>AI was supposed to make development much more efficient and fun, because now you could choose what to work on and what to outsource to AI agents to implement. Similar to how you manage a team of devs.
When was ever "managing a team of devs" fun?
Especially for people who liked coding, management was always dreadful. If you went for it, you went for career reasons, not because it was "fun".
I enjoyed the idea of giving boring work to someone else and working on stuff I want myself. But that only worked if I code at night, after I waste my whole day in meetings
It's bad at some many levels - but it's also a case of if you know, you know. There no universal "KPIs" one can check to say it's good or bad.
But it's full on the "AI slop" style, a mix of LinkedIn/Online Marketing/Listicle like writing. That it's part of hundreds of thousands of articles, all with the same mechanical style, makes it even more tedious than if it was just some individual blog post here and there.
"Set the scene" paragraphs to add suspense, repeated lists of three or more items, "and, honestly"/"does real load-bearing work" and phrases like this, mic drops after every few paragraphs, writing about mundane stuff like a marketing executive, the list goes on.
This one isn't even one of the really eggregious examples.
>And the only claim for the existence of such products is that one day Anthropic or OpenAI will jack the prices up until the model is unaffordable or the model will just become unusable.
The other claim is that open-source models exist and that they're cheap and hence we should be using them.
So the first claim is not "the only" - there's also the second claim. And both sound like great points.
>The problem with all these claims is that they assume that we use frontier models because we trust the companies or because the price is right. That's not why. We use them because we want the best model available.
What "we want" doesn't matter one shit if those are priced out of our reach.
Anyway, the whole post is AI written, we shouldn't even have engaged.
There's a singular very good reason to upgrade to 27 and any new version (at least when a .1 minor version is out and the kinds have been ironed): the whole ecosystem works better when you are on the latest stuff.
So unless you have legacy app needs, or need ultra stability for stuff like running a video shop or something, it makes sense to keep up, or at least don't fall more than 1-2 versions behind.
>They removed Rosetta and you have to reinstall that if you want it.
And they'll kill it entirely in a future version. It's been like 6 years for getting apps to Apple Silicon versions. And it's generally not ideal to run both AS and Intel programs if you can avoid it (wastes double the system lib memories, Intel misses some extra protections, etc).
>I switched to MacOS 3 years ago
OK, this explains it. So macOS is not like Windows where you can have even 15 years old version of the OS and programs still work forever.
There's no singular "right way to think" that applies in all cases. In many cases thinking of "radical alternatives" can be a bad idea (think people falling for "alternative medicine" stuff). Or some idiot coming up with shitty "radical alternatives" because they can't accept some specific setup or status quo is better - also related to the N.I.H. mindset in software.
This thread is dead, but my point was about radical alternatives to academic fads. These are stickier than most fads -- they typically last a generation to a lifespan, because the only thing that kills them is the death of their believers. And they are always wrong, and absurdly wrong, and destructively wrong. Any alternatives to them are either light paraphrases of them, or else radical; and the radical ones are mostly crank. But the only way to make actual progress is to find the rare radical hypothesis that is not crank, and the only practicable way to do that is to generate more radical hypotheses. (The impracticable approach is to wait for a genius. People have starved to death doing that.)
Thanks for expanding. That second paragraph is the "meat and potatoes" I was wishing for, while only cautioning that style doesn't necessarily falsify.
Yes I agree. My mind goes back to the Bob Cringely plane documentary recently posted upon the occasion of his sad passing,[0] as a prime example of how someone can elevate radical thinking too far. The Contrarianism Is The Product, essentially, which can be great as a tech writer, but it's lousy for an airplane PM.
The rest of the right way to think is the obligation to discard radical hypotheses that don't work out. Visionaries see things that other people don't. Mostly that is because they are not there. Every so seldom, they are.
Combined with N<20 and a definition of trauma that means literally any experience, we might be able to get a preprint published claiming trauma cures cancer as well.
> Yeah this caught my eye too. I can't imagine any technical reason why a Mac works but a windows/Linux computer wouldn't work.
Parent meant the macOS version restriction.
Why it's not for Windows/Linux is obvious. Unless it's to give Windows users access to some cross platform service or hardware (like the iPod used to be), it's not Apple's job to make software for other platforms. They're not a company making general business software like Adobe or MS.
This is to interface with macOS only Apple Wallet anyway.
Ah thanks, I didn't realise Macos 27 is a new version.
Ironically though, I think it's kinda the same thing? Apple software is first class if you follow their rules - including having Apple hardware, being on the latest version, and so on. It's still a nice and pretty walled garden of Eden.
Apple has several web apps, including tv.apple.com, podcasts.apple.com, maps.apple.com, icloud.com, music.apple.com, apps.apple.com, account.apple.com, card.apple.com, privacy.apple.com, reportaproblem.apple.com.... not to mention business oriented ones too.
It's 2026 and people still think it's not about convenience.
If they have to "maintain their own data under their own care" and it's inconvenient, they'll prefer the convenient option, even if it means ocassionally losing their stuff or lacking privacy.
So you don't get any disappointment or regret from the article at all? It reads to you like someone merely lamenting that it happened to be them while fully aware that they chose the risk and are fully responsible for this outcome?
Don't be silly, tech companies market themselves to people as reliable. They do their best to give people the impression that they don't need to weigh reliability against convenience and that they can have the best of both worlds.
Articles like this one exist because people need to be reminded that this isn't, in fact, the case. What they fail to do is present any proper alternatives to really give people the chance to make that decision.
And hey, if someone knows these things happen, knows how easy it is to set up a family nextcloud plus some backup system, and still choose to take their chances with the corpo slopware, then at that point it's really just their own fault.
>So you don't get any disappointment or regret from the article at all? It reads to you like someone merely lamenting that it happened to be them while fully aware that they chose the risk and are fully responsible for this outcome?
You don't have to be "fully aware" of the risk to chose convenience.
In fact, the biggest convenience is not even bothering to research and think about risks.
I didn't say the chose the most convenient option after researching risks and features. I said they chose the most convenient option, period.
When was ever "managing a team of devs" fun?
Especially for people who liked coding, management was always dreadful. If you went for it, you went for career reasons, not because it was "fun".
reply