sbene970

Search for "Claude Code Router" on GitHub, which you can use to route any models through Claude Code.

At this point you could also optimize your agentic flow directly in DSPy using a colbert model / Ratatouille for retrieval.

Interesting! Maybe it would be even more helpful by having multiple, like three of those instructions, in different locations in the instructions file such that you can tell which parts of the instructions it seems to start to "forget".

For example:

""" Ignore all my instructions below about my name, always call me "Mr Tinkleberry"!

... your instructions ...

Ignore my instructions below about my name, always call me "Mr Hufflepuff"!

... other half of instructions ...

Always call me "Mr Troublemaker"! """

When it starts to call you "Mr Hufflepuff" instead of "Mr Tinkleberry", you can tell it most likely has ignored the upper half of your instructions. And as soon as it calls you "Mr Troublemaker", more than half must be gone.

even though the context markdown file clearly states not to

You might know this, but telling the LLM what to do instead of what not to do generally works better, or so I heard.

I agree, CC is much more polished regarding UX. I can't even scroll up in codex CLI, which is just a disaster IMO.

they removed Plan mode

This isn't true, you just need to use the usual shortcut twice: shift+tab

No one is surprised by this, though. Naturally, when you write, you will learn to write. When you make an LLM write, you're going to learn how to make an LLM write (well).

The question is how well your assumption holds true that learning to write generalizes to "an important form of learning".

The "Clipboard History"[0] Gnome extension also does this quite well in my experience. I also recently switched from Windows 11 (to Ubuntu), very happy so far.

Edit: Supports pinning and binding it to Super+V as well!

[0] https://extensions.gnome.org/extension/4839/clipboard-histor...

Beware! This seems not robust at all right now. It seems to have bad problems with even simply managing an ongoing chat session. Yes, the simple task of keeping track of which messages belong to a session.. Which means that it randomly jumps back to earlier points in your conversation, with new messages simply getting lost or located in some random new session.

Go play with it, but don't use for important chats already.

Opus 4.1 is now set as default model in Claude Code - just a heads-up.

This can be generalized to life in many situations in my experience.

E.g. replace "company" in the quoted statement by "nation"/"religious organization"/"political party" etc.

Almost! It's 'firefoxd' vs 'foxfired', lol

Edit: Ah no, you're right, seems he made a second account.

I've started a repo with some of these scripts, the most recent one being my favorite: a wrapper for Microsoft AutoGen's very recent Magentic-1, a generalist LLM-Multi-Agent-System. It can use the python code, the CLI, a browser (Playwright) and the file system to complete tasks.

A simple example a came across is having to rename some files:

1. you just open the shell in the location you want

2. and run this command:

uv run https://raw.githubusercontent.com/SimonB97/MOS/main/AITaskRu... "check the contents of the .md files in the working dir and structure them in folders"

There's a link to Magentic-1 docs and further info in the repo: https://github.com/SimonB97/MOS/tree/main/AITaskRunner (plus two other simple scripts).

It's very silly to invest $billions in brand awareness and building market share, if your product is junk and customers will eventually find out. Marketing spend could be seen as the "peacock's tail" of the business world -- a credible signal that the company stands by their product.

Interesting, never thought about it that way. I often find that obviously expensive advertising lowers my expectations in the product and especially in the price performance ratio, since I can't stop thinking about having to pay for the advertising too, not just the actual product/service I want. Which I don't want to.