See also https://vibekeys.dev/ (no affiliation). Has a few more buttons

Can someone comment on the AirPods tweet making the rounds (i.e. the subvocal killer)? I am excited by Subvocal, but am wondering how much bespoke hardware is actually required?

I’ve been using Vercel’s Eve it’s really lovely for an agent that lives on a server somewhere (no affiliation)

what llms are you running with m5 pro?

I am not — I'm using a 64GB M1 Max. Mostly with Muse Glimmer, sometimes Qwen 3.8 27B, Gemma 4 12B QAT, currently playing with LFM 2.5 24B-2B a bit.

I'm just saying, the base M6 is not very competitive in terms of memory bandwidth (the M5 Pro has about 80% faster memory bandwidth).

Compared with my old M1 Max, the M5 Pro is radically better at prefill, only about 15% slower at most at generation (and probably performs better at DFlash and MLX than the M1, so it might close the gap or even outperform the M1 Max)

So the models I am using/testing should all be quite fun, assuming the 48GB RAM machine. The Gemma 4 12B and LFM should run well on the 24GB machine. Neither of those are great coding models but they should be pretty good assistants. There are harsher quantisations of those other models that may fit.

Whether playing with these things locally is worth the extra spend, I don't know. It's probably more cost-effective to buy a cheaper Mac, experiment with the smaller models, and use Openrouter for the bigger ones.

makes sense! was curious to see if there is any deepseek 4.1 flash class model that could concievably run well on the newer machines. muse glimmer, will haave to check that out!

My reaction to AI generated websites and READMEs: I ain’t reading all of that - but happy for you, or sorry it happened

Can't wait for the next generation of e-ink picture frames, Spectra 6 just doesn't cut it yet I'm afraid

Also consider Astropad Workbench. This app doesn't get talked about enough in AI circles but the performance is top notch (no affiliation)

$100/year - why can't it just be a one time purchase?

Hasn’t this always been the case with forks?

I think OP is saying that github, and forges like it, making pull requests so easy, have created a situation where the users bother them too much with said requests. This wasn't always the case with open source, back before community web-based forges were a thing, and it was much more difficult to collaborate from outside the project (where the project has their own regular contributors that do most of the work).

But I don't think most people actually want to just turn off pull requests like they are suggesting... I think that would be disastrous for the open source community as it operates now, and may push people further away to other (maybe even closed) projects.

Perhaps if they still want some form of pull requests, just not as much of them, maybe they could do something like require them become an official project contributor first, where most people work on other issues first before suggesting any of their own.

These questions could benefit from being rephrased to make it clear what is being voted for

An actual use case for Jev?