People want to use LLM hammers like screw drivers and are disapointed.
Also, since the quality of the context matters a lot for completion, copilot will be way more useful on clean code with beautiful variable names and nice structure flows than on code-golf snippets. Same got chatgpt: write good prose, and it will help you better.
I often though we would benefit from having some kind of shell, only a mix between ipython qtconsole et jupyter.
Not an editor like jupyter, rather a shell with a REPL flow. But each prompt is like a jupyter cell, and the whole history is saved in a file.
But if you don't create a file, it should work as well. One of the annoying things about jupyter is that you can use it without file on disk unlike ipython shell.
I would say the most popular way to speed up a python program today is not through a compiler anymore, but using rust for hot loops and maturin (https://pypi.org/project/maturin/) for seamlessly use it in Python and provide packages.
I get the benefit of being able to "just use python" and still gain a speed up though. Plus there are many situations with rust is no possible (no time to learn such a complex language, rust is not approved by the security team).
The other edge is where your heatmap is just all red, and so "speed up hot loops" isn't going to get it done, and at that point somebody needs to bite the bullet and {hire people who know / learn} an AOT language with good performance.
or
it turns out the reason the heatmap is all red is that you were solving categorically the wrong problem, e.g. you decided to use machine vision to figure out what's in the packaging, while your competitors are just scanning the EAN-13 barcode.
They don't have the same goals, nor characteristics.
Numba and cython are rarely used to write the entire program, you usually apply them on your bottleneck, and compile just this part.
Also JIT, like for numba or pypy, compile when the program runs, but projects like cython and nuikta compile ahead of time. The performance characteristics are very different, since the former need a warmup but the latter don't, however they need to be manually built.
Solutions like cython or mypyc require type hinting for getting the best speed up.
All of them except nuikta target better perfs, some in general (pyston, mypyc, cython), some targetting specifically numerical calculations (numba).
nuitka doesn't care about any of that, it takes regular Python code and compiles it. While it does provide some speedup (up to 4x), it's not the goal at all. The goal of nuikta is to able you to ship your python program as a standalone executable to the end user.
It's the most reliable way to do it, as, just like the tag line says, it's "extremly compatible". I've yet to find a Python program that nuikta couldn't compile, including one with PyQT + numpy. It's very good at it.
It's always been there. 20 years ago my father went for business in Turkmenistan, and all this was already in place. This country is one of the biggest producer of fossil fuel, it's an ultra-dictatorship, and has strong commercial partnerships with France. It's very well known.
So why is there sudden medias coverage of this?
I'm now doubting more and more that the politicians and medias really care about any environmental issue and that, like terrorism and pedophilia, it's used to motivate another agenda.
The problem is, unlike terrorism and pedophilia, climate change is an urgent issue that affects every single human in the world, so I really want this to be a genuine concern.
But some part of me now always wonder if somebody, somewhere, is not preparing us to invade Turkmenistan 2 years from now.
Media coverage of methane leaks is going up because:
a) progress on various other sources (e.g. coal) means remaining issues move up the priority list
b) new satellite measurement techniques let us pinpoint blame, rather than do back of the envelope calculations based on educated guesses and global averages, which were bad enough but less actionable.
2023-05-09: The Guardian publishes the follow-up submitted to HN here.
Lots of long-running situations are like this, constantly getting reported on at a level below most people's awareness threshold, and then when you stumble across one such report, it suddenly feels like a flood. See also the Baader-Meinhof phenomenon.
Twitter started like this as well, let's see their real policy in 10 years when investors, law makers, medias and advertisers pressure them before getting too enthusiastic.
Because insurers decide if who is at fault and to pay or not. This system is creating bad incentives since the entity at fault and the entity deciding who is at fault are the same on one side.
The problem with American society, as your comment reflects, is that culturally every discussion always goes back to money.
But money, while quite a good proxy for many things, reduces the human experience to a simplistic model.
It's like game theory: beautiful in a lab, but a parody of life interactions IRL.
You have to shape the system so that the emerging behaviors take in consideration justice, happiness, sustainability, etc.
> Because insurers decide if who is at fault and to pay or not.
In my experience, insurers decide the other party is at fault, unless they are proven wrong. As such, there's no reason for a company to _not_ self insure, if they have the bankroll for it.
People want to use LLM hammers like screw drivers and are disapointed.
Also, since the quality of the context matters a lot for completion, copilot will be way more useful on clean code with beautiful variable names and nice structure flows than on code-golf snippets. Same got chatgpt: write good prose, and it will help you better.