Larry Diehl

Founder @ColimitAI. Formal Verification & Neuro-Symbolic AI. Formerly postdoc @uiowa and PhD @Portland_State.

[Larry Diehl @larrytheliquid]

It's been over a decade since I've given a public tech talk, so I was super happy when @eatonphil said there was an opening at NYC Systems.

This talk is on a language called Dafny, which IMO does the best job of making formal verification accessible to programmers without previous experience in the area.

I hope this can act both as a quick intro to Dafny, as well as reference material to understand what's going on when you move beyond the basics.

[Larry Diehl @larrytheliquid]

this one is dedicated to @rob_rix

[Colimit @ColimitAI]

Types

[Larry Diehl @larrytheliquid]

@Keleesssss niiice, congrats!

[Larry Diehl @larrytheliquid]

@zeeg in some eval i even saw a 5.6 Sol Ultra, so the 3 tiers isnt even the end of it =/

[Larry Diehl @larrytheliquid]

@mitchellh if youd pair that with a harness that could formally verify your code, you might even be able to cut out the frontier model 🤔

[Larry Diehl @larrytheliquid]

@headinthebox ive seen this in a talk this smart guy i know once gave https://t.co/8idvgH4NVk also reminds me of this other one https://t.co/PQxUOl1Swd

[Larry Diehl @larrytheliquid]

@devabram Wait till you see view patterns

[Larry Diehl @larrytheliquid]

our tasting menu is almost ready to be served

[David Abram 🐊 @devabram]

Bros, @ColimitAI sounds extra tasty.

Operative @tritlo is working hard on this one.

[Colimit @ColimitAI]

Formal Spec-Driven Development

[Larry Diehl @larrytheliquid]

@GeoffreyHuntley @heavybit shortly? thats over a month away, we could have agi by then! =p

[Larry Diehl @larrytheliquid]

@glcst Insane work, you're getting rid of all the objections against using sqlite, congrats glauber!

[Larry Diehl @larrytheliquid]

One thing that I noticed about the original implementation of Devin was that it broke up its plan into a sequence of line numbered steps, and some steps had gotos (much like asm). This was over a year ago, and at least for smaller models this was very effective. I copied the strategy in one of my products as a latency and cost savings measure to be able to use small models in certain parts

[Prof. Anima Anandkumar @AnimaAnandkumar]

We’re excited to release TorchLean which is the first fully verified neural network framework in Lean. The Lean community has largely focused on pure mathematics. TorchLean expands this frontier toward verified neural network software and scientific computing. With the recent release of CSlib, we see this as another step toward a fully verified ML stack.

We support features:
1. Executable IEEE-754 floating-point semantics (and extensible alternative FP models) verified tensor abstractions with precise shape/indexing semantics
2. Formally verified autograd system for differentiation of NN programs Proof-checked certification / verification algorithms like CROWN (robustness, bounds, etc.)
3. PyTorch-inspired modeling API with eager-style development + export/lowering to a shared IR for execution and verification

Project page: https://t.co/YHpqhRbMQe
Paper: [2602.22631] TorchLean: Formalizing Neural Networks in Lean
Work done @Robertljg, Jennifer Cruden, Xiangru Zhong, @huan_zhang12 and @AnimaAnandkumar.

#MachineLearning #ScientificComputing #Lean

[Larry Diehl @larrytheliquid]

@headinthebox Fun fact, the overloaded functions in julia work similarly to those of castagna

[Larry Diehl @larrytheliquid]

@headinthebox i mean the true castagna-style overloaded functions, not vanilla structural subtyping interacting with set theoretic types

[Larry Diehl @larrytheliquid]

@headinthebox be careful if you plan to support overloaded functions via intersections, as that also implicitly exposes it

[Larry Diehl @larrytheliquid]

@headinthebox word of warning, the reflective subtyping in conditionals at runtime takes quite a bit of work to make efficient, if efficiency is a concern of yours

[Larry Diehl @larrytheliquid]

@headinthebox ah makes sense! our use case was a stored procedure language for a distributed database, which could have JSON-RPC endpoints exposed, so well typed meta-programming with JSON was indeed a sweet spot

[geoff @ 4th of july americana @GeoffreyHuntley]

"I've been carrying this little category theory library around for ten years, porting it from language to language, and every time, the experience tells me something about the state of the art. This time what it told me is that we're living in the future, and it's weirder and more interesting than I expected."

https://t.co/t4dLQ8MOJd

[Larry Diehl @larrytheliquid]

@tritlo @avi_press @openclaw Reminds me of the back in the day primary supported interface for twitter being text messages, hence the character limit