Mark Cuban Puts Agents on Healthcare Lawsuits and Flags a Possible AI Debt Bubble. Here Are 10 Lessons.
1. Key Themes
Theme A: Agents should take over the work humans skip, while humans keep judgment
Narrow, repeatable tasks are the right first agent
Cuban's own first agent handles a single recurring job: monitoring healthcare conglomerates and drafting records requests.
"I can use the agent to create an email to send to the Open Records Act to get me copies of all those contracts."
"That's a textbook first agent, because the job is narrow and it repeats."
Agents unlock work that previously went undone due to attention cost
The value isn't replacing staff but making previously uneconomical work possible.
"Before agents, he says, that work meant hiring and managing someone, so it stayed undone. Now he hires for higher-level work and spends less of his own time."
Humans retain "problem identification"
Cuban argues AI can gather and combine inputs but misses the step of defining the problem.
"I have to use my senses, look, take my experience and be smart enough to ask the right question."
"He defines smart as gathering and combining inputs, and he says identifying the problem is the piece AI misses."
Theme B: AI outputs are probabilistic drafts, not answers
Model variance is a feature to manage, not ignore
Cuban stresses that different models, or even rewordings, produce different outputs.
"It's predictive. It's not deterministic. Meaning, if you asked 4 different AIs the same question, you're going to get 4 different answers."
Operational implication: compare, reword, and rerun after model swaps
"I'd run his reworded-question test on every key prompt, and I'd rerun it after any model swap."
Theme C: Financing risk, not technology risk, is the AI worry
Hyperscaler borrowing could create a debt bubble
Cuban backs the technology but questions how it's being funded.
"We literally could get into a debt bubble in the United States of America because there's so much debt being taken out by all these hyperscalers."
Supporting data
"Fortune reported in September 2026 that the 5 biggest hyperscalers issued $121 billion in US corporate bonds in 2025, against an average of $28 billion a year from 2020 to 2024, citing BofA Securities. Their capital spending now runs close to 100% of operating cash flow, it adds."
Efficiency as the counter-thesis
"Cuban says a lot of data centers will turn into pickleball courts, as a metaphor, because AI peaks with less power doing more."
Theme D: Public markets trade on sentiment, not fundamentals
Outsized reactions to tiny misses
"A company might miss earnings by a penny or only beat by a penny and their stock will be down 18%."
Perceived value over intrinsic value
"Cuban says the company's intrinsic value changed by far less than that, because buyers and sellers trade on anticipated information instead of the net present value of future cash flows."
"We're trading them like baseball cards."
Theme E: Intellectual property strategy is shifting toward secrecy
Publishing now feeds every model
"It's immediately available to every model and you lose control of your intellectual property."
Trade secrets over patents and papers
"He sees more of them choosing trade secrets over patents and a private model over a paper."
2. Contrarian Perspectives
The biggest AI proponent says AI is "just predicting the next word"
Cuban pairs maximal enthusiasm with skepticism about AI's limits, rejecting hype around its reasoning ability.
"Mark Cuban says AI just predicts the next word. Then he points agents at the lawsuits and contract bids of the healthcare conglomerates he calls the problem."
"He also calls himself 'the biggest proponent of AI that there is,' and he still spends much of his long sit-down with Steven Bartlett on The Diary of a CEO explaining where AI falls short and where the money behind it looks thin."
The AI-security fear is overstated; defenders can use the same brute force
Rather than treating AI-enabled hacking as a novel threat, Cuban argues attackers just run the standard list at scale, and defenders can too.
"What we don't talk about is using that same brute force on the defense side, that you can go out there and find all the holes in yourself."
"The answer he reports: it's the same list every security person has, tried in every combination by an unlimited number of agents."
"He says he fears what people do out of concern and misunderstanding more than what AI will do."
Open knowledge-sharing is now a liability, and models will splinter
Against the open-science norm, Cuban says publishing erodes IP and fragments AI into models with different training data.
"Gemini and ChatGPT train on different data, so they give different answers, which he says defeats much of the purpose."
3. Companies Identified
- Description: Cuban's drug-pricing company
- Why mentioned: The testbed for his first agent, which monitors healthcare conglomerates and drafts Open Records Act requests. Also where hospitals and researchers are choosing trade secrets over publishing.
- Quote: "For Cost Plus Drugs, an agent tells him whenever a healthcare conglomerate he calls the problem gets sued or bids on a contract. Then it drafts the Open Records Act request."
- Description: Cuban's early PC software business
- Why mentioned: Case study in learning software deeply to gain a competitive edge.
- Quote: "He wanted a huge edge for Micro Solutions."
Anthropic / Google (Gemini) / ChatGPT
- Description: Leading AI model providers
- Why mentioned: Cited as examples of model variance and divergent training data.
- Quote: "Cuban says a model from Anthropic or Google is out of date the moment it ships."
Hyperscalers (unnamed "5 biggest")
- Description: Large cloud/AI infrastructure companies
- Why mentioned: Subject of Cuban's debt-bubble concern.
- Quote: "The 5 biggest hyperscalers issued $121 billion in US corporate bonds in 2025."
Lotus 1-2-3
- Description: Early spreadsheet software
- Why mentioned: What Cuban mastered via manuals to build his edge.
- Quote: "In the early PC days, Cuban learned packages like Lotus 1-2-3 by reading the manuals over and over."
- Description: AI operations consultancy (sponsor)
- Why mentioned: Sponsor advocating one tested workflow over multi-year roadmaps.
- Quote: "You'll get further with one workflow tested on real cases than with a multi-year roadmap."
Bitcoin, Micron, Google, Apple
- Description: Public assets/companies
- Why mentioned: Cited as moving like meme stocks on small news.
- Quote: "Bitcoin, Micron, Google and Apple move like meme stocks on small news, Cuban says."
4. People Identified
- Description: Entrepreneur, investor, Cost Plus Drugs founder
- Why mentioned: The subject of the interview and all ten lessons.
- Quote: "Like for me with AI, I see it as an opportunity. Wherever there's change, there's opportunity."
Steven Bartlett
- Description: Host of The Diary of a CEO
- Why mentioned: Conducted the long-form interview.
- Quote: "The interview itself shows the gap: Bartlett's next question kept surprising him."
- Description: Venture capitalist
- Why mentioned: Linked for his view on where the AI moat sits.
- Quote: "Marc Andreessen places the AI moat outside the model."
- Description: Venture investor
- Why mentioned: Recommended reading on a venture rule AI supposedly broke.
- Quote: "Read Martin Casado on the venture rule he says AI broke."
Scott Galloway
- Description: Professor and commentator
- Why mentioned: Presses the return question on AI spend.
- Quote: "Scott Galloway presses the return question on AI spend."
Andrej Karpathy
- Description: AI researcher
- Why mentioned: His workflow shift is cited as a resource.
- Quote: "Andrej Karpathy's workflow shift."
5. Operating Insights
Pick one narrow, recurring task for your first agent
"List 3 tasks you skip because staffing them costs too much attention, then put 1 agent on the first."
Run key prompts on multiple models and keep a fallback
"Run each key prompt on 2 models and keep a fallback, since Cuban says 4 models give 4 answers."
Study every new release the week it lands
Cuban's overnight-manual approach translates to reading release notes immediately.
"The creator first and everyone else tied for second... Release notes are the manual of 2026, so read them the week a model lands."
6. Overlooked Insights
Plan for the "painful middle" of creative destruction
Cuban's Terre Haute example (disc plants moving from CDs to DVDs) is a concrete reminder that transition costs fall on specific communities.
"Cuban points to Terre Haute, Indiana, where disc production moved from CDs to DVDs. He adds that people still make money on music, that new technology creates different kinds of jobs and that the path hurts."
Investors should stay patient and hold cash for the turn
A brief, actionable portfolio comment buried in the playbook section.
"He also says to stay patient and keep some cash for the turn."