Hmm I see https://peps.python.org/pep-0829/ is now merged in. I understand why it's necessary, but I was doing some cool things with codecs installing "macros" with `.pth` imports.
makes sense (and I somehow missed `mcpyrate` when I scanned the ecosystem for macro packages).
I ended up writing a custom package that expands on the one "macro" (really a small DSL) that does it very quickly, and uses the `.pth` import so it automatically loads, via entrypoint especially, rather than being a constant import in the files that uses the DSL.
To the point of using buckets of models: as long as there's >0 models available in a bucket, and we can order models in order of fit, we're resilient to different sets of models being available. With that said obviously cutting out some models has a much larger effect than others.
OpenRouter isn't strictly necessary but it does make it easier to not have to set up accounts/API keys with several different providers to get started.
Funny enough, I saw some tweet earlier today about their company trying to get past the legal hurdles with getting figma mcp to work and ended up bluntly giving up. Wonder if this is related.
I tend to disregard anyone who complains about foundational open source being funded like this. As long as the donor doesn't have dictatorial control over the product, why obsess over this?
Because not having dictatorial control doesn't protect anyone from anything; politicians take 'donations' (bribes) and then argue or enact laws for what the donor wants, no dictating involved.
Or in software terms, BigCorp sponsors you to develop an open source tool fulltime. If you make the tool more useful for them while making it more complex or worse or ignoring things for others, they might fund you next year as well. If you don't develop things that help them, they might stop and send that money elsewhere and you will need to get a normal job instead and drop your tool work to part-time. Even though they aren't dictating anything, what are you incentivised to do?
There's definitely a cost in terms of features BigCo doesn't need that get deprioritized. But that work still gets done if other people need it, it's just slower, and I don't think you could ever pull off making things worse on purpose in a big community run project like Rust.
The only times this actually happened are C++ and CORBA as far as I remember. Both of them were bureaucratic design by committee efforts from the start, the people doing it weren't users but implementors looking to stop competition, and it all happened before the open source era.
If I was going to be more verbose I’d say in the 10-15 years prior to AI there was such a demand for software engineers that people with poor social skills or with unique idiosyncrasies were tolerated so the odds of meeting odd people in software were higher than in the general public.
It may go further back than that but I don’t have experience before the 2010s
Unfortunately we live in a litigious society. The more you observe, you'll see that companies and government make decisions to reduce culpability and offload risk.
Trust me, only people who were unable to do it in the first place will substitute basic human skills with AI output. They never developed it in the first place.
reply