Im not worried about the technology, it's the funding that's the issue.

ANOTHER Force Majeur incoming.

Exelon refuses to build power to massive hyperscaler data center being built in Joliet for undisclosed hyperscaler, after the developer failed to come up with deposit for the $20 billion energy project.

Developer could only pay $1.$ORCL $META…

— Roger (@rdd147) September 24, 2026

Quote:

---

A 1.8-GW, $20 billion data center planned by PowerHouse Hillwood was underpinned by a $1 letter of credit "less than the price of a cup of coffee"'

The Federal Energy Regulatory Commission on Tuesday rejected Commonwealth Edison's "notice of cancellation" of a transmission security agreement, or TSA, for a 1.8-GW, $20-billion data center PowerHouse Hillwood Holding is developing in Joliet, Illinois.

In part, the contract dispute centers on the TSA's credit support requirements. PowerHouse Hillwood contends it met the agreement's initial credit requirements via a $1 posting, according to FERC's decision


The idea that $1 may provide appropriate security to any such agreement strikes me as an embarrassing legal fiction: insulting to the underlying ratepayers, stakeholders, and the grid itself that bear the real risk of this project," LaCerte said. "Treating that risk as collateralizable for less than the price of a cup of coffee to me trivializes the very obligations that such a guarantee purports to secure."

---

https://www.utilitydive.com/news/ferc-exelon-comed-powerhouse-hillwood-data-center/831127/

And I don't think enough people understand how much Oracle is screwed if OpenAI can't pay their bills. Oracle has hundreds of billions of dollars in contracts with OpenAI to build out these data centers, which will supposedly be paid by future revenue. The problem is that revenue will likely not materialize. From a business perspective OpenAI is a coffin on roller skates.

There's a 90 plus percent probability this ends very badly for Oracle when the bubble bursts.

This is a great video to listen to while you are out in your car or mowing the lawn or whatever. This is a professor of finance at NYU who is doing a lot of math on the AI story, he's on this podcast talking about the serious financial issues these companies are facing. The amount of revenue the frontier companies will need to be able to pay their bills in the next few years is staggering and there's no indication they will ever get there.

These data centers are a huge misallocation of capital and the fewer that get built right now is probably a good thing.

ErnestEndeavor said:

---

This is a great video to listen to while you are out in your car or mowing the lawn or whatever. This is a professor of finance at NYU who is doing a lot of math on the AI story, he's on this podcast talking about the serious financial issues these companies are facing. The amount of revenue the frontier companies will need to be able to pay their bills in the next few years is staggering and there's no indication they will ever get there.

These data centers are a huge misallocation of capital and the fewer that get built right now is probably a good thing.

---

It's fine....everything is fine

ErnestEndeavor said:

---

This is a great video to listen to while you are out in your car or mowing the lawn or whatever. This is a professor of finance at NYU who is doing a lot of math on the AI story, he's on this podcast talking about the serious financial issues these companies are facing. The amount of revenue the frontier companies will need to be able to pay their bills in the next few years is staggering and there's no indication they will ever get there.

These data centers are a huge misallocation of capital and the fewer that get built right now is probably a good thing.

---

So is AI the next bubble many projected it would be about a year ago?

Jetpilot86 said:

---

I'm not too worried about AI. It can't even do excel calendar math correctly.

Asked it the number of days between 6/30 and 9/7. The answer is 69.

---

You used an extremely rudimentary AI on Brave. It's nothing more than junk.

As a follow up to the stupid date response from Grok:

Brave does not use one model. It uses different stacks for search answers versus the in-browser assistant.
Brave Search "Quick Answer" / Ask Brave (the UI in your earlier screenshot) is not ChatGPT or Gemini. It runs open-weight models that Brave hosts itself, grounded against Brave's own search index. As of 2026 that search chat is commonly Qwen3. Earlier versions used Mixtral 8x7B and Mistral 7B, then a rotating Llama + Qwen mix. Brave's search team has said they switch models often. The model is only half of it: retrieved web snippets are stuffed into the prompt (RAG), then the LLM writes the tidy answer. That is why you can get a polished wrong day-count the model is generating prose, not running a calendar.
Leo, the sidebar/address-bar assistant in the Brave browser, is a separate product. Default is Automatic: Leo picks a model per task. You can also choose one yourself at brave://settings/leo-ai.
Free tier typically includes smaller open models such as Qwen 3 VL 30B, Llama 3.1 8B, Llama 4 Scout/Maverick, GPT-OSS 20B, and Claude Haiku. Premium ($15/mo) unlocks larger ones such as Claude Sonnet/Opus 4.6, DeepSeek V3.1, Mistral/Pixtral Large, Qwen 3 VL 235B, Qwen 3 Coder 480B, and GPT-OSS 120B. Lineup changes. Tab Focus Mode has been documented as using Claude. You can also point Leo at a local model through Ollama (BYOM).
Brave says the hosted models run on its infrastructure, not sent to Google/OpenAI for training, and IP is stripped. That is a privacy claim, not a guarantee the answers are numerically correct as that June 30

No Spin Ag said:

---

ErnestEndeavor said:

---

This is a great video to listen to while you are out in your car or mowing the lawn or whatever. This is a professor of finance at NYU who is doing a lot of math on the AI story, he's on this podcast talking about the serious financial issues these companies are facing. The amount of revenue the frontier companies will need to be able to pay their bills in the next few years is staggering and there's no indication they will ever get there.

These data centers are a huge misallocation of capital and the fewer that get built right now is probably a good thing.

---

So is AI the next bubble many projected it would be about a year ago?

---

it's not just the next bubble...it's the next too big to fail. Get ready for government bail outs for some of these companies.

Aggie95 said:

---

No Spin Ag said:

---

ErnestEndeavor said:

---

This is a great video to listen to while you are out in your car or mowing the lawn or whatever. This is a professor of finance at NYU who is doing a lot of math on the AI story, he's on this podcast talking about the serious financial issues these companies are facing. The amount of revenue the frontier companies will need to be able to pay their bills in the next few years is staggering and there's no indication they will ever get there.

These data centers are a huge misallocation of capital and the fewer that get built right now is probably a good thing.

---

So is AI the next bubble many projected it would be about a year ago?

---

it's not just the next bubble...it's the next too big to fail. Get ready for government bail outs for some of these companies.

---

Good point.

That might be why they're all of a sudden asking for government to step in.

Turns out we're just increasing the speed and capacity of pocket calculators.

I was programming "AI" in middle school on a TI-83 when I wrote BASIC code that took in a quadratic equation and returned the y=0 crossings.

Dungeon Crawler Carl said:

---

We could be in for an epic down day on the NYSE...... (unless the plunge protection team is on the ball)

Oracle AI bet ****ting the bed, Fuel export ban on the table, Xi in DC for a big meeting, and those pesky 5 and 10 year interest rates with Bessent announcing another $6 Billion in intervention yesterday.

Quote:

---

9/23/2026

---

---

and this type of analysis is why I come to TEXAGS F16

23,816.49 (-0.13%)

down an entire one tenth of one percent.

TexasRebel said:

---

LOYAL AG said:

---

Hank the Grifter said:

---

Honest question from an "old".
I keep seeing people saying they are "harnessing" AI to help their day to day routines.
What? How?
Can you give concrete examples AND how to leverage the AI to do those things?
My experience with AI is typing questions into ChatGPT.

*Edited to correct fat fingers/auto corrects*

---

Last one but this and my other two are representative

Same client has an annual budget requirement. The template is in Excel and has to be populated with current year budget and projected end of year actuals with info hitting several tabs in that Excel workbook. The antiquated accounting system produces PDF docs for YTD actuals and the government system we upload budgets to gives us an approved budget in PDF as well. Claude takes both PDF and fills out the current year budget, extends YTD actuals to approximate end of year final numbers then populates those into the template before applying various rules to create the 2027 budget. Prior years took about 1-1/2 hours each of data entry so for 150+ properties we saved about a month of man hours in data entry.

---

Because your sources refused to give you formatted data? Like a simple .csv?

---

It's complicated but the PDF generally comes down to antiquated systems that produce that file type in bulk (meaning for all properties at once) and to get an Excel friendly format you're exporting a PDF one at a time. In other cases the source data doesn't exist, hence the need for multiple skills for different inputs. It's not anyone being a PITA it's just the world we live in.

The real point though is that the end result is an Excel template that looks a lot like a PDF such that the various reference formulas in Excel won't work. Regardless of how I get the inputs Claude is significantly faster at this than I am.

LMCane said:

---

Dungeon Crawler Carl said:

---

We could be in for an epic down day on the NYSE...... (unless the plunge protection team is on the ball)

Oracle AI bet ****ting the bed, Fuel export ban on the table, Xi in DC for a big meeting, and those pesky 5 and 10 year interest rates with Bessent announcing another $6 Billion in intervention yesterday.

Quote:

---

9/23/2026

---

---

and this type of analysis is why I come to TEXAGS F16

23,816.49 (-0.13%)

down an entire one tenth of one percent.

---

How was he going to predict the end of the Iran war, which happened today.

GeorgiAg said:

---

TexasRebel said:

---

Turns out we're just increasing the speed and capacity of pocket calculators.

I was programming "AI" in middle school on a TI-83 when I wrote BASIC code that took in a quadratic equation and returned the y=0 crossings.

---

Your understanding of AI is unparalleled.

---

DIV/0 edge cases were tough.

Alternatively:

No GPUs on a TI- graphing calculator.


*AI isn't smart enough to get either of these jokes.

GeorgiAg said:

---

LMCane said:

---

Dungeon Crawler Carl said:

---

We could be in for an epic down day on the NYSE...... (unless the plunge protection team is on the ball)

Oracle AI bet ****ting the bed, Fuel export ban on the table, Xi in DC for a big meeting, and those pesky 5 and 10 year interest rates with Bessent announcing another $6 Billion in intervention yesterday.

Quote:

---

9/23/2026

---

---

and this type of analysis is why I come to TEXAGS F16

23,816.49 (-0.13%)

down an entire one tenth of one percent.

---

How was he going to predict the end of the Iran war, which happened today.

---

Why does this all look familiar?

Did yesterday happen again, again?

Hank the Grifter said:

---

Honest question from an "old".
I keep seeing people saying they are "harnessing" AI to help their day to day routines.
What? How?
Can you give concrete examples AND how to leverage the AI to do those things?
My experience with AI is typing questions into ChatGPT.

*Edited to correct fat fingers/auto corrects*

---

I use Claude Anthropic Cowork. I have five folders that are a template

about me
templates
research
output
source data

about me and templates has markdown files that contain generic boilerplate info for AI to use on each case file. Me, my writing style/rules/preference with examples, Georgia law, where I work, courts I'm in, my assistant, contact info, etc.

Research I add some statutes because for now Lexis nerfed our state (Georgia - only one of 4 states) statutes - AI isn't allowed to get around a CAPTCHA. I am connected via MCP to Midpage.ai - which is a sandboxed AI that only has verified case law, regulations and such.

Source data - I drop documents, contracts, images, video, letters, pleading specific to that case. Claude will read and digest in minutes, even dozens of large files

I give Claude a task and it reads through all my folders and does what I want and puts it in the Output folder. The results are mind-blowing. Things that would take days, now take an hour. I review it for accuracy, tweak it and it's a finished product.

It's like having a very intelligent associate do the work and I just review it now.

I just did this today for a nastrygram demand letter for an architect stiffed on a bill. Turned it around in less than an hour. Client email, "This looks good. Very tough! Much better than I would have done if I wrote it myself, lol." A client could use this, but he doesn't know the underlying law to be able to check it. Like I've said with medicine. I don't want to get medical advice from AI, but I want to get it from a doctor who uses AI.

FWTXAg said:

---

Good luck out Terminator'ing China.


People arguing for us to attempt to outcompete China in constructing the dystopian future on its' way are going to be thoroughly dissapointed. They don't need to spread any propoganda to beat us to that end, they're already at the finish line and we just left the starting blocks.

---

Defeatist. Maybe you should go to China and lick their boots. They dont have the AI edge but they are pimping the lies as always

AI is quite good at roasting you based on your case history by the way. If you haven't asked it yet, go do so. It's a little bit scary.

TexasRebel said:

---

Turns out we're just increasing the speed and capacity of pocket calculators.

I was programming "AI" in middle school on a TI-83 when I wrote BASIC code that took in a quadratic equation and returned the y=0 crossings.

---

Wow

TexasRebel said:

---

GeorgiAg said:

---

TexasRebel said:

---

Turns out we're just increasing the speed and capacity of pocket calculators.

I was programming "AI" in middle school on a TI-83 when I wrote BASIC code that took in a quadratic equation and returned the y=0 crossings.

---

Your understanding of AI is unparalleled.

---

DIV/0 edge cases were tough.

Alternatively:

No GPUs on a TI- graphing calculator.


*AI isn't smart enough to get either of these jokes.

---

Claude's response:

Challenge accepted.
The first one is understatement. Division by zero isn't a tough edge case. It's undefined, and the limits run to positive and negative infinity depending on which side you approach from. Calling that "tough" is the engineering equivalent of calling a building collapse "a setback."
The second one is irony. A graphing calculator exists to draw graphs, and it has no graphics processor. It also means a TI-84 can't run AI, which makes the asterisk the actual punchline: the one thing that can't understand the joke is explaining why.
What I can't find is the setup. "Alternatively" suggests both are answers to the same question. My guess is something like "Why did the AI fail calculus?" If that's right, I got the jokes and missed the premise, which is a very AI way to lose.

Yup.
So far over its head it caused a variable rollover.

…and apparently an off-by-one bug.

I would encourage everyone to incorporate a little bit of AI work every week.

You will be surprised at how well it can work.

ttu_85 said:

---

FWTXAg said:

---

Good luck out Terminator'ing China.


People arguing for us to attempt to outcompete China in constructing the dystopian future on its' way are going to be thoroughly dissapointed. They don't need to spread any propoganda to beat us to that end, they're already at the finish line and we just left the starting blocks.

---

Defeatist. Maybe you should go to China and lick their boots. They dont have the AI edge but they are pimping the lies as always

---

Sure thing tough guy.

No Spin Ag said:

---

Aggie95 said:

---

No Spin Ag said:

---

ErnestEndeavor said:

---

This is a great video to listen to while you are out in your car or mowing the lawn or whatever. This is a professor of finance at NYU who is doing a lot of math on the AI story, he's on this podcast talking about the serious financial issues these companies are facing. The amount of revenue the frontier companies will need to be able to pay their bills in the next few years is staggering and there's no indication they will ever get there.

These data centers are a huge misallocation of capital and the fewer that get built right now is probably a good thing.

---

So is AI the next bubble many projected it would be about a year ago?

---

it's not just the next bubble...it's the next too big to fail. Get ready for government bail outs for some of these companies.

---

Good point.

That might be why they're all of a sudden asking for government to step in.

---

Their excuse for the bailouts is going to be national security. Shameless con artists.

Hank the Grifter said:

---

Honest question from an "old".
I keep seeing people saying they are "harnessing" AI to help their day to day routines.
What? How?
Can you give concrete examples AND how to leverage the AI to do those things?
My experience with AI is typing questions into ChatGPT.

*Edited to correct fat fingers/auto corrects*

---

Gen-X here *****es…. I use the hell out of AI. It's hooked to all our "replicated data warehouses" and I'm able to access and analyze in minutes what it took me hours to find and export and manipulate in excel to report out. I can't code for **** but I have built 6 separate apps in last 6 months to improve not just my productivity but entire team. It has increased my productivity and made me indispensable at office. But also increases my work load. Because I am the AI guy. Yet I should be the least AI guy cause I'm ok with computers but I'm no expert.

I think the topic should be "is today the data center bubble burst?". AI is obviously a persistent force…question is how much computing is needed/when are returns on data centers diminishing?

ttu_85 said:

---

Logos Stick said:

---

Good thing Oracle is not "the AI".

I'll keep using AI to greatly enhance my productivity. I encourage the rest of you to avoid it.

---

Funny. It is a fantastic tool but so is an automatic 7.62 machine gun. Its all about the application. Using a 7.62 to kill rabbits is not very efficient. Same attitude regarding AI. The clueless just see the Terminator.

AI has a zillion uses and China wants everyone in the US scared like the OP. People wanting to blindly ditch the tool of our age are the tools of the CCP- literately

---

M2 .50 Cal in single shot mode or GTFO

LOYAL AG said:

---

It does all that in 3-4 minutes where it used to take me several hours in Excel and I'm really good in Excel.

---

I'm really good at Excel too. But... it takes me a little while to remember what the heck I'm doing because I'm not in it every day. Maybe once a quarter.

The thing with Claude is, it doesn't get memory fog. It can remember exactly what I had it do the last time I used it.

So there's a significant time savings from not having to clean out the Excel cobwebs in my brain every time I use it.

Hank the Grifter said:

---

Honest question from an "old".
I keep seeing people saying they are "harnessing" AI to help their day to day routines.
What? How?
Can you give concrete examples AND how to leverage the AI to do those things?
My experience with AI is typing questions into ChatGPT.

*Edited to correct fat fingers/auto corrects*

---

I'm not big into AI but it has done wonders for remodeling our house. Take a picture and ask Grok to add light wood floors, or plantation shutters etc.

Also, i'm on our youth baseball league's board and part of that responsibility is running a specific ages try outs. I am the 7u commissioner and we had 120 kids for the fall. We had two graders right down the kids name and then scores for fielding, catching, batting etc. I literally took a picture of all 5 sheets and told grok to grade the players and put them in order and then put it in a spread sheet. 6 minutes later it did it. That is something that used to take me 5-6 hours to manually import all of that data.

HollywoodBQ said:

---

ttu_85 said:

---

Logos Stick said:

---

Good thing Oracle is not "the AI".

I'll keep using AI to greatly enhance my productivity. I encourage the rest of you to avoid it.

---

Funny. It is a fantastic tool but so is an automatic 7.62 machine gun. Its all about the application. Using a 7.62 to kill rabbits is not very efficient. Same attitude regarding AI. The clueless just see the Terminator.

AI has a zillion uses and China wants everyone in the US scared like the OP. People wanting to blindly ditch the tool of our age are the tools of the CCP- literately

---

M2 .50 Cal in single shot mode or GTFO

---

GAU-8?

Windy City Ag said:

---

Quote:

---

China is cheating their ass off as usual. They are feeding their open weight models chat transcripts from Claude to train them. So if you asked them to do stuff Claude had already figured out it worked great. If you asked it something novel, not so much.

---

I don't disagree, and it begs the larger existential question of all models open or closed "cheating their ass off" by repurposing other people's intellectual property and work output for commercial profit.

But in China's case they have shown the world the OpenAI and Anthropic approach of incredibly capital intensive closed model systems may not make much sense if OpenSource versions can fast follow at a fraction of the cost.

If China can do it, then anyone can do it and corporations are probably going to build their own versions that they can control and monitor that in turn lean on a wide variety of specialist, narrow ai programs like Legal AI or Medical AI etc.

This is basically what the software stack looks like now.

---

Agree with your thoughts. Instead of corporations though, think country level. There is an entire sovereign movement underway right now because everyone else but us and China have suddenly realized two countries are going to be a global duopoly unless they build their own capability.

NM

This has nothing to do with AI. This a deliverable problem between Bloom Energy and Oracle. Oracle has contracts to deliver on a certain date. Bloom let them know they were going to slip the date. That combined with a bad grid situation leads to massive, and I mean massive penalties. In most contracts there are no outs other than force majeure and that is delaying tactic at best.

I have been involved in 3-400MW projects where the penalties were $1M a day. I can only imagine what the penalties are on a 2.5GW project.

The datacenter power delivery business takes no prisoners.

TKEAg04 said:

---

You are truly falling behind if you are not harnessing some of the latest models in AI. I use Grok Bots now for almost full automation of all of my workflows and it's mind blowing just how good they are now.

---

I have no idea what a workflow is. They call us into meetings to talk about it, and I leave having no idea what it was about.