There was a story of a flight that turned around because a passenger told the crew he'd left a power bank in his checked luggage. There was no suggestion of a problem - just that there was a power bank in the hold.

These sorts of things are "an abundance of caution". On the face of it, they're daft, but you'd be worried if they didn't do it. The alternative is delegating the decision making and domain knowledge of flight security, and thus the responsibility of the entire flight to a couple of hosties and the captain. Instead, organisations like the CAA, FAA etc make the rules and the hosties/captain just have to follow them.

Hosties look after the cabin, pilots *fly* and some people at the regulator think deeply about security. I'd personally rather all of those people stick to their best skills, because I sure as shit don't want some idiot from the CAA spilling drinks all over the place, the captain signing off on security and the hosties flying the plane.

> I don't think it can be profitable.

Oh it can be - but it'll be a shadow of its current form to do it.

IMHO, this is uninvestable (like SpaceX, for similar fundamentals, although different causes). That is, you can speculate on the share price going up and down - that's fine. But before you buy and hold, or buy in because you like them or you use Claude or some other emotional reason, think about how they're going to have to change pretty fundamentally to go anywhere - including losing their morals and internal safety nets. They'll have to cancel a lot of internal programmes, cut costs dramatically, crank up revenues, and as I say, likely have to do deals with people they'd (up to now) chosen to avoid.

So good for them - the senior managers and VCs are about to make some serious millions. Once the two year golden handcuffs are off though, it'll all change.

> Funny, though, that there has never been a need to extend DNS to support autonomous applications deployed to the internet, yet now with AI we need it?

There are people suggesting we need a whole new Internet for agents. Putting some SRV records in DNS seems like a pretty small ask in comparison.

(I can't imagine why the Internet isn't "good enough" for agents. I suspect it's mostly not good enough for VC money, but there we go)

With Spec Driven Development, the documentation is the "source of truth" and the code is just an expression of it. As a developer, you have to review the code, but if it needs modification, you modify the specs - not the code.

So in future, developers could be spending their time editing markdown descriptions of things instead of actual code. They're still going to need to know how to code though, because they need to review what is produced, but they won't need to spend the time to get the syntax just right - AI will have done all of that.

If (and it's definitely still an "if") SDD proves out to be a better way to write software (at least in certain domains), then it will indeed do away with the verbal story telling. If it's not part of the spec, it isn't part of the code. Whilst spec writers might be as slapdash as Jira ticket writers, there will have to be at least some mention of every "side conversation" that ever took place, if it's part of the end product.

There's no magic though - it's still garbage in, garbage out. Keeping the quality of those specs up is the subject of most of the research in this area right now. Likewise of course the model makers are trying to produce better quality code for ever more vague prompts. It remains to be seen if any of this is truly successful, but already today it is a way to get some software written - and crucially also get some high quality, up to date software written for it too. I'm sure there are people studying this sort of thing, but it's believed that the output of SDD is faster (maybe cheaper?) than humans hand-writing it all.

I honestly don't know my ar*se from my elbow here, but I'd imagine that the drones won't really do all that much. They'll tootle about listening for submarines. If they detect some, they'll alert the Admiralty who'll decide what to do.

For the most part, submarines are supposed to be 'invisible' and their locations aren't supposed to be known. Thus, if you detect one, you're one step ahead of your enemy. I'd imagine that if your enemy was tinkering with a cable, and one of your own subs just happens to come by and makes it known that they're there, and that they know your enemy is there (and who the enemy is, probably down to the exact craft and maybe even the crew), and therefore they know what the enemy is doing, that would be a fairly significant indicator that your enemy ought to leave, quickly, and attempt "lose their tail" while they do so.

My point is, if you're able to find your enemy, then your enemy loses their advantage. My guess is that in 50 years time, the entire seabed will be littered with fixed or moving drones of various types keeping a constant watch on anything that happens to float by. But until then, they're planning to just keep an eye on cable runs (or whatever else is of particular interest I guess - they're not going to tell us what those might be).

Just as they flag up copyright material, they'll flag up humans talking to each other as "AI". There'll be no way to talk to a human about the mistaken flag, there'll be no human verification of your appeal and your account will be closed without human oversight.

Honestly, self-serving or none, it's going to get harder to actually put any sort of content onto YT. On aggregate, that's probably a good thing because there's a lot of dross, but it's going to piss off the "influencers" no end.

I ditched 'em years back, probably when they got acquired, so just took a look to see what they were about. They've accepted that Spotify et al. are really where it's at, and now all they seem to do is track what you listen to on those platforms and then offer suggestions.

I suppose if you're listening to multiple platforms, then having all that history consolidated in one place could be useful, but otherwise, I wonder how their recommendations are going to be any better than Spotify's?

Back in the day, last.fm was pretty awesome though. I found a load of new music through listening to them, which back then resulted in me actually buying physical media. I suppose now, even if they found me something new, the best they/the industry could hope for would be that I bookmarked it or something for a couple of plays later on.

> A "tool" that lets one programmer do the work of 20 means that 19 will be laid off,

For me personally, I'm looking forward to a future where I get to do *engineering* a lot more than I ever have done. For a few decades I've been trying to get programmers to engineer solutions to problems instead of just patch them. I've even pushed devops folks to engineer solutions too, so it's not just programmers. The point is, there's a difference between a "software developer" and a "software engineer", likewise a "devops analyst" and a "devops engineer". Assuming the name has been used correctly, the latter in both cases possess skills that no AI has yet assimilated. Never say never, but I can't see the current crop of AIs becoming capable of engineering either - sure, they might start putting logging and metrics in their code by default and claiming it's because they "get" engineering, but that's really not what engineering actually is. I actually think most employers will need a lot more engineers than they currently have - sure, that might mean some programmers/devops/QA get fired, but the actual engineers will likely be collecting a decent salary and likely will be turning out work pretty constantly too.

I'd also disagree with the 1:20 ratio, I'd say it was 1:2 in some cases, but 1:5 is probably a more reasonable guess. Tech CEOs blaming AI for job losses isn't a reliable source of information, nor are the tech-bros claiming their new model can do the work of 20 people or whatever.

> Anything that tastes or feels good IS BAD FOR TOU.

Probably, but we keep hearing "occasionally is fine" - all things in moderation.

As far as this TFS/TFA go, I'd like to see the government embarking on some education rather than moving straight to banning things. Alcohol is already on something of a decline in British society, so a bit of education might be all that's needed. Likewise ultra-processed foods and so on, there have been some news articles and documentaries on the subject, but something more sustained would probably do some good. Education won't solve every problem of course, but it might as well help those that it'll help. We've all had our self-destructive phases in our lives, and I don't imagine that will change, so rather than turning those into some sort of underground, illegal affair, just let people do it - they pay plenty of taxes to do it, so it's not all bad for society.

As for sex... get married. That takes care of that problem in more ways than one ;-)

> Just what we need: the ability for anyone to publish a vibe-coded Android app.

No - AndroidStudio only lets you built the apps. To publish, you need to get through the labyrinth that is the Play Store. There you need to jump through several hoops to even get an account, and then you need to jump through several more hoops to publish an app to the test store. Getting to production... yes, more hoops. Oh, and then you *must* update something every 6 months or they close your account, leaving you with nothing to show for the many, many hoops you had to jump through.

Honestly, it's a wonder anyone ever actually manages to *publish* any apps at all.

I'd say 'diabolical' could fit here, for these reasons:

1) Github is owned by Microsoft
2) VSCode, and it's extensions marketplace are owned by Microsoft
3) The data lost was far in excess of what any one developer with a malicious extension should have had access to

They, more than anyone else, could have/should have vetted the marketplace before letting their people install random crap from it. They could have potentially made VSCode more resistant to malicious extensions. They could have publicly vetted the marketplace so malware never got into it in the first place (so we all get the benefit), or even just vetted it down so that only their whitelisted extensions were available for their devs or whatever.

They also could have limited individual developers access to small numbers of repos, rather than giving them access to thousands that they probably will never so much as look at, much less ever actually work on.

They chose to do none of these things, and so someone used something they shouldn't have done. Probably a bit stupid, but in fairness, it's hard for random individuals of varying levels of skill to be able to properly assess the safety of an extension to their IDE. Especially as their environment is likely deceptive because they don't have any access to production networks, they mistakenly believe they don't have access to anything 'important'.

I'm astroturfing this link at this point, but all this on this backdrop: https://blogs.microsoft.com/bl... - so yeah, 'diabolical' fits.

I wonder who, if anyone will really "invest". I can well imagine a lot of people will be looking to ride the share price changes, but who's going to *invest* for multiple years at a time?

I ask because I'd imagine the first thing the collective shareholders will ask for is that the crappy bits of Musk Inc. get divested as quickly as possible. They're all going to be loss making - whatever notional value (say) Twitter has, it'll actually make way less than that when it gets sold off. None of that looks like a sensible investment at all. Again, by all means ride the share price changes around the announcements and actions, but invest... I don't get it.

Once all of that is out of the way... yes, you'll be able to invest in SpaceX, and it'll probably be worthwhile too. Just maybe wait 5+ years.

Google have forgotten what good search is - I suspect the airy fairy 'designers' have taken hold, and are warping the managements minds with tales of how they're gonna optimise the user experience by leveraging some of the core assets, without boiling the ocean. or some such.

Aside from the search results being different from the thing you asked for, they have of course been adding in AI. I actually don't have a problem with that, in some cases it's good, albeit at the expense of the website that gave them the information in the first place. I also note that 'search' is no longer on the home screen of Google Play - that decision seems probably the most crazy to me. Even Cloud Console seems to be getting harder to find things in, although it does still have a useful search. Don't get me started on the so-called Search Console though - that thing is just a crapfest. I'm glad I don't work in SEO.

If I were running Google, I'd make it an internal law that all pages, everywhere in all circumstances need to have a search on them - that's the core reason for their existence, and so they ought to be promoting "don't remember, just search". They ought to be returning search results as honestly and quickly as possible - by all means come in with AI or the bullshit results afterwards, but give me what I actually came for right away.

As for the search market, I can't see Bing getting anywhere, and DDG needs a new name. There's no one else worth thinking about, and boring old search isn't gonna get any investment money these days, and you can't vibe-code a new search engine in a weekend, so sadly, they might be as good as we can have for some years yet.

> The CDC does say some patients can continue to spread Ebola in semen after they recover, although I'm not expert enough to know if there have been confirmed cases transmitted in this manner.

I can confidently say, there's no point asking that on Slashdot ;-)

If you're paying a million a year (and making 9 million) you definitely should be thinking how you can get off that platform. You could hire a bunch of AI-wielding programmers and devops to make a site just for you for less than half a million a year - and actually run it professionally enough too - no vibe coded crap. I'm sure there are other providers who would love to host your well visited content and so whatever other services for less than substack charges.

Usually sites offer discounts on volume, because there's no way it's costing substack a million, or even half a million to run the site for that one content creator. Without the discount, they're just opening the door to competition - which apparently is happening on quite a scale.