I think the main problem for Linux is the fragmentation and lack of focus. If you can live without the Adobe suite and such, any number of distros and desktops can serve you well, but it often tries to do so

An initiative like Omarchy got a lot of traction just by "picking one" of all the infinite options available, writing decent documentation for how it all works in Omarchy specifically, and having the whole thing install in minutes.

Omarchy and tiling VM's are not for everyone but I think the principles are great, and can surely be applied to other DE's as well.

I think Windows has improved as a basic OS tremendously in the past decade, with things like WSL, winget, terminal, storage spaces, fancyzones and such. The components of perhaps the best general purpose OS available right now are there, but they are buried behind a crappy car-salesman tabloid-press upsell circus that is frankly unbearable to some users.

What I don't get is why MS refuses to let you pay to avoid the circus - It's like a streaming service that only has the ad-financed tier!

I get that the people who care are a minority, but charging that minority a high purchase price for a just-the-os version of windows seems like easy money and would let them dodge all the badwill by presenting a choice.

Right now the official option is simply "suffer the circus or leave". It should at least be amended with "or pay".

My feeling is they would never do that because it would implicitly admit that stuff is bad and people don’t want it so much they’re willing to pay a lot to remove it.

>Apple is #4 in laptop sales. Lenovo, Dell and HP each have at least as much volume.

True but they divide their sales among several models. Gaming models, 2-in-1s, 13" to 17" and so on. Apple not only has fewer models they often keep the same case design between generations which also benefits economies of scale.

Volvo still design, develop, prototype and even mass produce cars in Sweden just like they always did. They haven't had Swedish majority ownership this century but still do their thing.

That could change of course, but so far so good for the Swedish economy and fans of Scandi design.

Since Chinese tech is clearly state of the art for EV's I think Volvo could be in a good spot if they get to continue as they have so far. Win-win.

As for realiability, 500 000 km is no problem for a decent EV and Geely makes good ones. I wouldn't worry about that aspect either way.

I think this old quote can come around to being accurate in a way, if you consider that from the user's perspective every cloud service is like one system. Aws, Azure, Google cloud...how many will there be when the dust settles? ;-)

When someone uses a violin to hammer nails, we don't laugh at the violin.

...although to be fair, LLM's are like violins that are really good at pretending to be hammers. :)

I think Uncle Bobs advice is mostly bad and am afraid to admit it because it’s like a (cargo) cult now.

Agreed. The quote from him in the article reads like a caricature of a comically overzealous SWE.

Add most of what Martin Fowler said to that list.

reading uncle bob's book was the singlemost harmful thing i ever did to myself as a software engineer. wasted so much time writing such worthless tests. TDD? ughhh no way.

Why no power? The forests, hydro dams, heat plants and such that gave power long before wind became a thing are still there.

>The option up here really truly is "do we use fossil fuels, or do we use nuclear". Renewables do not help.

Hey now - renewables gave us electricity up here long before Einstein started thinking about atoms!

We are very few people here, 250MWh helps a lot, but if we have to chip in to build a nuclear plant we'll be broke before the project planning is done. ;-)

I think the fragmentation is good, it allows many talents working in the same area at once and lets users choose what fits best. The problem is that this leaves a lot of the polishing work to the users, and the defaults - which most stick with - are often the most boring, safe choice.

I think what DHH did with Omakub (and Omarchy) was a constructive solution here - use the myriad of options to pick out a set of components and configs that work well. Polish the selections, hide and ignore the capabilities that don't fint in, and document how to use the resulting "bundle" in great detail.