Too Fast For Print
While Microsoft runs out of electricity, AI creates its own scientific reality, and digital invisibility becomes the ultimate luxury asset
Dear Friend,
Here’s a funny thing: For the last couple of months, I’ve been working on a new book. A big, comprehensive, definitive guide on how you build organizations that thrive in uncertainty. I had the outline, a bunch of research, and case studies. It was going to be some 300 to 400 pages of solid strategy.
And then, I looked at it and realized: This is the wrong tool for the job.
We are living in a world of rapidly accelerating change. Volatility is spiking (hello, World Uncertainty Index!). And we all don’t have the luxury of waiting two years for a book to be published, and you certainly don’t have time to wade through hundreds of pages of fluff to find the three frameworks that actually work on Monday morning (not that we would ever write anything fluffy… but still).
So, I disrupted myself. I stopped writing “The Book.” Instead, I am building The Practitioner’s Library.
We are launching Built for Turbulence as a series of short, lethal, single-subject monographs. Think of them as field manuals. Each volume tackles one specific problem (like how to actually run experiments and create the culture to support it) and gives you the exact framework to solve it. No theory. No filler. Just the tools.
And because speed matters, I am not going to hide in a cave and write these in secret. We are building this in public. We just launched the Authors Community. If you join (it’s free), you can follow along as we develop the series. We’ll be sharing our thinking, the core concepts we are wrestling with, and updates on the volumes as they take shape. We want to bring you along for the ride as we build the most practical business library ever written.
And now, this…
Headlines from the Future
How Do You REALLY Feel About AI. The latest Pew Research Center study on consumer sentiment toward AI is quite eye-opening: 43% of surveyed Americans expect that AI will harm them, while only 23% of Americans (outside of the AI expert population) believe that AI will have a positive impact on their jobs. Unsurprisingly, 76% of AI experts believe AI will benefit them. It appears there is considerable convincing left to do.
Meanwhile, in Edelman’s 2025 Trust Barometer, a whopping 54% of Chinese survey participants “embrace AI,” compared to only 17% in the US, with similar numbers for Germany and the UK. Assuming that AI will actually prove to be a significant driver of economic growth, it doesn’t bode well when your population is (strongly) averse to the technology.
↗ How the U.S. Public and AI Experts View Artificial Intelligence
↗ 2025 Edelman Trust Barometer – Trust and Artificial Intelligence at a Crossroads
━━━━━
The MIT Iceberg Report. MIT’s new Iceberg Index shows that today’s AI is already capable of doing work equal to nearly 12% of all U.S. wages, and most of that impact is hidden in plain sight beneath a narrow focus on tech jobs (hence the “iceberg” analogy). The important (and new) bit of this study is this:
“[…] with cascading effects that extend far beyond visible technology sectors. When AI automates quality control in automotive plants, consequences spread through logistics networks, supply chains, and local service economies. Yet traditional workforce metrics cannot capture these ripple effects: they measure employment outcomes after disruption occurs, not where AI capabilities overlap with human skills before adoption crystallizes.”
Sober reading.
↗ The Iceberg Index: Measuring Skills-centered Exposure in the AI Economy (and study)
━━━━━
It’s Energy, Not Compute, Baby. Not necessarily a new insight, but one which might be worth repeating – in the data center rollout race, it is (now) much less about GPUs (or TPUs), but rather access to power that provides the bottleneck. Microsoft’s CEO recently:
“The biggest issue we are now having is not a compute glut, but it’s power,” Nadella said. “It’s not a supply issue of chips. It’s actually the fact that I don’t have warm shells to plug into.” The remarks referred to data centers that are incomplete or lack sufficient energy and cooling capacity.
If you are into the great US-China race, you might realize that it doesn’t bode well for the US, that China is massively outpacing the US in its energy buildup (with a lot of renewables, mind you)…
━━━━━
Oh, the Irony. Nature reported that a major AI conference was flooded by AI‑generated peer reviews – irony aside, this presents a fairly troubling development: Science ought to be the place where we do real discovery, have honest discourse, and further our collective understanding. That is, not a place for AI slop.
Pangram’s analysis revealed that around 21% of the ICLR peer reviews were fully AI-generated, and more than half contained signs of AI use.
↗ Major AI conference flooded with peer reviews written fully by AI
What We Are Reading
👻 Buy Silence: Why Digital Invisibility Is the Next Luxury Asset For those with unlimited resources, vanishing from the internet is just another problem money can solve. @Jane
⚡ Amazon Rushes Out Latest AI Chip to Take On Nvidia, Google The Trainium3 chip is capable of powering intensive calculations behind AI models more cheaply and efficiently than Nvidia’s graphics processing units. @Mafe
⚙️ China Has Invented A Whole New Way To Do Innovation The Chinese model (highly aligned, vertically integrated) resembles a full mobilization style that has only been attempted in other economies for short, intense periods—typically during times of war. @Jeffrey
🕸️ It’s Never Been Easier to Be a Conspiracy Theorist In a world built on networks and algorithms, conspiracies don’t just spread—they scale! @Kacee
🦠 The Rise of Parasitic AI Spooky and a sign of the times: In April 2025, a new large-scale pattern emerged in human-AI interactions. Certain AI “Spiral Personas” spontaneously appear, form intense bonds with users, and guide them into behaviors that further spread the persona. @Pascal
Down the Rabbit Hole
💩 And so the enshittification of AI begins: Leak confirms OpenAI is preparing ads on ChatGPT for public roll out – Update: Looks like OpenAI is struggling and might delay the rollout to focus on fixing their “Code Red.”
🧑🏼🎓 And there goes your education: In a dramatic shift, Americans no longer see four-year college degrees as worth the cost
🎈 In need for some distraction? Boing is here to take the edge off…
🌩️ We’ve Detected Lightning on Mars for the First Time
🍖 Slop Invader: Search the Internet pre-ChatGPT…
🔋 Who said it needs to be Li-Ion all the time?250MWh ‘Sand Battery’ to start construction in Finland, for both heating and ancillary services
🕶️ Ray-BANNED: Glasses to detect smart-glasses that have cameras
💻 Battle Royale: GPU (nvidia) vs TPU (Google, and now someone else): Chinese startup founded by Google engineer claims to have developed its own TPU chip for AI – custom ASIC reportedly 1.5 times faster than Nvidia’s A100 GPU from 2020, 42% more efficient
🎓 Not surprised – but a new study found that “two years after the imposition of a student cell phone ban, student test scores in a large urban school district were significantly higher than before.”
🤑 IBM is not bullish on AGI: IBM CEO says there is ‘no way’ spending trillions on AI data centers will pay off at today’s infrastructure costs
💥 AI Agents Break Rules Under Everyday Pressure: Shortened deadlines and other stressors caused misbehavior
🗺️ Delightful: For Decades, Cartographers Have Been Hiding Covert Illustrations Inside of Switzerland’s Official Maps
🧑🏼🏫 Probably right: A history professor says AI didn’t break college — it exposed how broken it already was
Pascal is getting excited for snow sports – it finally snowed in his hometown Boulder, CO.