The fundamental problem is that the EU's regulatory bodies do not understand technology. Regulation in the EU has turned into a bona-fide industry, and it's a massive train of the blind leading the blind.

Whether it's with Article 45 trying to break PKI [1], or with the AI Act trying to regulate the transformer model itself [2], or with ChatControl trying to surveil everyone, EU regulation is clearly a clownshow at this point where you get cheered on as long as you propose some regulation, any regulation!

[1]:https://www.eff.org/deeplinks/2023/11/article-45-will-roll-b...

[2]: https://twitter.com/arthurmensch/status/1725076260827566562?...

There have been positive changes at EU commission lately regarding this Article 45 thing:

https://securityriskahead.eu/wp-content/uploads/2024/02/Mozi...

eIDAS is also a bad thing for privacy in my opinion because it's going to make it much easier for websites you request identification online. I believe identification should remain a barrier, the same way you don't ID yourself at every physical store you enter. eIDAS doesn't mandate this but it does prepare the technological foundations to do that online.

And it doesn't have much to do with PKI breaking, though it's nice that tangential link was removed, it's indeed good that that attempt was called out.

Removing the most objectionable content doesn't make it a good thing though.

And again this is a big business lobby championed by Thierry Breton. This is not something the voters asked for, it's something the industry wants.

> This was already attempt 2

Agree with almost everything else, but if you're counting attempt number one as the one a couple years ago allowing services to do the scanning if they wanted it's a bit disingenuous to put it in the same category as mandating that they do