Managing Partner at Autopilot (@hqautopilot). Investor @xAI, @replit, @fluidstack, @skyryse, @AyarLabs, @RadiantNuclear, @Mach_Industries Prev: @ARKInvest
Joined March 2021
- Tweets1.3K
- Following227
- Followers22.6K
- Likes11.5K
Will Summerlin retweeted
(1/6) From @GoogleDeepmind’s AlphaGo to AlphaZero, we saw machines find moves that reached beyond what we’d imagined for them. No one taught them those strategies. They emerged from reinforcement learning: playing, failing, and trying again under a single binary task reward for success or failure. Inspired by them, we couldn’t help but wonder: what might a robot discover if we gave it that same freedom to learn?
That question led us to AlphaNIST. Starting tabula rasa, with that same binary reward plus only a simple control cost, a single policy solves all 20 @NIST assembly tasks at a 98.2% success rate in simulation. More surprising still, AlphaNIST can complete 10 different tasks in a row, receiving that reward only once, after all 10 are done.
Just as AlphaGo’s strategies surprised its creators, AlphaNIST discovered dexterous strategies we couldn’t have come up with ourselves.
Paper: arxiv.org/abs/2610.13354
Website: alpha-nist.github.io
Another way to frame this: would you rather go to Italy during shoulder season, with lower prices, better weather, and fewer crowds, or battle the crowds and heat in peak summer?
Playing different games
In venture, there are a few themes at any given time that are broadly consensus. Given how much capital is sloshing around the ecosystem, a reasonably strong team building within one of those themes will get funded.
This is the game the large platforms are playing: identify a consensus theme, back anything “good enough,” and then double down on the winners with large late-stage checks. It’s private tech beta, and that’s fine. With strong tailwinds, beta is great.
That’s not a winnable game for smaller firms. The way to play as a smaller firm, in my view, is not to figure out which direction the wind is blowing today, but where it will be blowing in 6-18 months. Identify themes before they are broadly consensus, build conviction, and partner with the best teams you can find. In most asset classes, you generate alpha by being non-consensus and right. The same is true in venture.
This is often conflated with being at an earlier-stage. Sure, it can work at pre-seed, but it can also work at growth. I’d argue that growth is actually a more attractive place to play this game, given relatively fewer capital providers and current headwinds around pro rata that hamper early-stage investors’ ability to defend or build ownership as a company becomes consensus.
Looking at our own data, our best returns have consistently come from early-growth investments that went from “non-consensus” to “broadly consensus” in 6-18 months. The rounds went from relatively uncompetitive to highly competitive and sharp-elbowed.
Win with your brain, not your brand.
Playing different games
In venture, there are a few themes at any given time that are broadly consensus. Given how much capital is sloshing around the ecosystem, a reasonably strong team building within one of those themes will get funded.
This is the game the large platforms are playing: identify a consensus theme, back anything “good enough,” and then double down on the winners with large late-stage checks. It’s private tech beta, and that’s fine. With strong tailwinds, beta is great.
That’s not a winnable game for smaller firms. The way to play as a smaller firm, in my view, is not to figure out which direction the wind is blowing today, but where it will be blowing in 6-18 months. Identify themes before they are broadly consensus, build conviction, and partner with the best teams you can find. In most asset classes, you generate alpha by being non-consensus and right. The same is true in venture.
This is often conflated with being at an earlier-stage. Sure, it can work at pre-seed, but it can also work at growth. I’d argue that growth is actually a more attractive place to play this game, given relatively fewer capital providers and current headwinds around pro rata that hamper early-stage investors’ ability to defend or build ownership as a company becomes consensus.
Looking at our own data, our best returns have consistently come from early-growth investments that went from “non-consensus” to “broadly consensus” in 6-18 months. The rounds went from relatively uncompetitive to highly competitive and sharp-elbowed.
Win with your brain, not your brand.
Will Summerlin retweeted
Full conversation here: youtu.be/YgYiF86h0yU?is=PtLN…
BREAKING: Patrick Collison just told the Valley exactly where the next Stripe gets built.
And it's not an AI lab.
Asked what he'd build if starting from scratch, Collison pointed directly at legacy vertical software — decades-old systems still running healthcare, logistics, construction, finance.
Entire industries. Still on 1990s infrastructure.
Every ambitious founder is chasing foundation models right now.
Collison is pointing at the opposite end of the stack.
The most defensible businesses often come from the places everyone else decides are too boring to touch.
Is the Valley sleeping on the biggest opportunity in software right now?
Will Summerlin retweeted
Skyryse is partnering with @BAESystemsAus to advance software-defined autonomous flight. Together, we’ll explore next-generation aircraft opportunities across piloted, optionally piloted and autonomous flight. bit.ly/4AqsDvG
Will Summerlin retweeted
One lesson it took me way too long to internalize:
Cool technology is not valuable in and of itself.
Cantos founder and managing partner @IanRountree explains why a technological breakthrough does not automatically justify a massive valuation.
“What happens if you have two fusion energy companies succeed? … Why are you both going to be worth $10 trillion? You’re going to be in a price war, potentially.”
RT @RadiantNuclear: Thanks for having on to discuss the Janus program and our $750M award! @CNBC @torishiv cnbc.com/video/2026/08/27/to…
🚀🚀🚀
The @USArmy has awarded Radiant $750 million to deliver 15 Kaleidos microreactors through the Janus program, one of the largest contracts @DIU_x has issued to date.
The Army launched the Janus Program to strengthen the energy resilience of military installations and mission-critical infrastructure. Radiant was chosen after extensive evaluation of reactor design, technical maturity, manufacturing readiness, deployment strategy, and commercial viability. Radiant’s 15 microreactors will enable the Army to deploy nuclear power across multiple locations. In tandem, the operational data acquired through the Program will accelerate commercial deployments across a global footprint of commercial customers.
Learn more: radiantnuclear.com/
Will Summerlin retweeted
The @USArmy has awarded Radiant $750 million to deliver 15 Kaleidos microreactors through the Janus program, one of the largest contracts @DIU_x has issued to date.
The Army launched the Janus Program to strengthen the energy resilience of military installations and mission-critical infrastructure. Radiant was chosen after extensive evaluation of reactor design, technical maturity, manufacturing readiness, deployment strategy, and commercial viability. Radiant’s 15 microreactors will enable the Army to deploy nuclear power across multiple locations. In tandem, the operational data acquired through the Program will accelerate commercial deployments across a global footprint of commercial customers.
Learn more: radiantnuclear.com/
AI researchers in SF are doomsday prepping in anticipation of AGI.
The average Fortune 500 employee has used Microsoft copilot a few times.
Wildly varied views of the future.
Will Summerlin retweeted
Incredibly important (and well written) article that cuts through the hype. Nevermind data centers, American hospitals, grocery stores, warehouses, army bases and more need an energy upgrade. @RadiantNuclear recognized this long ago. Hats off to the team for the tireless, diligent work of making it a safe, practical, near-term reality.
IRR (because of the drawdown structure) makes comparisons to the S&P 500 a bit more nuanced than MOIC.
That said, you shouldn’t invest in a venture fund unless you believe it can return >5x MOIC and >30% net IRR.
We’re in a funny time for investors where you can be >99% confident you’ll be right over the medium term, and as a result, knowingly pass on a short-term 10x layup.
Some of the highest quality businesses are radically supply-constrained (i.e. more customer demand than they can serve).
A proxy is marketing spend - the less they spend on marketing, the higher the quality of the business.
There are a lot of great businesses that you’ve never heard of / have websites from the 90s.
You can make a lot of money in venture by being smart (contrarian and right) or by being cool (getting access to hot consensus deals). Neither is wrong, but know which game you’re playing.
I’ve found that it’s an easy trap to feel like you need to have an opinion on every company, stock, trend, etc.
The most unique insights, and the best investment opportunities, are most often derived from a combination of life experience, prior (often unrelated) areas of research, and relationships.
You only need 1–3 fat pitches every year to be a truly extraordinary investor.
Stay curious, but don’t succumb to the pressure of trying to be right about everything. Just say no to everything but the fat pitches where you have a unique insight and overwhelming conviction.
Will Summerlin retweeted
This is the best way to end America's entrepreneurial culture.
Most founders I know endure 100+ hours of work/week because they are obsessed with their mission and the impact they can have on the world.
They make < $100K / year, fly economy, and live modest lives. The average American doesn't want their life.
Just because their company's valuation is growing doesn't mean they are "wealthy".
When you impose a > $100M tax on illiquid stock, they will declare bankruptcy or flee the country.
It's time for a national billionaires tax and a new social contract.
10% of Americans own two-thirds of the wealth. Wages have stagnated. The cost of living has skyrocketed.
The system is fundamentally broken.
The federal tax code, a corporate code, and an inheritance code were written for a different set of Americans.
It’s time for an economic reset.
substack.com/@gavinnewsom