And after you then stopped typing the response, otherwise having to use a device that was made in China built from oil-based components using oil—based transport throughout its supply chain, send a postcard with your apology to the Kagi team and anyone else who does not fund wars and still uses traded goods, because this is how the world is.

If you want to throw a however tiny wrench into Putin‘s war efforts, complain load to your government to join sanctions and to have them support Ukraine and make your choice at the ballot box accordingly.

If a site works on Safari but not on Orion it is mostly due to ad blockers etc. Just flick the compatibility mode and it works. I have not encountered a single case where this did not fix it.

Also for a while now Apple Pay works, Apple password manager works, autofill works.

[1] https://github.com/kagisearch/kagimcp

On my personal test bench, when compared to other inexpensive models, GLM 5.1 provides the answers that I would consider most complete or satisfying (these are subjects that I consider myself an expert in). The answers tend to be more comprehensive, nuanced, and include references that I would consider the correct ones (if given access to web search).

I also find it a joy to code with, somewhere between Sonnet 4.6 and Opus 4.6 (have not tested Opus 4.7 yet).

Finally, just gauging by pelicans, it kind of stick out: https://simonwillison.net/tags/pelican-riding-a-bicycle/

I like the idea of using small local model (or several) for tackling this problem, like low rank adaptation, but with current tech, I still have to piece this together or the small local models will forget old memories.

I have experimented with a lot of hacks, like hierarchies of indexed md files, semantic DBs, embeddings, dynamic context retrieval, but none of this is really a comprehensive solution to get something that feels as intelligent as what these systems are able to do within their context windows.

I am als a touch skeptical that adjusting weights to learn context will do the trick without a transformer-like innovation in reinforcement learning.

Anyway, I‘ll keep tinkering…