Why does it require write access to change my Github profile and act on my behalf before I've decided if I want it to make any changes?

Hello, thanks for the feedback. I need to only read your profile but GH permissions only have read+write options for this data. I am researching workarounds. That said, I have no workflow or code that changes the profile in any way.

Interesting, that seems very short sighted on GitHub's behalf - I wonder what their reasoning is?

That is an insane amount of money for just 32GB of RAM! That's what we were paying back when it was hard to use more than 32-64GB in a desktop setting. These days with all the electron and node bloatware, containers everywhere and AI - 32GB doesn't get you far.

Handy with parakeet v2 is excellent

Managed to get 6.4tk/s with some modifications: https://github.com/PureBee/purebee/pull/1

That's the postMessage bottleneck - PR #1 replaces it with Atomics-based dispatch which should push utilisation much higher. Early numbers look like 6.4 tok/sec on M2 Max

That's my PR

Interesting idea, FYI M2 Max gets around 4TK/s with only around 14% CPU utilisation

Patents

Really? Feels like there's a bunch of companies

It is complex. The early patents have expired, so a bunch of companies wouldn't surprise me. I can't follow the chains, but it appears that there are not really a bunch of companies, it is one company (maybe two) that makes just the displays and sells to others.

I don't really get what is different about this from how almost everyone else uses Claude Code? This is an incredibly common, if not the most common way of using it (and many other tools).

Looks like the blog is down, but I'm going to assume it's someone complaining about the number of AI related projects in Show HN, personally I often look at Show HN for interesting AI related projects.

Handy is nothing short of fantastic, really brilliant when combined with Parakeet v2!

There is almost no point in telling an agent to build a skill without augmenting it's knowledge on the thing it's writing about as you're just piping output to input without expanding the information in the system. If you get an agent to perform a bunch of research online, distil that down to information that the models tend not to get right or is newer than what is in their training data or simply better aligns with your desired workflow than what they generate out of the box - that's going to create a far more useful skill. I use a skill that gets activated when creating a skill to help guide this approach: https://github.com/sammcj/agentic-coding/blob/main/Skills/sk...

Absolutely, they didn't give the agents autonomy to research or any additional data. No documentation, no web search, no reference materials.

What's the point of building skills like this?

I find it useful for it to automatically update skills after trying them out in the wild. It can then improve the skills with real feedback. Seems to work well but I didn't do real research on it.