Utensil

[Utensil @_utensil_]

In the process of migrating from GitHub to https://t.co/hQWCKE8IFh :

Featured repos are already migrated via API (agents directly work with them based on the tangled core source).

Still lots of knobs to tweak for CI and websites, but feels home already.

[Lean @leanprover]

Recordings and slides from Software Verification in Lean 2026 are now available.

The one-day workshop and multi-day hackathon, held in Paris in April, brought together researchers from the Beneficial AI Foundation, @CryspenHQ, @ethereumfndn, @GoogleResearch, @leanprover, @MSFTResearch and others to share recent advances in building verified software.

Recordings: https://t.co/KzlJZJwKHf
SVIL website: https://t.co/2NfzbWb8n4

#LeanLang #LeanProver #FormalVerification #SoftwareVerification

[Mitchell Hashimoto @mitchellh]

I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out.

I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really).

It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely.

The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture.

We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying.

I worry.

[Leo Alt @leonardoalt]

We can now fully rewrite most software in @leanprover and prove it correct:

- Compiler module rewrite (AI) from Rust to Lean
- Full FFI integration
- All unit and integration tests pass
- Formal spec and proofs!!
- Under 20h wall time (unnoticed pauses)

https://t.co/u601dZ8wph

[Utensil @_utensil_]

@steipete Anyone ever tried to back up agent transcripts via git would find out about it sooner (don't ask me how did I ... 😂)

[Utensil @_utensil_]

@SIGKITTEN any chance this runs on iOS? then it might be embeded into Kitty Litter to connect to tailscale 😃

[Utensil @_utensil_]

@mitchellh @ashekhirin @zack_overflow Even cooler if the uniforms also work for normal mode in Helix (currently only insert mode and command line prompt)

[Utensil @_utensil_]

@dynomight7 It would be nice to add a section to compare with einops, which solves the numpy hassles https://t.co/ifhYAQOYUq

[Utensil @_utensil_]

slides: https://t.co/wKSSnkR75r

[Utensil @_utensil_]

Rendering complex implicit surfaces using interval arithmetic etc. The ideas behind mkeeter/fidget well explained. https://t.co/bWhCThTfBk via @YouTube

[Alberto Maria Pepe @albertomariape]

sharing Torch-GA: a rendition in Torch of TFGA, a library to build neural layers in Clifford/Geometric Algebra.

Torch-GA allows to work with Algebras of any dimensionality + weights/biases with fully customisable grade. Do try it out!

https://t.co/eCMkrDkL9s

[Utensil @_utensil_]

An introductory post to the talk: https://t.co/qM9jtOQjJb

[Utensil @_utensil_]

"Short theorems with long proofs" by Prof. Colva Roney-Dougal: A proof has an area, not just length - cases run sideways avoiding direct assault. Proof's curvature must also be carefully tracked - you can't sweep it under the carpet, it always pops up somewhere.

[Utensil @_utensil_]

@tom_doerr @paulgauthier It would be nice to have a way to put suggested commands with comments explaining their purpose in an editor so I can tweak before execution instead of just confirming or copying them out of aider. Ideally they are suggested in batch, but executed one by one.

[Utensil @_utensil_]

Really amazing talk by Tomasz Stachowiak: Rendering Tiny Glades With Entirely Too Much Ray Marching https://t.co/N8Nka6i1Qo

[TooMuchVoltage Software @TooMuchVoltage]

Presenting: SDFs on a CWBVH!

With much thanks to https://t.co/Sxs9KI4GdM and @j_bikker's https://t.co/foCiv4KQgk .

Technical details and comparisons available here: https://t.co/gvuBaPH7E1

#graphics #programming #rtxoff #raytracing #vulkan #vulkanapi #glsl #shader

[Minhyuk Sung @MinhyukSung]

#SIGGRAPHAsia2024
🔥 Replace your Marching Cubes code with our Occupancy-Based Dual Contouring to reveal the "real" shape from either a signed distance function or an occupancy function. No neural networks involved.

Web: https://t.co/4NijU2pPJR

[Christian Legnitto @LegNeato]

I ported @zach_nussbaum 's matrix multiplication blog post to Rust GPU:

https://t.co/YNyB8fgIB8

Please upvote on HN! https://t.co/W8rlm3YVDu

[Wenzel Jakob {deprecation notice} @wenzeljakob]

Following over 1.5 years of hard work (w/@njroussel& Rami Tabbara), we just released a brand-new version of Dr.Jit (v1.0), my lab's differentiable rendering compiler along with an updated Mitsuba (v3.6). The list of changes is insanely long—here is what we're most excited about🧵

[Utensil @_utensil_]

@tensor_fusion Nope, but maybe it's just due to the prompting. After Deep Think, it gives a reasonable informal proof for a more trivial theorem on concrete cases, and the formal proof in Lean 3.