Welcome to Black Sheep, a spin-off publication of my serialized memoir. SMIRK. While SMIRK was a deep dive into my unusual personal and professional relationship with one unique white-collar fraudster— Martin Shkreli — Black Sheep takes a broader view and tells the stories of a wider range of business crimes and failures.
This publication will examine cultural themes and motives that contribute to lying, cheating, stealing, and related self-inflicted disasters; the impacts of those events; and the people who play starring roles in these dramas. I find these tales both cautionary and fascinating; I hope you will, too.
If you’re looking for SMIRK, you can find the full table of contents and links to all the posts in chronological order here. Paid subscribers can access all the posts; free subscribers can access select posts. Thanks for reading!
Inflated Expectations
Long ago — ok, six months ago, but feels like decades in tech years – I wrote a Black Sheep piece about an AI startup founder with a chatbot that gained rapid traction in the sports entertainment industry. But customers didn’t want to pay very much for it. So the founder, Alexander Beckman, allegedly turned to fraud.
Today, Beckman and his wife, Valerie Lau (who allegedly helped him with the fraud), are facing an ongoing criminal case and the prospect of prison. Meanwhile, the AI industry’s trillion-dollar promises of value creation have yet to yield results. Numerous other founders are undoubtedly battling pressures that could push them down the same path.
Let’s look at some recent headlines, shall we?
WSJ, Aug. 2, 2025 – “The AI Boom’s Hidden Risk to the Economy.” The rush to build AI infrastructure is “straining companies and capital markets,” but hey, at least it’s propping up the economy while consumers pull back on spending.
CNBC, Aug. 18, 2025 – “OpenAI’s Sam Altman Sees AI Bubble Forming.” Trying to get ahead of the inevitable, the CEO himself drops the “B-word,” and talks about “smart people” getting “overexcited about a kernel of truth.”
AIInvest, Aug. 24, 2025 – “Massive Credit Fueling AI Boom as Debt Risks Rise.” Tens of billions are flowing into AI via…CMBS (commercial mortgage-backed securities). I feel a twinge of 2008 real estate crash deja vu.
The Information, Sept. 5, 2025 – “OpenAI Says Its Business Will Burn $115B Through 2029.” In other words, it will consume roughly the entire GDP of the Dominican Republic. That’s just $80B more than previously forecast. No biggie.
Motley Fool, Sept. 24, 2025 – “The Buffett Indicator and Shiller P/E Ratio Are in Rarified Territory.” Spoiler alert: The “rarified territory” is not a good sign. The article doesn’t mention “AI,” but we all know it’s the main driver behind those huge multiples.
Axios, Sept. 25, 2025 – “Nvidia’s $100B OpenAI Investment Fuels Bubble Fears.” Capital markets strained? No problem. Papa Nvidia cuts a check. The circular nature of the transaction prompts a flurry of memes of animals eating their own tails, and concerns about “interconnectedness” in Silicon Valley and what happens “if the AI bubble pops.”
People who have been through a few popped bubbles can easily see what’s going on here. There will be at least a temporary downturn. Investors will get burned. Hard-working founders and employees will end up broke or even bankrupt if they take on debt they can’t repay. The tide will go out, and you will see (as Warren Buffett famously quipped) “who’s been swimming naked” — in other words, who’s been faking success, or possibly worse. It’s familiar terrain. But how did we get here?
Whenever the “bust” happens, your social media feeds and email inboxes will be inundated with thinkpieces from technology experts, stock market gurus, and highly credentialed academics trying to answer this question. Here is my attempt as a lowly journalist to front-run all of them by explaining what I think is the blatantly obvious problem with the AI boom, staring everyone in the face: AI — in its current form, anyway — simply does not deliver enough value to ordinary people to justify its staggering cost.
Trenches Vs. The Clouds
Lest I sound like a Luddite or Grandpa Simpson shaking his fist at a cloud (meme below), let me be clear that I don’t hate AI. I find generative AI tools like ChatGPT, Claude, Gemini, Grok, and the like fascinating, fun, and even sometimes really helpful. I also agree with the notion espoused by many technology executives that smart people eventually do something really amazing with these tools.
For now, this is what generative AI platforms have provided for most people. At the positive end: help with research, brainstorming, line editing, compiling or assembling data, and other tasks requiring little higher-order thinking. At the negative end: We get loud warnings about diminishing job security, paradoxically paired with heavier workloads and frustration. That’s because all the productivity gains that were supposed to magically result from adopting AI for everything have simply not happened. Quite the opposite, in many cases (see the Harvard Business Review article on “Workslop”).
Meanwhile, seemingly unaware of the chaos in the trenches, Silicon Valley thought leaders still breathlessly compare AI to every great technological innovation in human history, waxing philosophical about its alleged ability to cure cancer, solve climate change, power “hypergrowth” across all industries, and make human labor obsolete. In the long term, some of these pronouncements may be true. But they all ring fairly hollow, for now, when generative AI can’t even reliably take drive-through orders at a Taco Bell.
From Fire, To Electricity, To Internet
Silicon Valley may insist that it’s on the cusp of the next great leap, similar to humankind’s mastery of fire or electricity, or the invention of the internet. But here’s the thing about true technological revolutions: They happen not just because the technology exists and it’s cool, and has the potential to do great things, but because it creates immediately recognizable value.
For example, it’s hard to imagine many Pleistocene cave-dwellers remained wary for long after being introduced to fire, stubbornly chewing raw meat in the dark. Fire’s benefits in a survival context are obvious: It provides warmth, light, makes food more digestible, and discourages predators from sneaking up and mauling you to death. No early human needed to hear a treatise on how fire’s transformative properties would allow our species to endure an Ice Age and spread throughout the globe. They just used it. The rest is history.
Likewise, electricity was an easy sell once it could be brought, at scale, into people’s homes. There were bankruptcies and missteps along the way — streetcar schemes, failed utilities, the costly DC-vs.-AC fight — but the value of electric lighting was instantly clear. A single live demo, the 1893 World’s Columbian Exposition in Chicago, captivated the public. More than 27 million people, over a third of the U.S. population, attended the fair to experience the glow of AC-powered incandescent bulbs. By 1900, most major cities had adopted electric streetlamps, gas lighting was in sharp decline, and the number of power plants across the country had tripled.
The internet was a bit trickier. Tethered to a bulky personal computer and a landline, it was seen as a novelty at first, used mostly for chatting on AOL, playing basic games, looking up trivia or celebrity gossip, or (of course), downloading porn (at glacial speeds). After the dot-com bust cleared away the hype, broadband took hold and e-commerce exploded. A tipping point happened: people realized they could research, shop, bank, and interact with the world without leaving their homes — and later, with the iPhone, not even their toilets.
To be sure, there are enterprise and niche applications of AI that are producing real-world value today: accelerating drug discovery, detecting fraud, helping coders write at least some boilerplate code, optimizing logistics, and streamlining certain back-office workflows. But revolutions aren’t built on niches. They happen when the value is obvious to everyone.
I could go on and talk about other paradigm-shifting technological leaps like the printing press or the steam engine, but I think my point is clear: What drives widespread adoption and fundamental shifts are the users themselves — what individual human beings find important, useful, and valuable, not what “thought leaders,” self-interested technology executives, tech bros, or pundits tell them they should care about.
My humble prediction: Until the focus returns to humans, AI’s power to reshape life as we know it will remain a hallucination.
