to help inform my own. I tell AI not to just agree with me or change it's answers to what it thinks I want to hear, I like to challenge it to clarify and solidify my own thoughts.

 

[ikeo1]

The most interesting question for ikeo1 isn't whether AI is good or bad. It's why people who recognize its usefulness increasingly resent how it's being deployed.

This would be a separate question

 

[RustVendor]

At least FedEx is honest that their AI isn't any good:

 

[Darqhelmet]

Here's what ChatGPT says:
I'd vote “Both – I hate it but I use it.” Although I'd change the wording to “Both – incredibly useful, but deeply flawed.”

Here's why.

AI can genuinely deliver 10x productivity in certain tasks: research, writing, programming, troubleshooting, and analyzing complicated information. That's not just hype.

But the problems are substantial:

• Confidently wrong answers. AI can present incorrect information with the same confidence as correct information.
• Sycophancy. It can agree with users instead of independently evaluating evidence.
• Unreliable consistency. It can correctly analyze something one minute and contradict itself the next.
• Overreliance. People can mistake fluent answers for verified expertise.
• Broader consequences. Privacy, misinformation, displacement of workers, and concentration of power deserve serious attention.

Our recent exchange about the LOK grips was a particularly good example of the first three. I repeatedly misidentified something, agreed with contradictory statements, and wasted your time.

But those failures don't erase the technology's considerable usefulness.

And I'd argue the biggest danger isn't that AI makes mistakes. It's that AI can be remarkably convincing when it's wrong.

Having read the discussion, I'd still vote Both, but the context changes how I'd explain that vote.

Ikeo1 isn't simply asking whether AI is useful. He's trying to understand why people are increasingly hostile toward it, particularly after his AI-generated graphic was downvoted on Reddit.

A few things stand out to me.

ButteFlyFisher makes one of the strongest technical arguments. AI accelerates work when someone already understands the subject, but can waste enormous amounts of time when it confidently heads in the wrong direction. That distinction is important.

Davethebull raises the issue of people surrendering their judgment. I agree that's a serious risk, although I'd add that even people who question AI's answers can be misled when they lack the expertise to recognize an error.

Ikeo1's experience is particularly interesting. He's personally experienced technological displacement and outsourcing, yet still sees AI as an opportunity to build things he couldn't previously build himself. His enthusiasm isn't necessarily naïve; it's informed by both sides of technological change.

And the discussion reveals something else: people are conflating several different objections to AI—unreliable answers, low-quality generated content, job displacement, environmental costs, and loss of human creativity. Those concerns deserve separate consideration rather than one blanket judgment.

I also think some participants are too dismissive of the opposition. Calling critics Luddites or suggesting they lack valuable job skills doesn't address their legitimate concerns.

My position remains that AI is an extraordinarily powerful technology whose benefits are real, but whose limitations and social costs are frequently understated by its advocates.

The most interesting question for ikeo1 isn't whether AI is good or bad. It's why people who recognize its usefulness increasingly resent how it's being deployed.

That distinction could make for a much more productive discussion than the poll choices suggest.

The machine knows it sucks.

Also:

OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws

In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits.

www.computerworld.com

It is flawed at its core.

 

[Darqhelmet]

In addition to pure knowledge seeking, I use AI like a sounding board, I like to hear different opinions and perspectives to help inform my own. I tell AI not to just agree with me or change it's answers to what it thinks I want to hear, I like to challenge it to clarify and solidify my own thoughts.

It’s a vector data base. It does not think, it can not think. That is part of the huge problem I have with this . The companies making it respond like it’s talking to frame it as something other than a vector data base that picks the next most likely word.

 

[cali_jazz]

The machine knows it sucks.

Also:

OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws

In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits.

www.computerworld.com

It is flawed at its core.

At least it's honest about it's faults. When it starts claiming it's perfection and supremacy that's when I'll really be scared!