“Oh, finally, someone with 20%. Wait… is that a negative sign?!”

Unfortunately there’s still plenty of unencrypted traffic from normal usage (e.g. DoH is still fairly rare).

If every AI company steals the public data separately, it means massively increased costs for everyone who is getting their data stolen. If the AI companies “steal” from each other it’s much better for everyone else.

Hm, alright, I can see that - but to me, this is an example of business practices that the GDPR is explicitly trying to restrict. Of course it will be difficult to delete someone’s data if you’ve been sharing it with many other companies.

Could you expand on some of these challenges? We haven’t had these issues in any companies I’ve worked at, but those were mostly on the smaller side.

As a passionate Golang hater, I can gladly explain!

• It’s not just the repeated if err != nil, even though that’s already bad enough. But the really fucked up part is the := bullshit. It makes moving code around unnecessarily annoying, and it’s telling that few other languages share Golang’s approach.
• The lowercase/uppercase rules for private/public stuff is theoretically not a horrible idea, but it makes the code look much more inconsistent. I find _ much easier to read, and this leaves upper/lower to signal other details. But I see that this is mostly personal preference.
• Fairly basic operations take much, much more code than they should (e.g. deserializing JSON while handling extra args, or basic functional operations - though that should change sooner rather than later with the new generic methods, right?)
• The decision to initialize every non-pointer primitive with the “bottom value” (or whatever it’s called again) makes sense in isolation, but it’s really unfortunate that they don’t support additive types, because this means a bunch of common tasks need to use pointers, and unfortunately the type system is worthless when it comes to preventing segfaults caused by bugs with pointers.
• I find that most Go libraries have basically no documentation, if you’re lucky you get an example that might vaguely be related to whatever you want to know. I’ve had much better experiences in other languages.

All in all IMO most Go code is 5x longer than necessary to actually express itself in a readable manner, all because the language still doesn’t have proper error handling or generic support (until recently at least). At the same time it’s fairly inflexible, the type system is still shallow and basic, and it’s still way too easy to shoot yourself in the foot.

The only good thing Go has going is the single file deployments, but I’ll gladly spend one hour of every remaining day of my life setting up containers, if it means I never have to touch anything Go again.

A lion does not concern himself with the opinions of sheep.

So please don’t be mean, sheep, or the lion will start crying.

Weird question to ask through radio, over.

AI could be useful if it could orchestrate more complex tasks that the relevant apps don’t natively support. But of course that’s not what we’re getting.

I use lots and lots of timers every day to structure my chores and work. For example, I’d sometimes love to be able to start one timer, and have another one set to automatically start as soon as I stop the ringing from the first (which might take a few minutes).

Very simple requirements, but specialized enough that I’d have to write it myself. But I can’t see phone AIs supporting something like this in this decade.

That’s good! But I’d still prefer if I and the other 44% could at least say “please stop producing and sending us trash”.

They should really make mailed advertisements opt-in, but of course the industry won’t do that. And this what makes it cancerous IMO: they never give up trying to inject ads into people’s lives.

OTOH, it seems like ~44% of junk mail gets thrown away without ever being opened: https://zerojunkmail.org/environmental-impact

That’s a bunch of environmental destruction and personal annoyances caused, so the label “cancer” seems fair to me.

Yeah, I find it hard to believe this wasn’t their goal, especially considering they found this bug in testing and still released the update: https://news.ycombinator.com/item?id=47994193

I’m also not a fan of Slack, but it doesn’t break new things every week while also putting ads in your face, so that’s a plus.

Teams is also only communication.

Slack isn’t great, but at least a few years ago it was much, much better than Teams

Yes, it is. No, the delays on Youtube don’t come from the performance of the adblocking code, so you won’t notice many differences. But more efficient adblocking is good for everyone - noticeably more so on devices with batteries, but still helpful for everyone.

Well yes, but have you also considered that line must go up?

Firstly, the GNOME Foundation’s financial woes long predate the rise in AI scrapers.

Sure. Does the Gitlab -> GitHub redirect also long predate the rise in AI scrapers?

I know of no such increase.

If you haven’t been staying up to date, maybe do that before dismissing explanations? The increase in scraping is such a massive issue that it’s been discussed to death again and again, and big parts of the open internet has changed.

You’re basically calling bullshit on a charity having increased fuel prices, because you’re not aware of any Iran conflict!

Ah, and why isn’t it reasonable to assume that the increase in hosting costs which is hitting everyone for the last few months, is also hitting GNOME? Why do you think they are exempt, when even big commercial platforms are struggling with the amount of traffic generated by AI scrapers?

Given that hosting costs have objectively increased across the board due to AI scraping, and the GNOME project just introduced these changes, it is obviously sensible to assume that GNOMEs hosting costs have increased too.

Not righteous at all, bro :(