I think it’s more of a “how you use it” thing, but you’re definitely right that AI agents can’t design systems properly.

Some people I know have produced way more code, removed tech debt, and all without introducing any bugs since they started using AI. That’s because they’re not using it to do anything beyond their skillset, understand everything it’s doing, and are using it to catch mistakes they otherwise would have made. Other people are using it without reviewing the output, or are using it to try and do things beyond their skillset, and that’s how you end up with infinite tech debt and a whole host of bugs.

Personally I’ve recently started heavily using the AI code review bot we have at our company, both for my own code and other people’s. While 50% of what it says is hallucination or wrong, that’s not an issue because I know it’s wrong or a hallucination so can just tell it no and to focus on other things, like catching bugs or issues that most reviewers would just glance past, and also gives you a rubber duck that talks back.

Same in English…

Can’t, won’t, you’dn’t’ve, and if we get into dialects then you get things like owt (anything), nowt (nothing), ont (on the) and int (in the).

I don’t even support blocking propaganda and fascism. Maybe you could block malware sites, but even that’s a push. Let moderators of individual communities figure it out, otherwise you’re just going to turn whole instances into echo chambers.

I mean it’s also kind of true though…

AI is taking some jobs, in situations where the limiting factor is the rate at which work can be done rather than the skills required to do it. Say you have five people in PR, of which three are responsible for trawling through sources to find out what’s actually being said about the company, and two are responsible for writing press releases. The jobs of two of the people trawling through sources could be replaced with AI, as the limiting factor is the amount of posts, documents and stories you can read. While you’d still need an overseer to fact check and collate, that sort of work can be done much faster than actually reading and finding sources. If the company also lays off one of the people responsible for writing press releases, however, that would be unrelated to AI as that sort of job isn’t replaceable by AI (right now at least) due to the majority of the work being something that probabilistic models just aren’t correct enough to do, so that’d be an unrelated layoff being blamed on AI, even if whoever orders it genuinely believes that AI has replaced the job.

I mean one of those faces consequences for its actions if it kills a protester, the other less so

I think you’re underselling just how evil he was…

They found lots of bones from women and girls when excavating the torture chamber beneath his former mansion, and Stalin distrusted him so much that he dropped everything when he learnt his daughter was alone with Beria.

Most people in this thread were either evil in their career, or in their personal lives, but Beria managed to excel separately in both.

Worse is relative, a proportion of the requirement increase will be due to worse code, but much more will be for features to make the software more accessible to more people, and adding features without needing to remove old ones, neither of which are a bad thing, otherwise everything would be a command line tool that removes options every few months and only has one way to use it

Sure, but corporations which keep profits up by providing a good service that people are willing to pay for should be celebrated and rewarded over those which keep them up by wringing every last penny out of you while simultaneously degrading their service.

AI image generation is amazing for replacing stock photos, and not bad at replacing clipart and porn images.

AI video generation is ok at replacing very simple videos without continuity or physics, but their only real applications are for spreading misinformation or mindless scrolling, there’s just no real way to get anyone to pay for them.

That’s aside from the fact that sora could’ve been great for generating generic stock footage/b-roll, but the way they implemented it was to generate a script, then audio, then video, which meant that it really struggled to generate anything without a focal point, ie what it would actually be useful for.

It’s fundamentally hard for AI to make things dirty.

It can do messy, but it’s based on denoising, and it’s very hard to distinguish a layer of dust or grime from noise, so unless some new technology comes up we’re still going to have that as an indicator well after scale and limbs are fixed.

Twitch is more of a parasocial website in most cases though, unless you’re watching exceptionally small streamers

Regarding rigging voting machines, that’s conspiracy theory territory, however you did miss that Democrats are more likely to be working jobs where you can’t easily take time off to vote, or it’s that or lunch, or the time they give you isn’t enough, so doing it in advance by post is pretty much the only way you can vote.

There hasn’t been an insignificant number of times I’ve found out about a new product or event I’m interested in, a sale for something I already want, or something like that through advertising.

It’s rare as a proportion, but it definitely does happen.

I believe that ads are just yet another tragedy of the commons type of thing, where bad actors not only ruin it for everyone, but also convert good actors to being bad actors.

I’d say there’s three tiers:

• Ads showing you things you actually want or need, and providing you with new information.
  • These are going to have high CPMs, so you don’t need many per page, and having more per page will decrease their value, but kind of require tracking to ensure their relevance.
• Ads showing you things you might not want or need, but might cobsider buying, or information that isn’t immediately relevant.
  • This is the baseline for reasonable quality, untargeted ads, and CPMs for these are going to be fairly low, but much higher if you click on them
• Ads promoting scams, malware, and things you neither want nor need.
  • In this case, the CPMs will be virtually zero, so the site is forced to cram as many on the page as they can. They’re also encouraged to get you to click by mistake.
  • This makes people block ads or trackers, reducing the number of ads in the first category and forcing more sites to adopt these patterns.

It’s kind of sad that it’s going this way (and has been for a while) but I guess it’s going to end up with just a return to paying for media with money rather than ads.

The surge starts around 2016, could be political instability following David Cameron leaving/Brexit, smartphones being common enough that muggings increase, changes in policies due to Sadiq Khan being elected, children who grew up during the recession growing old enough to join gangs, or a number of other things

I think the drop is 2020 and then it never recovered actually?

Ah yeah, I was looking at the mail/password manager/docs side instead… They’ve also got Proton Drive on there as well though

It’s interesting they haven’t recommended anything proton here when proton as a whole works very well in a bunch of these categories… I imagine it could be because they’re a competitor to Tuta

Probably Switzerland.

All AI is good for is giving instructions on how to make bombs, and generating images of tits, but they caught on so now we just end up with search summaries saying it’s not physically possible to [xyz].