Introducing our first-ever softcover: a lighter, more portable edition of Poor Charlie's Almanack.
The hardcover has sold more than 400,000 copies and we've spent years figuring out how to make it even more accessible.
We hope it's the first of many.
Kelly Johnson, the legendary Lockheed engineer behind the SR-71 Blackbird and the U-2 spy plane, ran the company's most secretive division: Skunk Works. The planes he built defined an era.
But he carried one regret to the end of his life. He never managed to build a military aircraft that could take off and land vertically.
Brandon Tseng, co-founder of @shieldaitech , once shared a clip of a Johnson interview on X. And Shield AI's X-BAT, an AI-piloted fighter drone that takes off and lands vertically, is now making that exact idea real.
So I opened my interview with him from that post: the problem Johnson couldn't solve back then, how are you solving it today?
His immediate reaction: "I love that you brought that up."
He told me the X-BAT concept is far from new. The US military built a vertical takeoff and landing aircraft called the X-13 back in the 1950s, and it was nearly identical to what X-BAT is doing today.
Two things killed it. The engine lacked thrust, so the aircraft could barely carry any payload and could only fly for 45 minutes.
And it relied on a human pilot. As Brandon put it, you're asking someone to look straight up at the sky while landing a plane vertically back onto an exact spot. That's nearly impossible.
76 years later, two changes make it work. The engine is now the GE F-110, the same one that powers the F-16. And the pilot is now AI.
Then he made a point that stuck with me: a lot of old ideas had the right direction, the technology just hadn't caught up.
AI algorithms were already being written in the 80s and 90s, but there wasn't enough compute. Rockets that could land were discussed in the 60s and 70s, then largely forgotten, until Elon Musk made them real 40 years later.
In this episode, we talk about how an idea shelved for 70 years is finally being built. New episode out now.
Cool bit of history my HBS professor found.
A media interview asking Kelly Johnson (legendary Skunkworks engineer and leader) what his biggest disappointment was and him stating it was not getting their VTOL airplane to work.
Am pumped that @shieldaitech gets to build and make XBAT operational.
Little backstory here — weeks before sitting down with Brandon, I'd just finished Project Maven by @KatrinaManson (great read btw).
One detail stuck with me: @PalantirTech's software once saved him in Afghanistan during his SEAL days.
That's exactly why I brought it up in our interview.
Shield AI co-founder @brandontseng2 on use of Palantir $PLTR software as Navy Seal and future inspiration:
(Palantir documentary OTW)
“Palantir organized all of the intelligence information that we had and made it digestible so we could make better decisions”
“Gave us significant activity around IED placements. Was able to use this information to inform commanding officer”
“Palantir trail-blazed the defense industry”
@shieldaitech co-founder @brandontseng2 : "I've lost a lot of years of my life stressing out."
A few weeks ago I interviewed Brandon Tseng, co-founder of the defense startup Shield AI.
My last question stuck with me: after 11 years, 30 VCs saying no, and now a 1.5 billion dollar Series G, what do you wish you'd known on day one?
He paused, looked up, then said:
"Because you stress out a lot. I've lost a lot of years of my life stressing out about these problems we had to solve in year one, year two, year three, year four. If you have the right culture, the right mindset and team, you're gonna solve those problems. They can be painful and stressful, but don't worry about them too much."
👇 Full interview in the latest episode.
I've read The Power Law more than ten times, and The Infinity Machine the day it landed.
Getting an hour with @scmallaby for the podcast was a conversation I'd been waiting for since the book was announced.
One story stuck with me:
In 2018, AlphaFold's team lead told Hassabis it was over. They'd hit the ceiling. Time to move on.
Hassabis didn't argue with the data.
He sat in the room with the researchers and listened to how they talked to each other. The ideas were still flowing.
That room was his signal to keep going. Three years later, AlphaFold won the Nobel Prize.
👇 Listen to the full episode
https://t.co/4Q2v3mOvUg
The founder of one of America's most important defense companies just issued a chilling warning that should stop you cold: Taiwan is not China's final destination. It is merely the opening move.
Palmer Luckey, founder of Anduril Industries, the autonomous weapons powerhouse building tech for the U.S. military, made the terrifying reality plain. America cannot outmanufacture China. Beijing's industrial capacity in missiles and warships already dwarfs the U.S. annually. But Luckey argues that matching them weapon-for-weapon is the wrong metric. The strategy must be absolute denial, making an invasion of Taiwan too costly to ever attempt. "There's a world where they have a hundred times more ballistic missiles than us and we still have enough between us and our allies to deny them access to Taiwan."
But the most alarming part of Luckey's warning is what happens if Taiwan falls. He isn't just worried about Taipei; he is terrified of what comes next. "If they take Taiwan, history and the words out of Xi Jinping's own mouth show they're immediately going to hop over to Okinawa, part of the Philippines, maybe part of Vietnam."
Luckey exposed a critical detail the public is completely missing. Xi Jinping has personally targeted the Ryukyu Islands, the Japanese chain including Okinawa that sits between Japan and Taiwan, by citing 650-year-old archives claiming they were a Chinese tributary state. As Luckey bluntly interpreted: "He's not saying this because he loves wandering the national archives. They are constructing a national narrative that would allow them to convince their people to start with Taiwan and end with a lot of our Pacific allies."
This isn't theory; it is documented. Chinese state media and academic journals are already actively questioning Japan's sovereignty over Okinawa. It is the exact, deliberate pattern of narrative construction the CCP deploys before it strikes: establish a historical grievance, build public consensus, then move. Beijing has already sanctioned Luckey and Anduril for supporting Taiwan's defense because they know he sees their playbook clear as day. Taiwan is just the first domino.
#UnveiledChina #TaiwanInvasion #PalmerLuckey #Anduril #NationalSecurity #Okinawa #CCPPropaganda #IndoPacific #Deterrence
I’ve always believed the No.1 application of AI should be to improve human health.
That work started with AlphaFold, and now at @IsomorphicLabs with the mission to reimagine drug discovery and one day solve all disease!
We are turbocharging that goal with $2.1B in new funding.
Three books that give permission and guidance for why and how to be deeply optimistic all have "infinity" in their titles:
INFINITE IN ALL DIRECTIONS (1988), Freeman Dyson
THE BEGINNING OF INFINITY (2012), David Deutsch
THE INFINITY MACHINE (2026), Sebastian Mallaby
Celebrating 36 years of discovery with the Hubble Space Telescope 🔭
Collected on April 23, 2026, by one of Vantor's WorldView Legion satellites, this remarkable non-Earth image showcases Hubble from just 61.8 km away—an incredible perspective of one of humanity’s most iconic scientific instruments. With a space sample distance of 4.0 cm, Hubble’s signature cylindrical body, gleaming thermal shielding, and extended solar arrays are clearly visible, along with the open aperture door at the front of the telescope.
For over three decades, Hubble has expanded our understanding of the universe—delivering breathtaking imagery and groundbreaking science that continue to inspire.
Proud to support the technologies and teams that make moments like this possible.