Founder of @ValstadShip. We're building the machine that builds the ships.
Austin, TX
Joined July 2021
- Tweets5.3K
- Following1.1K
- Followers9.7K
- Likes6K
Pinned Tweet
We're hiring an Operations Lead!
Come help us change the way ships get built in America.
valstad.com/careers/operatio…
Lots of new followers, so I should probably explain what we're doing at @ValstadShip:
We're decomposing the common operations required to build ships, and building autonomous systems to carry out those operations at least 10x faster.
Let's make America competitive again.
Testing the new welding sensor package we assembled & integrated in-house.
Weld camera + leading and trailing lasers.
The dataset we’re building is incredible. Will be able to completely automate inspection as well.
Counter-offer: sell me Mercor for $500k.
I am available for a 15-minute call this week to discuss.
Two huge hires joining us in the next few weeks:
* Sr. Welding Technician from Northrop Grumman Marine Systems
* Sr. Welding Engineer from General Dynamics NASSCO
Building a GOATed welding automation team.
I actually think this approach maps much better to industrial robotics.
In an industrial setting, you want deterministic, accurate and fast. An 86% task completion rate is unacceptable.
“Skills” that are just testable code will be much better suited to this environment than squishier models, and don’t suffer from the data scarcity problem.
Basically you get all of the benefits of the world’s best LLMs coupled with the reliability of rock-solid classical approaches.
We are open-sourcing V1.0 of our agentic robotics framework, ASPIRE, with full real robot skill discovery and self improvement capabilities! On the day Astra arrived, our lab switched the backend model and saturated all the benchmarks to ~99% over night.
And the frontier is already on the move:
- Opus 5.5 slightly edges out Astra on our robot simulation bench.
- Sonnet 5.5 hits the Pareto frontier of cost vs. robot capability, $12/task.
We first introduced ASPIRE and /robot skill library in June, way before the big models could reliably do robotics. A rising frontier tide lifts all boats, but Astra was a 0->1 tsunami on many tasks.
Now you can also feel the Physical Bitter Lesson at home like we did on that fateful day ;)
links in thread
Cybercabs everywhere in Austin - I’m in one and can see two others right in front of me.
Lots of people trying them for the first time, too - everyone seems puzzled by the doors at first 😂
Hell yeah. Competitive open-weight models from US companies.
Meta is cooking.
Today we’re releasing Muse Spark Open, the most capable US open-weight model to date.
Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning.
Key Model Specs:
→ 816B total / 42B active parameters
→ 1M token context window
→ Natively multimodal across text, image, video, audio
Automated weld procedure development - how much interest would there be in this as a service?
Tuning weld parameters can be a PITA, but we could test potentially hundreds of combinations in a short amount of time on our robotics platform.
I read the article.
The argument they make for wealth taxes is extremely flimsy.
When faced with overwhelming evidence that wealth taxes fail miserably everywhere they've been tried, their response is "But this time it's different!" (it's more strict! applies to fewer people! it's retroactive!)
They confuse wealth with GDP in a way that distorts how much tax revenue would actually be raised.
They don't even attempt to zero-base the budget and prove that money is well spent, it's simply assumed that more taxation + more spending = better.
The Laffer Curve doesn't exist as far as they're concerned.
They acknowledge that it's likely unconstitutional federally, but may not be at the state level - which would definitely result in capital flight, although they deny this too.
They don't attempt to quantify, *at all*, the massive positive contribution these companies make to state coffers through payroll, income, property, and other taxes, nor through reinvestment of capital into other startups.
They claim it's a one-time tax, but openly admit they'd like to make it permanent - and that the original version targeted much lower levels of wealth but that targeting billionaires was more politically popular.
They even say the quiet part out loud:
'Tapping that wealth, though, was an obvious solution to an urgent problem. “It’s the Willie Sutton thing,” Regan says. “Why do I rob banks? Because that’s where the money is.”'
The academics don’t actually think about what will happen to startups. They also have no idea how new innovations occur and how much of a miracle that is for CA.
Their ignorance results in misguided policy that will kill the golden goose. No on 40
nytimes.com/2026/10/14/magaz…
