Peter Morales

Founder, CEO of Code Metal. Interested in development at the edge? DM.

[Peter Morales @PeterMoralesX]

“The future is already here. It's just not evenly distributed yet”

[Code Metal @Code_Metal_AI]

Working on formal methods, program equivalence, or verification? We want to celebrate your research and support what’s next 🚀

University faculty: apply for one of four Code Metal research awards of up to $40K each by Oct. 8.

https://t.co/BtvNpQgwb8

[Peter Morales @PeterMoralesX]

@sh_reya 100%

[Peter Morales @PeterMoralesX]

Excited to see people talk more openly about leveraging formal methods to validate software. At Code Metal, we deterministically autogenerate those conditions so that tools like Opus 5.5 can meet them.

Happy to see the market come alive here

[Boris Cherny @bcherny]

I used Opus 5.5 to formally verify the Claude Agent SDK using Lean. A couple short prompts = 16 PRs fixing various bugs and race conditions. Video attached.

TLA+ also works well. I sometimes combine Lean and TLA+ to look for issues around data flow, concurrency, and state mgmt.

I don't know either language well, but Claude is excellent at both. This approach is super useful for formally modeling your code and finding bugs that a human probably wouldn't have spotted.

Is formal verification the future of coding (or at least, bug finding)?

[Peter Morales @PeterMoralesX]

@sabarinathc One is marketing, the other is masquerading as public service ... both highly overestimate the capabilities

[Peter Morales @PeterMoralesX]

Much of what passes for AI safety today is PR fearmongering dressed up as science. There’s legitimate safety work to do, but “propagandist” is a more honest title for too many of its practitioners.

[Vishal Misra @vishalmisra]

The paper he is referring to is BitWhisper from 2015. The cpus had to be 0-40 cm apart, both had to be compromised and the bitrate was 1 to 8 bits PER HOUR.

That’s how misaligned AI will kill us?

Maybe we will all die of boredom waiting to decipher what they are doing.

[Chad Wahlquist @chadwahl]

Karp 16 years ago…….Sometimes you have to focus on the core issues and iterate against them before you release a product…….

“If you are iterating on a problem that you want to be important 3 years from now, it’s better to have engineers figuring out what the core issues are and then iterate against them.”

“We’re long on dealing with the most important problems that we can find, dealing with them in a productized way so that they scale for the client… and because we’re long on that and short on what happens in near term, we are not planning to hire salespeople.”

He has been consistent for a very long time

[Rohan Paul @rohanpaul_ai]

Jensen Huang completely destroys Jacob Coxon’s (ex-Anthrpic employee) 10% extinction prediction. no data, no scientific grounding, no defensible 10%.

"First of all, we shouldn’t [try to explain a 10% chance of extinction], because it’s made up.

These are well-educated people. They’re called researchers. Obviously, they’re working in a lab, and so the confluence of these words—and then the prediction—is alarming and troubling.

It shouldn’t be done. It’s irresponsible."

----

Full video on "All-In Podcast" YouTube channel (link in comment)

[David Sacks @DavidSacks]

“I think they're having AI psychosis… I'm not sure they're dispassionate critics, observers of what's happening.”

[Code Metal @Code_Metal_AI]

Some places are more fun to deploy than others...

[Rhea S @rhea_sideris]

Code Metal CEO @PeterMoralesX joins us at the @NYSE to talk about building verifiable AI-enabled code translation for mission-critical systems, an $80M Air Force contract, and going from a $250M valuation to $1.25B in just three months.

Full interview, we covered:
- What Code Metal is actually building
- A real-world look at Code Metal in action
- The $80M Air Force contract for
WarMatrix
- Going from a $250M to $1.25B valuation in three months
- Why raise another $125M while already generating revenue?
- What happens when AI gets better
- Why AI can't simply verify itself
- What happens to software engineering as AI develops
- Where Code Metal could be three years from now
- And... should we be scared of AI?

His answer to the last one: he's more scared of humans than AI...

[Rhea S @rhea_sideris]

@PeterMoralesX and I talking about @Code_Metal_AI at the @NYSE today!

[Rhea S @rhea_sideris]

Tomorrow is going to be a good one!!

@AlexaLiautaud
@BaijuBhatt
@PeterMoralesX
@PhilipJohnston

[Peter Morales @PeterMoralesX]

Big day for Code Metal. We’ve been awarded an $80M OTA agreement with the Department of War to advance WarMatrix, built with the U.S. Air Force, and bring AI to trusted modeling and simulation systems.

Thanks to Preston Fore at Fortune chatting with me about it. More ahead!

https://t.co/Al8pPfK3eL

[Peter Morales @PeterMoralesX]

@Andercot What an event, Andrew. Niranjan Hasabnis and I had a blast running between electronics stores to grab whatever we could to help teams make it to the stage with something working. It was a great time.

[Peter Morales @PeterMoralesX]

@reidhoffman The thread says nothing and only deepens my doubts about their sincerity. Their CEO’s past comments already showed they’re far from “the good guys” (apparently open source stops at AI).

[Peter Morales @PeterMoralesX]

Excited to announce our paper has been accepted to #AI4Sys. We benchmarked LLMs on predicting NVIDIA CUDA kernel performance—early results that pave the way for solid cost models in our GPU model optimization. This is part of our open research collaborations, expect more soon!

https://t.co/5YT8HZa3oy

[andi (twocents.com) @Nexuist]

i get the sense the people most bullish on vibe coding are people who have spent their careers talking to SWEs over 15-30min calls and haven’t quite grasped how many hundreds of daily decisions are hidden from them out of respect for time

[David Hinkle @Drachs1978]

Because we've made it illegal for people to tinker with RF without significant licensing no children play with RF. Because no children play with it very few grow into adults that pursue it for a career. As a consequence; the field progresses very slowly, easy things are occult.

[Machine Learning Street Talk @MLStreetTalk]

Jakob Foerster @j_foerst at @UniofOxford arguing that the AI community needs to avoid being goodharted by benchmarks.

[Peter Morales @PeterMoralesX]

@ycombinator @sdianahu Hey, check us out

https://t.co/kZJRaV4DWa

And whole heartily agree @sdianahu