You seem to contradict yourself. You say platforms cannot prevent AI (slop), but then say it is helpful when they are required to label it automatically. How are they supposed to label something they cannot detect?

In the backend ads are using AI (and more general data mining) since at least a decade. What do you think why they grab every data about you that they can get? Because they use it to predict what ads you're likely to click before deciding what ad to serve to you. That are systems that are specially trained to predict your next move, not generalist chatbots.

I think there are many people much more advanced than the first year graduate who cannot do your task. You may be a bit biased, remember: https://xkcd.com/2501/
You are comparing someone who worked in a very niche of the field (or used time to learn that niche) to a generalist. If you need your high-throughput kafka pipeline and want to compare to a kafka expert, you would need to compare to a specialist model fine-tuned on kafka and related tech.

Aren't bubbles usually imploding?

Ask average joe to write a script that shows the time elapsed between two given dates. One time joe is allowed to use AI, one time not.
You will see a 100% increase in productivity for that task.

Nano has no line select. Thats the most basic feature I use every day. But there are a lot more. What about deleting up to the next quote?
dt" (delete to ") in vim. Most other editors can't do that with so little input. What about deleting content inside a HTML tag? That's dit (delete in tag). What about jumping between opening and closing braces?

Nano is (old) Notepad. vim is not comparable with any other editors, because it is so much more powerful. Can have similar power with emacs, but you cannot compare the two from their philosophy how to use them.

emacs has a vi mode.

It says "E37: No write since last change (add ! to override)".
I pressed "!" and it appears in the lower right of the last line.
I pressed it again and now there is ":.!"
When I press enter I get "E34: No previous command"

Thanks for nothing!

They cannot quit development.

Who doesn't know anyone with WhatsApp?

I am pretty sure they agreed on that mutually.
A few days before ChatGPT was refusing to do anything that specifies a known artstyle and had prompts like "If you are asked to draw a popular person, first generate a description and then generate using the description". Then boom Ghibli style popularized by their CEO himself. I am pretty sure that was a marketing stunt for both.

I heard it works pretty well in the downloadable versions (I don't have the hardware to have much fun with video). The guardrails are in the frontends, not in the models for Chinese stuff.

The EU does not require cookie banners. They require privacy. Cookie banners are the attempt of websites to tell users "Please forfeit your privacy voluntary because we are not allowed to just take it". The problem is not the banner, but the users accepting them. We would not have them, if users would have said right from the start "Fuck off, I value the privacy the law gives me".

The point to disable infinite scrolling is more nuanced, though. The right implementation is correct pagination, that allows you to start reading where you left off. The whole reverse-chronological timelines are a feature to bind your time.

With current systems, you start with recent posts and scroll down in the hope to find where you left off. You have no idea how many posts are in between, and you probably read new to old while scrolling back.

With the systems demanded, you have a permalink to your reading position, and continue reading forward (more useful and natural), have a page number how many pages you have to catch up (if you want to read all) and no FOMO about gaps you cannot bridge because you can only scroll back a certain number of posts.

You have little time the next two days and a whole day off after? Just catch up with the news on the third day, no need to try to catch up every evening to not miss out on something.

I am not using these thing yet, but from what I heard, the whole selling point is that they are not single-taskers, but organize their tasks more or less autonomous, finding sub-tasks, things they need do before and so on.

The downside of such things? Your bot may be writing rants about software developers behind your back and you learn about it when it already is debated in social media. And I think your own security is also often neglected, many tools seem to use way less sandboxes than they would need to be secure.