Okay friends, new side quest! I've started a Python core development newsletter!
If you've ever wanted a regular summary straight to your inbox about all the cool things happening in CPython (and adjacent areas), this is it! Like and subscribe!
First edition, room to grow! 🌳
The technology is very much a case of "I'll believe it when I see it" for me, but the Lake Nyos disaster released hundreds of thousands of tons, whereas these plants hold about 2000 tons. So unless something's getting misreported, the risk of serious harm seems far smaller.
Multiple interpreters look like a nice abstraction. I'm looking forward to giving them a try.
Personally I like multi threading. Free threading is a minor improvement for me - I have *one* use case for cpu-bound threads. What I'd like is better abstractions. No-one likes explicit locks.
The Board can do with experience, or with new blood. So many great candidates! Make your own choice!
Except DO NOT vote for Franz Király. Even ignoring his bad faith arguing and bad ideas, he's clearly not a team player and does not act in the Python community's best interests.
If you do decide to vote, please read the candidate details.
If you do read the details, do notice how one stands out.
If you do notice one standing out, do check out how that person is acting on discuss.python.org right now.
If you do all that, then please cast your vote.
Looks like wheel next is about to hit Discourse. There goes my free time for the next few months :-( #python #packaging
Principle, mainly. LLMs are based on (potentially illegal) massive use of content that wasn't paid for. So charging for the results feels wrong, IMO. But also the hype misrepresents the situation by not being explicit that good results require payment and not addressing the ethics of that.
AI companies lost almost all chance of trust when they trained on data without ensuring that the author(s) of that data were OK with them doing so. That's not just "illegally use copyrighted material" but also "assume consent based on licenses written before AI training even existed as an idea".
Regrettably, the (lack of) credibility of AI companies makes it hard to believe such claims...
Wow, I didn't know that. So that's another reason I want transparency over costs. "Don't train on my interactions" is a non-negotiable baseline for anything non-trivial (*certainly* for code assistance) in my opinion.
And *my* comment was in the context of having only just discovered that a lot of the pro-LLM posts I'd been reading were referring to experiences with $200/month plans. I was pointing out that the "power tool" analogy feels like a bait-and-switch in a world of generally free dev tools.
To be clear, I have no problem with people posting about what paid LLMs can do. I just want it to be clear that's what they are talking about. I don't read posts on PyCharm (£200/year) - why should I be interested in posts on LLMs that cost ten times that???
I'll freely admit that my experience is solely with the free ChatGPT (and specifically the "stay logged out" version). I use exclusively free development tools, and I don't see why LLMs should be any different. And I don't want to deal with my interaction history becoming part of the context.
What frustrates me most of all is that people praising LLMs aren't transparent that this is *not* the experience you get from free ChatGPT or Gemini, or whatever. Maybe everyone is talking past each other, but IMO it's pretty basic that a tool not being free is worth being explicit about.
Given the situation with nearly every other developer tool, anything non-free is expensive (especially for non-professional or open source developers). $200/month is ludicrous, and even $20/month is more than I pay for any software that I use - for *anything*.
It also feels like "here's a free power tool that cuts your arm off most of the time, but that's ok because there's a hugely expensive one that mostly works once you learn its quirks, and keep a surgeon nearby just in case"
Is it just me, or does this feel like a weird selling point?
Ah yes, that one. I agree - a ridiculous argument. I had a go at rebutting it, but they don't seem to be really responding to concerns/feedback :-(
Sounds like you should (and add a link to the post on the PEP thread). I haven't read the PEP yet, but I'm not even sure why "changing a class at runtime" is relevant (except in the context of "... which makes immutability hard, so this is what we've done to deal with it").
Cool. When free models are effective I'll take a look. Until then, it would be nice if pro-AI posts (in general, not specifically you) included some sort of "how much I pay to get results like this" information...
100USD/month??!!! When the majority of development tools are free, that's a ridiculous amount of money. I don't believe I've ever paid that much for a service. If the people supportive of LLMs are paying this much, I'm not surprised there is such a huge divide between promoters and skeptics:-(
Hey #PyConUS attendees--You’re invited to participate in the Python Packaging Ecosystem Survey! How do you interact with Python packaging? What do you wish were different? Take this survey and let us know! Thanks to @anacondainc.bsky.social for sponsoring and coordinating❤️🐍