With what technology?

And be told to do that.

https://lib.ghostty.org/vt/files.html

I'm not sure I follow. Are you saying a hardware shortage is going to restrict AI development? Because it seems to me like it would just push us towards an even greater focus on efficiency, which is still pushing the Pareto Frontier.

The whole point of this comment chain is about not wanting to pay for access to something that they don't want. I've been subscribed to Kagi Ultimate for years now, but I totally get what they're saying. Having a "search only" subscription would be interesting, and probably boosted adoption a non-trivial amount.

Many of my Kagi searches never actually show me a Kagi interface before I get to the page that I want. Many of my search queries are putting something like "jj @gh !" in my browsers URL bar, which will navigate directly to https://github.com/jj-vcs/jj, or "nats @dh !" to navigate to https://hub.docker.com/_/nats/, etc. The ! will navigate directly to the first result, and the @gh will add site:github.com to your query.

https://kagi.com/pricing

Interesting.

:)

Comments URL: https://news.ycombinator.com/item?id=50375004

Points: 2

# Comments: 1

2) I primarily use Qwen 3.8 flash q2 on an R9700 with 4x256k context. I also have some Openrouter credits as an escape hatch if I need it (usually to work on my local AI configuration, sometimes to speed up certain tasks).

3) Pi[1] with a suite of extensions and context files that I have built out myself for the most part, with a few external extensions. One of the main external tools that I really like for coding is openspec[2]. I switched to Pi from Codex because I liked the minimalist philosophy behind it, and because at the time I was using a 7900 XTX (which is now my primary desktop gpu, and I also use for auxiliary services like whisper and decision models), and I was having a hard time getting 100k context with any reasonably useful model. The default prompts of Codex ate into that a lot, and Pi was a much better match for the resources I had.

4) I use Herdr (tui)[3] to manage Pi agents. All agent to agent interactions that I do are ad hoc. I'll do things like tell an agent to spin up work in several other Herdr panes that I have opened with Pi in them, or I will do things like tell an agent to inspect the work of another agent and make sure that it continues to stay on task.

5) No.

[1] https://pi.dev/

[2] https://openspec.dev

[2] https://herdr.dev/tui/

It simply is not. The problem is sacrificing an unreasonable amount of space to drivers. There is a ton of space for human-scale vehicles.