It's time for an AI Solarium Project. Why? Current "AI safety" debates straw man against just two outcomes: (i) a ground stop or pause on research/ development (as if stopping will somehow allow us to imagine all possible end-states and engineer to them), or (ii) acceleration, because "China" Reality: It is unmistakably clear that we are climbing the escalation ladder I am not sure there is a single example of humans understanding that a scientific possibility exists and not working to see if it is possible. From weapons of maximum power to medical breakthroughs to information superhighways and attention algorithms, if you can dream it, someone is working it. A Solarium Project gives America the best chance at controlling the outcome, leveraging the incentives of the market without sacrificing our future. What is a Solarium Project: Project Solarium was a top-secret national security exercise convened by President Eisenhower to create America's Cold War strategy against Soviet expansionism. The project used 3 independent task forces to test competing policy options and present their paths to the NSC. How it Would Work: The CEOs of the three biggest AI labs, @sama, @elonmusk and @DarioAmodei would accept commissions or otherwise be brought in for up to 24 months under Direct Hiring Authority. They would each draft 100 colleagues / teammates to bring with them. Each would be physically separated and forbidden from collaboration... perhaps at the original Manhattan Project sites (perhaps Los Alamos, Oak Ridge, and Hanford as a callback). Under the umbrella of secrecy, and with a classified (but sufficiently large) budget, they would be encouraged to accelerate to build the most capable AIs imaginable. They would not know the approaches / strategies the others are pursuing. They would have access to data that no one else has. The commercial benefit (to prevent VCs from crying bloody murder) would be their home offices retaining commercial rights to any/ all technologies they develop when declassified. Opposition: - "Regulatory Capture." Nope. This is about scientific leadership. Strip away the commercial considerations, which are arguably slowing progress at all 3 major labs. - Civilian Control of Technical Architecture: in America, civilian control of technical experts is something of a democratic ideal. Civilians control the military. Generalists run great companies. There's a difference between technical experts controlling their development trajectory and their obligation to communicate implications to others. - Resources: Commercial firms have been able to attract private capital. That could continue under this framework, given the commercial rights. - Is this nationalizing OpenAI, Anthropic, XAI, or any other commercial entity. No. It's giving the top experts in the world the protection and resourcing they need to go as fast as humanly possible without recklessly endangering our nation - and therefore - humanity. This is about as far from my free market priors as I'd hope to go, but "societal conversations" simply don't occur on their own, and I'm fairly confident we can pull this off without sacrificing economic progress.

Brian Taylor spent 15 years building satellites at @SpaceX, @amazon Kuiper, and @LoftOrbital. Every one of them was designed to burn up. So he started Lux Aeterna to build the ones that come home. Check out my full conversation with Brian on this week's #CrossingTheValley @SpaceNews_Inc

Today on Crossing the Valley... @method_security CEO and c-founder @___sjones. - How to raise a seed round from @a16z - How to build conviction in a market - How the "Mythos Moment" has upended the cybersecurity industry -...and a whole lot more Check out the full episode and subscribe for more!

Two guys sat down for a first check discussion and discovered they were life journey twins: - Both got into West Point. - Both left. - Both became Marines. One built a private military company. The other is building the brain for autonomous warfare. Today @paigecraig backs Philong Duong in building @NODA_AI_atx ... check out this episode of Crossing the Valley

The slowest gate in defense acquisition is test & evaluation. 3 years on a range. $2M a flight. It's where good ideas die. @Nominal_io's @CameronLMcCord built a company that turns that gate into a growth engine. On this week's Crossing the Valley: it's time to stop arguing speed vs. quality. Make risk a number.

When a C-130 loses a panel, the @usairforce had two options: 1) Wait 4 years for new tooling. 2) Go to a junkyard, find a wrecked C-130, and cannibalize the part. The workhorse of the U.S. military was being kept alive by scavenging. Not anymore. Today, @MachinaLabs_ robots are forming these panels in a week. New #CrossingTheValley episode with @EdwardMehr out now👇

Can confirm... you should read this

i kinda thought the free palestine folks mightve been celebrating khamenei's death. dude kinda sucked at "opposing israel" if we're being honest.

How do you get up to speed as a nontechnical teammate joining a hard tech startup? @ursamajortech Chief Business Officer Ben Nicholson shares his perspective on this week's Crossing the Valley.

What should young people be studying in school right now? @ursamajortech Chief Business Officer Ben Nicholson explains why data and analytics are top of his list

Too much focus on new tech development. Not enough on tech insertion in critical industries. @geckorobotics Chief Product Officer Troy Demmer explains

Deploying unproven technology to a nuclear silo isn't always an easy sell. That didn't stop @geckorobotics Chief Product Officer Troy Demmer explains how they got to yes on this week's Crossing the Valley

The future of warfare is now. Drone swarms and AI-enabled capabilities will be embedded into our entire fighting force. The @DeptofWar will be AI-FIRST.

Honored to be a small part of this with @usnda_org ... and if that footage looks familiar, stay tuned for more... ;)

Why take a company like Maxar private? @vantortech CEO Dan Smoot lays it out, plain and simple, in this week's Crossing the Valley recorded at the Reagan National Defense Forum! Full episode now live!

I asked @vantortech CEO Dan Smoot why rebrand an iconic company like Maxar? He lays it out, in this week's Crossing the Valley. Live tomorrow...

Fare thee well, @torybruno Getting to spend a launch day together this spring was a privilege. Clearly a deeply knowledgeable, patriotic leader.

Important

This week's Crossing the Valley: Nathan Diller, Head of Aerospace & Defense for @Divergent3D We discuss: 00:00:00 Introduction to Divergent and DAPS system 00:02:00 Nate's background and career journey 00:03:00 The Divergent vision for digital manufacturing 00:05:00 Why digital manufacturing matters for America 00:09:00 From Pentagon to Congress to private sector 00:12:30 Go-to-market strategy and partnerships 00:18:00 General Atomics partnership 00:22:00 Parts qualification process 00:25:00 Customer acquisition and sales approach 00:27:00 Family business culture at Divergent 00:28:30 Vision for weapons at volume and prosperity 00:29:00 SBIR program and government shutdown 00:34:00 State of American manufacturing landscape 00:36:00 Contact information and closing @KevinCzinger @lukasczinger

MANUFACTURING TWITTER: Small US manufacturing co's (20-50ppl) that do ANY defense work and want to grow, LISTEN UP I'm building a pilot program to get you CMMC lvl 2 Compliant at NO UPFRONT COST and pack you w/ as much work as you can handle Retweet to save 🇺🇸 DMs are open