They can doesn't mean they should. Letting an unsupervised agent register domains and build websites exposed to the public is yet another recipe for a disaster waiting to happen.
In the short term, yes.
Over the long run, I think it's good that we move away from the "seeying is believing" model, since that was already abused by bad actors/propaganda
Hopefully, not too much chaos until we find another solution.
Because the whole point of the startup economy is not "find profitable companies" but "find unicorn companies that will pay back 1000 times what you invested". With this goal in mind, you can afford to lose your investment on 999 shit companies and find 1 successful startup and still break even.
Hence, you're looking for quantity over quality. Ideally, you'd have both, but the talent pool is only so deep.
As an Italian, I can tell you this is just a move to grab some votes from their supporters. Lab meat is nowhere close to mass production so there is no real impact of this ban right now. When (and if) lab-grown meat becomes commonplace and, hence, cheaper than regular meat then the ban will be reverted as it would have negative implications on the country economy.
For an example in CS, we currently use machine-learning algorithms that work very well in practice. However, why they work so well is still unclear in many cases (although we are making progress).
I disagree with your definition of "well". It found an optimal solution only 30% (roughly inferred from the bar plot) of the time and only on graphs of very limited size (up to 14 nodes). I'm pretty sure even the most basic path-finding algorithms we have can do better than this.