There is Biscuit from OSDI 2018 https://pdos.csail.mit.edu/projects/biscuit.html

The 24% more PRs was merged, not «landed». Presumably after human code review (but I didn’t check the paper…) Not sure about this particular study, but I «think» my own productivity (coding wise) has improved by more than 30%. In many ways I wouldn’t even have started on many of the things I’ve completed these last six months because I would have viewed their effort to be insurmountable within my time budget as a professor (with all my other duties).

Does anyone know how this compares to Supacode (https://supacode.sh)? I just started using Supacode a few days ago and have been enjoying it so far; it is based on libghostty.

Supacode is macOS only and is a terminal emulator itself. Herdr is cross-platform and is just a multiplexer. You can use it in any terminal emulator. If you want you can even run Herdr in Supacode

I’m in academia and I try to convert everyone I know… I only recently started using it for a few things, and I used Claude to do some pretty advanced stuff that I wouldn’t have ventured to even try with latex. That is, I think a lot of the relevant CTAN packages can be developed using LLMs more quickly than you might expect.

For next year, I plan to prepare a thesis template for our university and encourage students to try it instead of latex (most of our students use latex now).

I’ve been using Typst lately and it has been great. I’ve made an exam template for my university and made an export feature so that I could generate the exam in the json format that our online exam system (WISEflow) expects, with support for multiple choice and essay style questions.

It is so snappy and with great error messages. I encourage people to try it out. The typst tutorial is very approachable. Thanks to the Typst devs for this great piece of software.

I’ve been using Typst lately and it has been great. I’ve made an exam template for my university and made an export feature so that I could generate the exam in the json format that our online exam system (WISEflow) expects, with support for multiple choice and essay style questions.

It is so snappy and with great error messages. I encourage people to try it out. The typst tutorial is very approachable.

I should note, it's still not on par feature-wise compared to TeX ecosystem, but it gets there with incredible speed. As for UX - it beats anything TeX-based ten times over.

Curious if people here have looked at the upcoming SIMD support in Go: https://go.dev/doc/go1.26#simd

Currently experimental, but looks like the first Intel arch will arrive in the next release in about 3 months. They are also going to support a portable layer.

Wondering what people here think about the approach the Go team is taking; I think they would appreciate more eyeballs on their design. (I’m not competent in this space (yet))…

I did a lot of experimentation with the Go1.26 experiment. It's easy to use and produces good code but only supports x86 ATM. (See https://andrewwphillips.github.io/blog/go1p26.html#simd-expe...)

I think there will be a "portable" wrapper that will also support other architectures (arm, riscv, even wasm) in the future based on the Highway C++ library.

Looks like that isn't a portable SIMD abstraction, but more similar to adding architecture-specific SIMD intrinsics support to go, with nicer syntax.

Sorry, I didn’t explicitly link to the issue for the portal layer.

Here is the issue discussing the portal simd package: https://github.com/golang/go/issues/78902

Also here is an example of some Go and the Asm code generated:

https://godbolt.org/z/n8hKhc7rY

(click the recompile button if you don't see the Asm code)

For the record, Go’s telemetry is local by default (not uploaded): https://go.dev/doc/telemetry

Call me when they have broken ECC with a real quantum computer.

That would be about 10-15 years after the moment it would have been wise to migrate to PQC. You won't have the time to migrate before breach when you start after ECC is broken.

Why is your use case interesting?

There is a $2T dollar use-case.

What’s my remedy when Google’s product (Gemini 3.1 pro high) makes a “grave” mistake? This is unrelated to the bans that’s been happening recently, but wanted to share …

This morning I asked Gemini to “save” its output to a local file. However it did more than that … it committed the file (along with several unrelated staged changes that was not ready to be committed) and even pushed the changes to GitHub. I’ve never asked any model to commit, let alone push… I’m not impressed; actually a bit disappointed that it would do this without any warning up front. This happened in Antigravity.

Tool confirmation and controls, every LLM with any instructions is susceptible to this "grave mistake", it's in the "nature"

Or just don't give it the keys, this is my strategy. Put them in a box with specific tools and access.

See also The Lethal Trifecta