OpenTube

Banks Just Found Out What AI Debt Is Virtually Worth: Zero

House of El: AI

493,338 views

Enjoy 1 month of Wispr Flow Pro, on us from House of El AI:
https://ref.wisprflow.ai/house-of-el-...

AI companies are asking investors for sums of money so enormous they are starting to reshape the financial system around them, even as public trust is falling, employees are walking away, and the evidence on productivity remains far more complicated than the headlines suggest. The technology is real, the gains are real, and so are the failures. But underneath all of it sits a financial structure built on hundreds of billions in debt, future revenue projections, guarantees, and one very important assumption: that AI becomes valuable enough, fast enough, to pay for everything being built around it. This video follows that money and asks what happens if the technology succeeds, but not on the timeline the financial system is betting on.

☕ Support the channel on Ko-fi:
➡️ https://ko-fi.com/houseofel

📌 Chapters
00:00 Introduction
01:42 Why Are People Losing Faith In AI Companies?
08:50 Is AI Delivering On Its Promises?
13:32 What's The Underlying Financial Structure
16:35 Follow The Money
22:22 What El Thinks About This

📚 SOURCES & FURTHER READING

Brynjolfsson, Li & Raymond (2025). "Generative AI at Work."
https://academic.oup.com/qje/article/...

Rowan, Mittal, Ammanath & Perricos (2026). "State of AI in the Enterprise."
https://www.deloitte.com/content/dam/...

Becker, Rush, Barnes & Rein (2025). "Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity."
https://arxiv.org/abs/2507.09089

David (1990). "The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox."
https://www.jstor.org/stable/2006600

Dikici & Bilgin (2026). "Small Language Models: A Systematic Review of Computational Trade-Offs, Privacy Advantages and Deployment in Intelligent Systems."
https://onlinelibrary.wiley.com/doi/f...

Wang & Chen (2026). "Boom, Bubble, or Buildout? A Multi-Method Evaluation of Whether Artificial Intelligence Is in an Ongoing Financial Bubble."
https://arxiv.org/abs/2606.01575

Kedrosky, Wahi & Preston (2026). "The AI Bubble: Hidden Risks and Opportunities."
https://theideafarm.com/wp-content/up...

#AI #ArtificialIntelligence #OpenAI #Nvidia #AIBubble #WallStreet #TechIndustry #Economics #AIInvesting

20 Comments

@houseofel-ai Creator 8 days ago Pinned

Enjoy 1 month of Wispr Flow Pro, on us from House of El AI: https://ref.wisprflow.ai/house-of-el-ai-yt

@Daemonworks 8 days ago

"Give us more money!"
"What's your business model?"
"That is our business model"

@MitchAshdown 8 days ago (edited)

Ask customers about customer support, don't ask CEO's. It is a stretch to say customer support even exists now.

@Alfaomegabravo 7 days ago

"Our value is Trillions"
"So you earn tens of billions?"
"No, we lose tens of billions!"

@ninamiller001 8 days ago (edited)

Measuring "productivity" is meaningless when the quality of the work produced with AI varies widely. My boss, a very smart and capable person, uses AI to produce reports and policy papers more quickly. He then hands them off to staff to revise. I am sick to death of spending countless hours unpacking these documents like elaborate puzzle boxes to figure out if statements are actually true, word choice is appropriate, and figures are accurate. It is an extremely time consuming process. If these documents were written by a human they would not have the volume of errors, including hallucinations, exaggerated claims, incorrect citations, and inaccurate figures. Now, it takes me that long because I am scrupulous. If I had no scruples, these published documents would be riddled with false claims and errors. And because they sound plausible, it likely no one would notice.

So the question is - what does productivity mean in relation to quality, customer satisfaction, and effort level.

@timshort2601 7 days ago

I am 70 and have watched this play out several times during my life. When there is a glut of cash banks will loan for anything, because they think they cant lose, as cash gets tighter they end up getting stuck holding the ball no one wants and they act like oh its not our fault

@HectorDiabolucus 8 days ago

Automated stupidity is faster stupidity.

@the-last-codebender 8 days ago

"Not more valuable than the people watching this video" And I had to switch screen and press like when I heard that. Thank you.

@feedmatt 8 days ago

You hear that!? We are more valuable than Open AI everyone!

@siverblack2593 8 days ago

Important to remember that its not just Flock. They are just the most talked about name. Other ALPR companies need to be stopped, too.

@joewestla 8 days ago

Turns out replacing employees with workers who lie repeatedly, require a blank check to complete a task, and accept zero responsibility is a tough sell to the CFO.

@sandradernerova4352 8 days ago

"Vibes based accounting has not yet been accepted by SCC... just give it a year, it´s gonna be fine." that cracked me up :).

@spinesheath 8 days ago

Let's be clear: Flock would not have been acceptable even if it did what it promised.

@entropyachieved750 8 days ago

I'm all for these AI bros losing money. Hopefully they won't have any left and can't influence our elections or government.

@2a5x1retired 7 days ago

Some lady in Florida just spent almost 2 weeks in jail because of Flock. Completely innocent of the thing she was accused of.

@miranina1585 7 days ago (edited)

Around 11:20 : what you're discussing is something I teach my clients : do not bring a technological solution to a process problem. Technology will only do your mistakes faster.

@coywoolley5734 8 days ago

I can confirm as an experienced developer, I have become less productive since some coworkers started using LLMs for both code and documentation. The automated Claude commands are slower and less reliable than deterministic scripts, and the slop content that appears in documentation makes my review of coworkers' committed code take significantly longer. At best it is way too verbose, and at worst it completely misrepresents what the code is doing. It makes some of my coworkers feel better I guess, but they're trying to force it into tasks it doesn't belong in.