In the Netherlands, a lot of government systems aren't procured from the Microsofts of this world. There are a lot of middle men (consultancy agencies) involved that over the years have helped build a strong ecosystem with lots of expertise around Microsoft and related suppliers.

So indeed, it's not like you can just replace a software product (or service) by some EU or open alternative. And there are huge vested interests.

Same in Germany. I think in any European country.

I think the heart of the matter is this section in the blog:

> Yes, I blame AI for this.

> I am currently writing much more, and more complicated software than ever, yet I feel I am not growing as an engineer at all. [...] (emphasis added by me)

AI is a force multiplier for accidental complexity in the Brooks sense. (https://en.wikipedia.org/wiki/No_Silver_Bullet)

(I know these days it feels silly to bring this up, but...)

That is not how the separation of power is supposed to work. If a law is bad, politics (preferably a democratic process representing the people) replaces the law with a better one. Until the new law comes into effect, everyone is supposed to abide by the old law, even if it's bad.

Do you think this also applied to stuff like e.g. Fugitive Slave Laws?

Laws are created for lobbyist not people. Nothing works in the real world like in the text book democracy. Everything that was written about democracy is as naive as young adults fiction. But I find it refreshing that the real world events finally force more and more people to realize that.

Not really. I grew up in a socialist (communist?) society. A form of democracy closer to ideal than whatever US is doing for the last 2 centuries or more arrived in my teen years. My personal freedoms were mostly unaffected. Most treasured freedom I acquired was the freedom to gtfo which I'm happily using right now.

I think build for, here means was intended to work at or with human scale in mind. Not be limited to.

Laws usually don't describe the bigger societal context in which they were conceived.

LLMs are often framed as lossy compression, and surely converting some copyrighted Sonic the Hedgehog image from PNG to JPEG is considered copyright infringement, no?!

Relevant: What Colour are your bits? https://ansuz.sooke.bc.ca/entry/23

I suppose both copyright and AI are problematic. And the article is thought provoking.

Might have been related to this:

https://news.ycombinator.com/item?id=46834441

From the Netherlands (eu), Firefox on Linux - youtube.com shows modal about cookies, "reject all"... Videos just work.

I suppose you just ended up on wrong side of google's captcha. Which basically relies on ... Everything.

...Good luck

I skimmed the article and it feels like they're comparing apples (Visio) to oranges (MS teams, zoom). One begin a software product, the other being a service.

There's a fair amount of work to be done before the French infrastructure runs this service as reliably as current commercial offerings.

(Not saying it's undoable, but this isn't the drop in replacement the article seems to imply)

They aren't comparing apples to oranges because they are not making a comparison, they are just informing that French government built their own meeting app to replace Zoom and Teams.

And they are already running it, replacing both Zoom and Teams, and not only them, but La Suite Numerique includes other software as well.

> As long as there are still safety drivers, the data doesn't really tell you if the AI is any good.

I think we're on to something. You imply that good here means the AI can do it's thing without human interference. But that's not how we view, say, LLMs being good at coding.

In the first context we hope for AI to improve safety whereas in the second we merely hope to improve productivity.

In both cases, a human is in the loop which results in second order complexity: the human adjusts behaviour to AI reality, which redefines what "good AI" means in an endless loop.