> That common knowledge, nothing revolutionary here.

I've never read about that. So it's not "common knowledge" - except maybe in the running community.

I like your comment for putting some facts into place (how far you can go with common options). But as I never heard of this before, I have no idea how common it actually is and the effects and the science around it, what research does say to this, how and why this is used in other sports - or why not.

So it was at least concrete / tarmac instead of mud?

Concrete. Used the opportunity to do some doughnuts before continuing on our journey.

I've fiddled around with https://docs.getmoto.org/en/latest/docs/server_mode.html

It didn't support the one thing I wanted but it was so easy to find the right place in the code, I was happy. Never got to continue it though or turn it into a PR

I can try!

I currently use ChatGPT for random insights and discussions about a variety of topics. The memory is basically a grown context about me and my preferences and interests and ChatGPT uses it to tailor responses to my knowledge, so I could relate better.

This is for me far more natural and easier than either craft a default prompt preset or create each conversation individually, that would be way too much overhead to discuss random shower thoughts between real life stuff.

This is my use case and I discovered that this can be detrimental to specific questions and prompts and I see that it can be more beneficial to have careful written prompts each time. But my use case is really ad hoc usage without the time. At least for ChatGPT.

When coding, this fails fast. There regular context resets seem to be a more viable strategy.

I see what you mean, but I like having a clean slate even for those one off questions. I don’t want a differing answer to a philosophical inquiry just because the LLM remembers a prior position I’ve written about you know?

I have all the history settings off for this reason, but something that worries me is that there's a fair bit of information about me trained right into the model weights. I'm not "famous" by any stretch but claude has awareness of some of my HN-front-page-hitting projects, etc., which I think should be enough to bias responses (although I haven't tried to measure it).

I set my name to "User" in the settings, so in a clean-slate chat it has nothing to go on, but the moment claude code does something like `git log` it knows who I am again. I've even considered writing some kind of redaction proxy.

FWIW, both OpenAI and Anthropic have a toggle to do a “Temporary/Incognito Chat” that does not use or update memory. I too wish this was the default, and then you could opt in at the end of the chat to save some long term aspects into memory.

That would be interesting, also at the start. As an option what to pull in. ChatGPT memory "improved" and now you normally don't even see anymore what it commits to memory!

At some point free markets become fiction. There's no financially viable way to start competing businesses in markets as entrenched as mobile OSes. Otherwise this would have happened. And if that becomes anti consumers, then the consumers start changing the rules the companies operate under. Because in a democracy we have more consumers than CEOs,so they vote with majority.

(This obviously simplifies things, but ultimately we as humans still haven't found the one and only true philosophy or moral, and maybe that's not possible (I'm no philosopher))

But this thread is about the option to install apps on your device regardless of OS vendor approval, and that's not possible either with iOS nor is iOS open source. And that's what this is all about. If you don't care about open-source and user freedom, then this change wouldn't matter to you anyway.

Until it does ;)

But I still agree - if the benchmark was in memory, Stoolap might be optimized for speed. Sqlite is optimized for persistence, so you have to benchmark on disk and compare how it performs when writes fail.

Not a security expert and also curious about implications:

I always considered it the best solution to have both: VPN encryption and TLS encryption over the VPN. Different OSI Layers. Different Attack Surfaces.

Not sure if that is a recommended pratice though (see initial remark ;) )

When looking at the git history, that suggests yes, that xoscript is a fork of citrine

That in itself is an answer :D