Amazing turnout in Dillon, Montana! 🙏 Thank you to everyone who came out for the book signing and Q&A on June 21st. The energy was incredible! Missed it? Catch the full Q&A: youtu.be/u-1jcFpggC0?si=cUED…

Thank you 🙏

The AI changes happening right now are making it much easier for bad actors to break into accounts, steal money, and lock people out of their own lives. Most people won't notice until it happens to them or someone they care about. I don't want that for any of you. You don't need to understand the tech. Just do these basic things this week: 1. Update everything. Phones, computers, tablets, and apps. Turn on automatic updates so it stays current. 2. Stop reusing passwords. Use the free password manager that came with your phone (Apple or Google) or something simple like @Bitwarden or @1Password. Give every important account its own password. 3. Turn on extra login protection on email and banking. Prefer an authenticator app or passkeys over text-message codes if you can. Save any recovery codes somewhere safe first. 4. Never give out login codes or download anything from a call, text, or email you weren't expecting -- even if it looks official. 5. Back up your important photos and files somewhere separate from your main devices. 6. If you need to share files with people and you use a Mac, consider downloading qURL desktop: qurl.link (tap "send a file"). 7. If you have an iPhone or Mac, look up "Lockdown Mode" and consider turning it on. It limits a few conveniences but closes a lot of doors. If that list feels like a lot, just do #1 and #4 today. Those two alone stop a huge percentage of the problems. This isn't theoretical anymore. The people who get ahead of it might be fine. The ones who wait usually aren't. Do these for yourselves.

We received 3 million attack attempts in 10 days. Nice try!

New announcements are up at treasure.quest Starting today, you'll also notice a change in how the site works. These are the first steps toward protecting the treasure hunt and Beyond the Map's Edge in the era of AI superintelligence. This hunt has always rewarded human curiosity: your research, your instincts, your time in the field. I intend to keep it that way. There's more to come. Head over and read the latest, and if anything gives you trouble, let me know in the comments.

I've helped ship software used by hundreds of millions of people. None of it matters to me as much as what I'm building now. The list: → Apple Pay v2 → CrowdStrike NG-SIEM → Disney: helped launch a plethora of experiences all across the company → Internet Explorer, back when it was the world's most-used piece of software. Yes, that Internet Explorer. Get the jokes out now. I'll wait. Here's why qURL tops that list. For 30 years, the internet has run on one rule: if it's online, anyone can find it. Every bot, every scanner, every attacker. qURL reverses that rule. Your app, your server, your AI agent's tools stay invisible until someone you chose shows up with the right credential. Everyone else finds nothing there to attack. As agents take on more of our work, they'll need to reach private things without exposing them to the whole internet. I believe the future of the internet is invisible and ephemeral, and qURL is how we get there. If you've ever opened a port or a tunnel just to show someone a half-built app, try this: qurl publish https://localhost:3000 What would you keep private if invisible were the default?

AI agents just sprayed 2,900 logins a minute through a compromised vendor into thousands of Microsoft 365 tenants. Decade-old bugs in PHP got turned into working exploits with models. Sites went dark. A popular AI tool got a second malicious release this year that stole cloud and AI keys. None of this was sophisticated. It was just visible. Scanners, agents, and password sprays still work because almost everything still answers when you knock. We keep bolting on faster detection and response while the first step (finding something to attack) stays free and automated. That sequence is losing. The only thing that actually changes the math is making the infrastructure that shouldn't be public stop answering at all. No open ports. No banners. No response until someone proves who they are with a cryptographic knock. Temporary links when something does need a window. Everything else is just trying to outrun machines that don't sleep. LayerV.ai is the way to do the hide-first part. The rest of the industry still mostly optimizes the old sequence.

An unused crypto address is already a bunker. Bitcoin and Ethereum addresses are hashes. The public key stays hidden until the first signature. After that, it sits on-chain forever. If someone later recovers private keys from public keys, every address that has ever signed is a target. Every address that has never signed is not. That used to be a quantum problem, dated in decades. It is now a math problem with an unclear clock. This week OpenAI posted 722 manuscripts from an unreleased model: claimed breaks of the n log n integer-multiplication bound, the 3SUM conjecture, and a May disproof of the Erdős unit-distance conjecture as the warning shot. Most of it is not peer-reviewed. Some of it will be wrong. The planning assumption does not require it to be right. Elliptic curves are full of structure: Schoof, Frobenius, pairings. Hashes are built to have almost none. A jump in mathematical search hits the structured object first. ECDSA is not broken today. It is also no longer obvious that a classical break waits for a quantum computer. The hedge is exposure control rather than a new wallet. Large holders move the bulk of funds to addresses that have never signed. When one of those addresses does sign, the remainder moves to a fresh one, same seed. No new cryptography. A rushed migration is worse than a slow one, because the stampede creates the signatures you are trying to avoid. The same rule does not stop at cold storage. A published public key is an open port: exchange signers, oracles, L2 councils, SSH, code-signing, TLS. Post-AI security will be less structure, and less advertisement of whatever structure remains.

Anthropic’s estimated run rate has stalled at about $77 billion after a climb from roughly $9 billion in late 2025, up only 1.8% quarter over quarter. That is the first visible crack in the assumption holding up the AI trade: that frontier revenue keeps compounding fast enough to earn back the buildout. NVIDIA, Microsoft, Amazon, Google, Meta, and Oracle are not priced on Anthropic’s books. They are priced on that ramp. Flat at $77 billion is still enormous. The growth rate is what a reported ~$2 trillion IPO valuation and the capex plans require. Two reads remain open. Capacity may be the constraint, which only delays the ramp. Or customers may be refusing frontier prices, which hits payback math directly. Some estimates already put a 1 GW AI data center at $50–100 billion to build against $10–30 billion of annual revenue, with GPU life closer to four years than ten. The chart does not reprice the complex tomorrow. It is the question the roadshow has to answer.

AI just proved the equations that run the physical world can explode. Navier-Stokes is the math behind air over a wing, blood in your veins, storms, oceans, engines, climate. For 90 years nobody knew if a smooth flow could blow up to infinite speed in finite time. An AI system found the crack in days. A vortex tightens, stretches like spaghetti, and the math itself breaks while energy stays finite. This is the model we use to design planes, forecast weather, and keep people alive. The equations that describe flowing reality just got a known failure mode. The machine that found it can now be pointed at everything built on fluids. That's the leap. Keep up.