orbis-tertius

Seconding that, gorgeous animations

It was nice while it lasted

That's very nice

Actually reading through a 35 pages planning doc. One word at a time.

And ctrl+r status.claude.com like I'm Jesse Eisenberg at the end of The Social Network

can someone ask codex to fix claude code please

Maybe Cloudflare related? There were some (possibly unrelated) outages on CF around the same minutes

Forgot to make it public! Should be working now

Thanks a lot for flagging this

If anyone is interested we built a TUI to use this outside Claude Code, with integrated spaced repetition and anki flashcards export

https://github.com/nodo-labs-uk/socratui

This works really well.

A positive take: this works much better than the many projects on gh that did this before Jane Street. This would be evidence that creativity and execution still matter, even when the product is just an md file

A cynical take: network effects are everything. tens of projects doing the same thing and this only gains traction when an already popular org does it

With Captchas? But I do understand your sentiment

So Fable 5.2 max technically failed the pelican test?

Yes, the comment was intended as a playfully hyperbolic nod to the lack of novelty and originality in the idea, but I guess the intended tone didn't carry through :)

Wouldn't learning a calibrated probability distribution just be a consequence of the CE loss being a proper scoring rule? So, by construction, they are approximating the distribution of the labels. Of course in practice there might be issues with overfitting the log-likelihood and the with the labels being a finite sample of the population distribution

Thanks! For now we just keep separate Claude Code sessions as separate independent windows. What feature would you like this to have? For example, the ability to organise windows/sessions together and have them push to the same attention queue?

If you're interested in these topics, the Sutton & Barto RL book is excellent, and Chapter 2 has a lot of discussion on exploration vs. exploitation strategies

Would it be a crazy unimplementable unenforceable idea to demand legal requirements for agents to identify themselves to servers?

Weight-space programming doesn't have an established definition yet afaik, but it would implement weight changes using a trained model using forward passes, no derivatives needed

Nice solution

I did read the article but missed that. Paranoia can definitely co-occur with distracted reading, they're independent dimensions ;)

Thanks for the info, will give it a try

Fair enough, I missed that. Will give it a try!

I love Atuin, but have some reservations about a shared server keeping all my shell history and agent sessions. Am I being too paranoid and missing out on convenience?

> Me: That's awesome! Please check which of these villages haven't been discovered yet. Search both the literature and the archives.

> Agent: One major village doesn't seem to have been registered.

The idea of just brute forcing discovery with agents on lidar is really cool, but doesn't the story ultimately rest on the agent making a mistake in its literature search?

We run everything in separate worktrees, and built https://github.com/nodo-labs-uk/attention-queue to manage our Claude Code sessions

Sorry for the snark, was meant in good spirit :) The unoriginal subtext being a reference to the correlation between LLMs capability and number of Rust repos.

Thanks for the link!

*just let your LLM write rust

(sorry)

Agree, I wonder if this is just a case of trying to meet white collar workers where they are? I guess most non-technical white collars are familiar with sheets, docs etc and prefer sticking with it

Very beautiful poems, thanks for sharing

The website doesn't help understanding what the product is, so hard to comment on it. Something OS/LLMs related? Anyway, it inspires an aside: now that website building is commoditised, the status/expertise signal seems to be anything that moves opposite to Claudified design rules. Hence the videos, the leftfield animations, the lack of brand carousel etc

Very nice, and congrats on launch! It would be helpful if you could add to the README a summary of the key differences with other TUI orchestrators.

I understand the feeling, and definitely share it sometimes. Feeling like we're losing expertise and ownership of our work sucks!

Maybe programming is just not an intellectually satisfying vocation anymore; but whichever format your ideas will take in the future, please don't stop having them and working on them! That would be such a waste