> colors

I'm always amazed when people go through the trouble of adding a light/dark theme switcher rather than just respecting the user's system preference with `prefers-color-scheme`[0].

[0]: https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/A...

You can do both, because some people like having per-site color controls.

Like I prefer to read prose and literary website content in light mode while keeping my OS in dark mode.

Admittedly it's a niche use case.

You can do both. And then you help out people who don't know that they have a meaningful system preference (or maybe all the other sites they visit don't honour it, or maybe their web browser isn't picking it up for some reason).

My problem with pay-by-the-token is that it discourages me using the thing ("oh the prompt will cost me $0.1"), so I pay a subscription which I'm pretty sure costs me about two-three times what I'd pay just for the api costs, but encourages me to use it more ("oh I have a subscription already, better make use of it").

> in 3D Clifford algebras it repeatedly confuses exponential of bivectors and of pseudoscalars.

I have no idea what any of those words even mean. I'm sure LLMs make similar obvious-to-professors mistakes in all the domains. Not long ago, we didn't even have chatbots capable of basic conversation...

Ironically, it's sort of the other way around! Every frontier chatbot since GPT 4 (at least) has had a pretty good understanding of even very esoteric technical concepts.

Bivectors and pseudoscalars (in a 3D context) are "just" signed areas and volumes. Easy!

Back around the GPT 3, 3.5, and 4.0 era I used to ask the bots to explain "counterfactual determinism", which is one of the most complex topics I personally understand.

Then I would lie to the bot about it, and see if it corrected me or not.

This test is useless now, the frontier models can't be fooled any longer on such "basic" concepts.

Conversely, LLMs are basically useless at anything that doesn't have enough (or no) public information for their training. Think: obscure proprietary product config files and the like, even if the concepts involved are trivial.

Similarly, Clifford Algebra is a relatively niche (even "alternative") area of mathematics and physics, with vastly less written material about it than the competing linear algebra. Hence, the AIs are bad at it.

Very low quality article. Makes a lot of unsubstantiated (though probably true) claims. Spends unnecessarily many words shitting on AI generated images and the people who prompt them.

Czechia? I always have to explain that Allegro is not a bad company per se, just absolutely messed up the attempt at conquering Czechia...

> Typical salary for a senior dev is around 20-30k PLN per month

Typical, yes, though it's possible to earn way more than that. Not that 20-30k PLN per month is bad, the average Polish salary is perhaps around 9k and the median around 7k.

20-30k PLN goes a long way in Poland. Some seven years ago, I was spending around 7k PLN a month, living in the beautiful Warsaw old town, 50 metres from Kolumna Zygmunta, eating out all the time, and generally felt I was living like a king. Good times!

It does go a long way, but note that 2019's 7k PLN is today's 11k (or 15k if you use the Big Mac index).

> Czechia is richer per capita.

Are you sure about that? I'm Czech but have lived in Poland for 8 years and visit regularly. Poland used to be way poorer than Czechia, but these days it looks the other way around. I think the stats are either lagging behind or computed wrong. Note I regularly visit both the cities and the countryside in both Czechia and Poland.

Btw, the article has a "GDP per capita growth in post-communist countries" table, with Poland at the top and Czechia at the bottom.

Yes, indeed, indentation is one of the very few things in Python which aren't problematic!

They coulda made it Scala!

I like that we, the humanity, have started paying attention to virus outbreaks. Compared to "real" pandemics, COVID-19 was rather mild, but it helped raise awareness. I think we're now much better prepared and equipped for the eventual real pandemic.

Are we? I feel like we're cruising and flying a boat full of potentially infected people to 23 counties.

I don't really think so. Already in the first stages of this outbreak we're not doing any quarantine, instead we're infecting airline passengers and personnel and let them spread the virus uncontrolled. That doesn't indicate a proper prepared response.

Official knowledge is that Hanta transmission required prolonged close contact, but there are increasingly indication that Hanta can be transmitted through the air. That is going to be ignored in favor of the official but possibly outdated mode of transmission, leading to wrong or insufficient response.

Also I feel like people will be more hesitant than in 2020 to adopt behavior that avoids virus transmission.

If mutated Hanta variants turn out to be very effective at transmission, and if we don't have the luck of a quick vaccin as we did with Covid, we're cooked.

Hanta is a lot more deadly than Covid, and that can possibly be a good thing because that's the one thing that could lead to proper effective response. It has the potential to lead to rigorous measures to stop transmission instead of allowing it to spread to the whole population, leading to fewer cases and fewer deaths.

Millions of people died and the main takeaway for the US seems to be giving up vaccines and cutting programs that would mitigate future disasters. Not sure how you've come to that conclusion.

There’d’ve been be no cases if we just hadn’t done testing.