chris_st

FWIW, Cognition has all the Sonnet/Opus/Fable models, and all the GPT ones, as well as GLM, Kimi, and Gemini.

What "non-house" harnesses have you found to work best?

Interesting to see (roughly!) what percent of each car is engine.

I have this setup, roughly, with UTM rather than orbstack. I think I have it set up safely, curious how you see it has the wrong permissions?

Another way I'm "going slower" is to have the AI implement individual sub-steps of the current task, and review each one. It's slower than having it yolo out the whole thing, but it's much smaller incremental bits to review, so my brain doesn't glaze over in a huge review, like I had if I had it do the whole task.

I'm following an Ideas -> PRD -> Issues -> Tasks methodology, where each task has a bunch of sub-tasks. I have it just do one (or a few, I'm having it do Red/Green/Refactor as separate sub-steps, so I review the Red case, and then once that's good, do the Green and Refactor steps, and review those).

Yup, just like people!

Good!

Nope, not better quality if you don't like the smell of cigarettes.

Please add support for the Windsurf editor as well. Thanks!

Awesome, stealing that!

And they're clearly marked as `unsafe`, so easy to find, which gives them a nice list of issues to address.

I asked Claude to tell me why something was implemented the way it was, and got an excellent response. One data point, would love to hear more examples.

...it came as a surprise that [leaving a Petri dish out with a window open] would end up with interesting [molds] (called [penicillin]). _It was not planned at all_.

Well, to be fair, people cheat by remembering what they did last time. I think the idea here is to run the models from a "clean slate" and see how often they succeed/fail.

They are, like people, non-deterministic, so giving them several "fair" trials makes sense to me.

Possibly because they just haven't been able to manufacture enough of them yet to be a viable business to others? They're fighting everyone else for foundry space and time.

But it's pretty cool that LLM bug hunting is pretty cheap... the 1-person projects can do it themselves, don't have to contract out to some huge security company.

Well, maybe not... see Simon Willison's ongoing reporting [0] on all the bug reports for `curl` people are finding with LLMs.

Interesting to see them go from "DON'T GIVE US AI SLOP!" to "Wow, lots of actual bugs found, including [ed: at least one] bug found by two people!"

[0] https://simonwillison.net/search/?q=curl

From the article:

As we noted at its September beta release, a windowed version of Tailscale’s macOS app doesn’t replace the menu bar app, but runs alongside it. It can be pulled up from the Dock or a Spotlight search, and makes a lot of Tailscale data and features more accessible.

Seconding Maple Mono - it's very nice.

Just out of curiosity, which version of Claude?

Good nerd stories, alas it was cancelled, so no new ones:

- Uncharted with Hannah Fry

Some great fiction:

- Achewillow - horror, but not excessively horrible.

- Desert Skies - humor, about folks who work in the first sphere of the afterlife, folks who are recently dead and arrive in Buick Skylarks are equipped with microwaveable burritos and information about the spheres to come.

Fantastic poetry:

- Poetry Unbound

Really fantastic interviews, alas, it's no longer updated:

- Partners by Hriskikesh Hirway

Linguistics and language:

- The Allusionist

Tabletop RPG:

- My First Dungeon

Turns out the apps in /Application/ run as you. Problem (between keyboard and chair :-) solved.

Honestly, I don't know! I should write an app and see who it runs as. I did an `ls -l /Applications`, and while every file is owned by `root`, none has the `suid` bit set.

LM Studio doesn't have an installer. Those often have to run as admin, and who knows what they're doing then, so that probably wrongly set my concerns about putting stuff in /Applications/.

I'll dig around the interwebs and see if this is answered elsewhere.

Thanks!

Nope - on macOS, almost all apps are just "drag this to wherever (usually your own personal application folder)" and they work perfectly, since they don't need admin privileges. But this one insists on running from /Applications - the root application directory - and for reason. To install there, you have to be admin. I really don't want apps installed as admin, and possibly then able to get admin privileges. It's just basic security.

There's a thread on their Discord that was reported in February of last year. No fix, no comments.

My complaint is that LM Studio insists on installing as admin on my Mac. For no apparent reason, and they refuse to say why.

To set the bar for other websites, to show how it should be done?

Or maybe just "for the love of all that's good"?

I've found asking GPT-5.2 High to review Opus 4.5's code to be really productive. They find different things.

Thanks very much! Glad to hear it was so good for you.

Curious about that book - did it help you? Silly questions, but serious: How much did it help? What, specifically, did you get from it? Asking for a relative who has serious anxiety issues (no, really, it's not me... I have depression, and Burns' "Feeling Good" helped me substantially, specifically by identifying brain-spirals I get into, which lead to depression, and once known can be avoided). I'd love to recommend something good for them. Thanks!

Recommended this to a friend who hates keeping track of "how many" kinds of things, and was dismayed at having to say how many books they've read (which they really don't want to do) and found they couldn't skip that question.

Maybe make it optional?