@unclebobmartin Starship could be the most important infrastructure breakthrough of our lifetimes, because it makes everything else in space cheaper to attempt.
It's not just a spacecraft, it's an epoch-making thing.
@esrtweet I think it’s mostly because the U.S. is a unique case. It’s probably the only country on the planet capable of thriving without the rest of the world. It’s basically a continent, not a country—and it practically invented and developed most of the tech that matters today.
@dhh You can't really live in a world that no longer exists—you can fantasize about it, but you can't live in it. Yet worlds don't disappear overnight; they usually shrink first. Sometimes their margins shrink faster, causing more people to fall off sooner.
@elonmusk Rather than responding with an immediate yes or no, governments could bring frontier AI companies, their CEOs, researchers and engineers into an open discussion to establish whether a genuine problem exists and, if so, what narrowly targeted intervention might address it?
@unclebobmartin@fadeleusdev China’s GPU constraints could push it toward getting more intelligence per unit of compute. The frontier may not be determined solely by who has the most powerful GPUs, but by who gets the most "intelligence" per unit of constrained resources. Necessity breeds invention.
Managers rely on static artifacts, but software moves in rates; only regular, subordinate‑owned 1:1s keep them aligned with real work. https://t.co/mIo27D7GyB
@dhh I tried it just for fun on an old Acer laptop with a dual core Celeron and 4GB of RAM, and I can say it's usable, works quite well, actually much better than Windows worked on that laptop when it was brand new.
Monoliths avoid network cost; microservices only pay off for real independence. Without that, you get a distributed monolith. Modularity and ownership come from discipline, not distribution. https://t.co/uENHyBmske
@elonmusk How do you get huge numbers of people excited to tackle ideas orders of magnitude beyond today’s tech — and actually deliver — when most instantly resist anything even 50% ahead?
@Grady_Booch Brains are likely computable machines. Minds aren’t. The ability to grasp abstractions, contemplate infinities, and experience thought points beyond Turing limits. Computation simulates behavior — it doesn’t explain the ghost in the machine.
Software teams chase visible technical activity because real outcomes are delayed and unclear; only redesigned incentives and falsifiable decision‑making can restore meaningful learning and reliability. https://t.co/GJtWkhnKsa
“People management” hides that managers shape environments, not individuals; incentives and constraints drive behavior far more than interpersonal skill. https://t.co/uvgNXyHXJK
Weak, delayed feedback in software pushes developers toward quick wins like complexity or churn. Strengthening real‑world accountability is the only way to rebalance incentives. https://t.co/mzeE9MDN69
AI lowers coding friction, but not the real complexity of business systems. SaaS isn’t dying—its value shifts to domain expertise, reliability, and long‑term accountability. https://t.co/hZ4UvfeZ22
"Training a superintelligent AI correctly" is flawed. True intelligence can’t exist in isolation—it emerges from networks of diverse minds with different experiences arguing and surprising each other. https://t.co/2Kj4ng9shm
My book, 𝐓𝐡𝐞 𝐂𝐨𝐧𝐣𝐞𝐜𝐭𝐮𝐫𝐚𝐥 𝐄𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞, is officially available on Amazon — and it’s now also accessible through wider distribution channels that make it orderable by bookstores in the United States and the United Kingdom.