Sponsored by
Today's theme is who pays when AI gets real. Washington is pulling a Chinese model out of government systems, investors are pricing in the chance that a major AI data center company defaults, and Google is learning that a chatbot's mistakes can end up in court. Here is what each one means for your business.
State of Product reveals what AI still hasn’t solved
80% of product professionals say AI helps them ship faster, but customers aren’t seeing value any sooner. Atlassian’s State of Product 2027 explores the gains AI is delivering and the challenges that remain, from decision-making that hasn’t kept pace to gut instinct overriding customer evidence.
The Federal Government Banned DeepSeek, and Anything Built on It
The Department of Homeland Security ordered federal civilian agencies to stop buying DeepSeek products and remove them from government systems, according to TechCrunch. The notice also covers third-party tools built with DeepSeek's technology, not just DeepSeek's own apps. DHS cites national security concerns, warning that sensitive data from U.S. government users could be accessible to Chinese authorities under Chinese law. Agencies can request exceptions, and China's foreign ministry said it opposes politicizing technology and trade.
Why it matters to you: The ban applies to federal agencies, not to you. The lesson is in the third-party part: many AI apps are wrappers around someone else's model, and you may not know whose. If you sell to government agencies or their contractors, expect questions about what is under the hood of your software. Everyone else should still know where customer data goes.
What to do about it: Ask each AI tool vendor you use which model runs behind it and where your data is stored.
One AI Giant Owes $50 Billion, and Bond Markets Are Getting Nervous
CoreWeave, a company that rents out computing power for AI, saw its shares fall more than 15% in one session last week, then another 6% in premarket trading the next day, Reuters reported via BNN Bloomberg. Yields on its bonds due 2031 rose above 16%, a level usually seen as a sign of distress, and the cost of insuring its debt implies roughly a 63% chance of default within five years, up from about 30% a year earlier. The company carries about $50 billion in debt and is not yet profitable. CEO Mike Intrator told employees its deal backlog more than covers its obligations and that markets can act irrationally.
Why it matters to you: The AI tools you pay for run on data centers built with borrowed money. If lending gets harder, the likely result is not that AI disappears. It is that prices rise, free tiers shrink, or weaker providers get bought or shut down. Gartner, cited in the same report, projects that about 80% of these newer cloud providers could shut down by 2030, though that is a forecast, not a fact.
What to do about it: Keep your key data and workflows portable so you can switch AI providers without starting over.
German Courts Say Google Is Responsible for What Its AI Says
A Munich court ruled in May that Google was directly liable for false claims in its AI Overviews that tied two publishers to fraud, treating the AI summaries as Google's own statements, The Decoder reports. Google appealed, but the case ended in a settlement, so no higher court has ruled. On September 24, a court in Aachen issued a preliminary injunction over AI answers that falsely linked a hospital physician to a sexual assault allegation, with fines of up to €250,000 per violation. That decision is not final, and Google can challenge it.
Why it matters to you: Few things hurt a small business like a false claim sitting at the top of search results. These rulings are German, preliminary, and do not bind U.S. courts. But they point toward AI answers being treated as the company's own speech. In the meantime, AI search can say wrong things about your business today.
What to do about it: Search your business name in Google and ChatGPT this week and screenshot anything false.
Most business owners I know share one problem: the sales process lives in their head, so it only runs when they have the energy to run it. That is exactly what the Epic Marketing Summit is built to fix. It is a 2-day live event in Miami this January where you build a working AI-powered sales funnel and vibe code your own AI app on-site, so you walk out with a finished asset instead of a notebook full of frameworks you still have to go implement. Last year's event sold out, and it returns with a bigger AI focus.
I wasn't there last year and I had major FOMO. I know a lot of the trainers, and they weren't talking theory, they'd actually done the work. I'm going in January, and if you're serious about building leverage into how you sell, I'd like to see you there too. The closer we get, the faster seats go. Grab your spot here.
WhatsApp's Support Bot Was Told Not to Mention Privacy Regulators
Privacy software firm Redact found 102 hidden blocks of text across 80 pages of WhatsApp's help center, written as instructions for a support chatbot, The Hacker News reported following WIRED's coverage. The notes told the bot not to recommend contacting a data protection regulator, and not to cite GDPR or other privacy laws once a user mentioned a lawyer or court. WhatsApp's European privacy policy says users can complain to a regulator. A WhatsApp spokesperson told WIRED the text was meant for the chatbot and was caused by a bug, and the blocks were removed October 18 after WIRED contacted Meta. Redact did not test whether the chatbot actually followed them.
Why it matters to you: If you run a customer service chatbot, you write rules for it too. Rules like "don't give legal advice" are normal. Rules that steer customers away from their legal rights are a lawsuit and a headline waiting to happen, and hidden text in your page code can be read by anyone who looks.
What to do about it: Reread your own chatbot's instructions and ask whether you would be comfortable seeing them printed in the news.
Gartner Says AI Agents Will Cost 5 Times More by 2028, Even as Tokens Get Cheaper
Gartner predicts the cost of running an AI agent workflow will rise more than fivefold through 2028 even though the price per token (the unit AI is billed in) keeps falling, a pattern it calls the "Inference Paradox," VentureBeat reports. Agents use 5 to 30 times more tokens than a basic chatbot for the same task, and AI labs are reinvesting efficiency gains into pricier next-generation models. In a VentureBeat survey of 109 respondents, 21% of enterprises using an orchestration platform had nothing that automatically stops agent spending.
Why it matters to you: A lower price per unit does not mean a lower bill when the AI does 30 times more work per task. Gartner's fixes are plain: track spending, tie each AI feature to a measurable result, and send routine jobs to cheaper models. That survey is small, so treat the 21% as a hint, not a verdict.
What to do about it: Set a monthly spending cap on any AI tool billed by usage, and review it.
Quick personal note. I've been a member of the Ignite Mastermind for over 3 years, and it's been a vital part of how I've grown my business. Some of my closest friends and a few real business partners came directly out of being in that room, and that's the part I'd lead with.
The practical side is strong too. Ignite is an annual membership for entrepreneurs still under $1 million a year who want structure, tools, and peers to get past that mark faster. You get 6 to 8 live virtual workshops a year plus 2 in-person events, each normally priced around $97 to $997 on its own. You also get a marketing software suite that replaces over $5,000 a month in separate tools, and weekly calls with Perry Belcher and his team. Join through me and I add monthly accountability and business development calls with me personally. If you want a mastermind to carry you to your next milestone, this is where I'd point you. See how Ignite works here.
An Anthropic Co-Founder Told the IMF the Singularity Has Begun
In an October 17 lecture at the IMF meetings in Bangkok, Anthropic co-founder Jack Clark argued that the early stage of the singularity, meaning AI speeding up its own progress, has arrived, according to his Import AI newsletter. He cited scores on Humanity's Last Exam rising from 8% to 67.7% and FrontierMath from 2% to 94%, plus an Anthropic scenario with U.S. GDP reaching $44 trillion by 2030. He admitted the GDP figure "sounds outlandish" and depends on AI completing a broad range of tasks and spreading fast. He also urged building power, compute, and lab capacity, and keeping human oversight on biological research.
Why it matters to you: Clark co-founded a company that sells AI, and much of his evidence is his own company's data. Benchmarks measure test performance, not whether software will handle your invoices. Take the signal that serious people are planning for fast change, and ignore the countdown clock.
A Startup Will Fix Your Plumbing for Free, as Long as It Can Record the Repair
AI lab HUMXN is subsidizing plumbing, electrical, HVAC, air-duct cleaning, and carpet cleaning in Minneapolis, Chicago, and Miami, Tom's Hardware reports. Technicians wear sensors that record how they diagnose problems and choose tools, and HUMXN turns that footage into training data for AI robots. Homeowners get free service and contractors get paid. I could not find eligibility rules or much of a public track record for the company, so treat the details as early.
Why it matters to you: If you run a trades business, this is a preview of how your know-how becomes training data. If you hire tradespeople, a camera inside a customer's home raises real privacy questions.
What to do about it: If anyone offers to record your work, ask who owns the footage and whether customers have to consent.
The Bottom Line
Today's stories share one theme: someone always pays in the end. Governments pay in access, data centers pay in interest, and businesses pay in higher prices and more accountability. None of it is a reason to panic, and all of it is a reason to ask boring questions. Whose model is this, who is paying for it, and what happens when it's wrong? That is the whole job. Keep your head, keep your options open, and stop the rollercoaster.
Enjoying the Ride?
Forward this issue to one person who runs a business and keeps hearing about AI. If someone forwarded this to you, you can subscribe at stoptheairollercoaster.com.
Talk tomorrow,
Mark Shilensky
Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social