kannanvijayan

I am in no way trying to undermine your point, but I have to say as I read past that first sentence, it became impossible to not imagine Tobias Fünke's aspirational role as "Frightened Inmate Number Two" as the subject of your commentary.

I came of age with the personal PC revolution and developed my career during the internet revolution, and was well into my career when the mobile computing smartphone revolution occurred. I did not see or feel even a fraction of the amount of social dread with those technologies that I see with AI, and I think that's justified given the circumstances.

Saying that this revolution is another one of those is true in a general sense, but misleading in the specific sense. Just as saying that adding 1c heat to some body of water is "more of the same" can be true in a general sense, but misleading when the water's basis temperature is 99.5c.

That is to say, periodic or gradual change induces large changes at inflection points.

It feels similar to the progression from post-WW2 technological utopianism that arose with the adoption and deployment of massive advancements in materials and mechanical, and electrical technology, to the cold-war dread of nuclear and ballistic technology.

I think the 90s saw the end of the cold war which lifted a bunch of that dread. The optimism of those days has been worn down over time.

What I remember back then is at least the pretense being sold that the technology would improve people's lives by giving them more leisure and enjoyment. That the productivity gains would be reflected in some improvement in quality of life. I think the main difference now is that this expectation has been lost by most people.

This has been very informative. With object property/method lookup optimized the next big leap in JIT-style perf is actually a hard pull. The next big perf win lies behind local type inference, and a method-jit with your standard bevy of SSA optimizations to hoist the typeguards.

I suppose the various JIT projects are trying to come at that from different angles incrementally, but that's a genuinely hard problem as you don't see real benefits until multiple complex pieces of infrastructure are in place and with good coverage over the language.

Anyway, my hot take was misplaced.

I see you brought up the freedom of speech group, but didn't address the potential implications of other similarly important groups such as fursuit afficianados and PETA and, most egregiously, the Taylor Swift fandom.

Ah, the `__slots__` feature makes it make a lot more sense.

> Sorry, that sounds a bit vague and not too different from what CPython already is doing.

With slots being available, I can see where I was misled. I went looking for a type structure that modeled the dynamic properties on an object which would be used as a typeguard, but the class itself functions as that if you're restricting the optimization to slotted properties.

Appreciate the clarfication!

Gonna dump a technical hot take here.

Python is a mature production language with a lot of users. Even before it was huge - like 20 years ago when I was in college - getting a small backwards compatible API change into a python extension module (gnu readline in my case) required convincing GvR personally that the API change was necessary.

Every JIT attempt I've seen for python just tries to cut too deep in the first pass, and scope themselves too large.

I believe any successful attempt at optimizing python performance from JIT will FIRST start with a 100% semantically compatible hidden type model for optimizing object property access and method dispatch.

The single biggest performance delta in most dynamic language jits come from that, not from optimizing guards off of bytecode operations or from compiling to machine code. It's from converting hashtable lookups into constant-time guard + offset lookups.

The hidden type model is also a hard nut to crack. It needs to perserve dict semantics which have very dynamic behaviour via the various `__*__` methods, and handle exotics scope objects since scoped name lookups are the most common ones which have their own exotic behaviours (I remember python providing some ability for callee frames to inspect caller scopes in some limited ways).

The "good enough" problem is a real biter. The personal PC market overtook the mainframe market long before PCs were even close to beating the top supercomputers. Cheap and good enough is the threat.

What does good enough look like for various use cases, when do we hit it, and can these companies guard that territory?

I moved my projects over to GitLab about a year ago.

GitLab's model is different - at least it was when I moved over to it. They charged for public repositories and gave you a free private ones as a sampler for trying out the platform. That approach is far less prone to getting overrun by millions of slopositories.

GitLab never tried to be the social media of software development. Theirs is a much simpler, classical model - one of the reasons I switched to them: you pay them money, they provide you services for that money.

Oh I thought you were going to give an example of the prior "regime" telling CFTC staff to retract analysis of suspicious oil trades.

I guess you are asserting that this other thing is equivalent to that?

You seem very sensitive if you think someone asking you if you are trolling and asking you if you are being serious is "an insult". If this is what gets under your skin, I shudder at the thought of what the rest of the internet feels like to you. It must be a living nightmare.

Also, where did you get the impression that trolling required the underlying opinion to be falsely held? Whoever taught you that was doing you a disservice.

Who exactly claimed to support anyone saying anything? Or are you claiming that supporting "freedom of speech" implies that one must support anyone saying anything?

The US also considers prosecution of the Epstein child rapists as unnecessary.

Point being, I don't really think that "the US considers" carries a whole lot of weight these days.

The idea of foreign entities holding American leaders to account when Americans are unable to do it themselves is a fundamental emotional taboo for many.