Navier AI retweeted
agree with most of this. models will get good at design fast. but most of the time on a real program goes to building the thing, measuring it, and figuring out why it didn't work as expected
more sims won't remove that, but they can cut how many times you have to loop. @NavierAI
AI is on track to become a superhuman hardware designer, but design is only the beginning. The interface between bits and atoms is lossy, and turning designs into real hardware gets harder at every level:
1. Sourcing off-the-shelf parts that work together (a 2006 Falcon 1 engine test was delayed by one capacitor, in stock 6,000 miles away)
2. Custom parts, where a master machinist translates CAD into machine instructions
3. Assemblies, where every part can meet its drawing and still not fit
4. Outside vendors (even SpaceX has 591 employees with supply chain titles)
5. Doing all of it at a scale of 1,000,000
"Superhuman AI might be great at pure design, meaning at defining the platonically ideal part for a particular job, but struggle with 'design for manufacturing,' or finding a point design that meets the requirements while remaining easy to build/procure and inexpensive."
Full piece from a16z's Christian Keil on what it takes to turn AI-designed hardware into the real thing: a16z.news/p/superhuman-hardw…
You can just simulate things.
has @Tesla tested a backflip yet?
the inertia about the pitch axis is ~3-4x that of the roll axis so rotational energy increases by ~3-4x
however, the majority of the energy required still comes from the initial jump, so this should theoretically be possible
Something fun our team ran this morning: an engine firing into supersonic flow, the same problem a booster solves when it slows down to land.
One prompt to our agent. It set up the case, ran it on the GPU solver we’ve been building in house, and rendered the start-up.
This sort of sim used to be reserved for a CFD specialist. Now, these sorts of complex cases are becoming more accessible to everyday engineers.
More updates to come soon! 👀
Navier AI retweeted
the bots are grinding on this one.
immersed boundary layer solved on gpu with @NavierAI agents
This morning, the team ran a robot gripper and arm through FEA and Project Chrono, all inside Prism.
Structural analysis and multibody dynamics, one agent, one session.
Prism acts as one interface across every domain, with every file staying in context. When you move from stress to motion, the agent doesn’t start from scratch. It already has the geometry, the setup, and the results from the previous step. That is what makes multiphysics design loops practical.
More on Prism soon 👀
Over the weekend, we ran some 2D tank sloshing simulations with Basilisk in Prism, our agentic engineering platform.
Basilisk, one of Prism's pre-bundled solvers for free-surface flows, paired with our agent, allows for rapid experimentation.
Navier AI retweeted
Supersonic retropropulsion temperature field. Different angle on the same case run with @NavierAI's solver.
Heat comes from the hot nozzle exit gas + shock compression.
What fields or post-processing would you add here?
Engineering is deeply coupled. Our tools aren’t.
CFD, FEA, and GNC all live in silos. Every handoff = lost context.
Prism brings them into one workspace, one source of truth.
Multidisciplinary design, natively.
Learn more about Prism at navier.ai/prism
This GTO mission was built end-to-end by Stella in <30 minutes.
From trajectory design → production-grade Rust flight software → fully simulation-verified GNC.
When iteration takes minutes, you don’t run one sim, you explore hundreds.
Learn more at navier.ai/prism
At Navier AI, we're building the simulation layer autonomy depends on. Agent-driven GNC, CFD, and FEA. All built for the pace this budget implies.
Building the systems DAWG is about to fund? Let's make sure your simulation stack can keep up. Book a demo today:navier.ai/contact