spinningslate

True. But that’s advertising the product, not advertising in the product. It’s the latter that’s happening here. Imagine spending hundreds of thousands on a Ferrari, starting it up, and … finding an ad on the touch screen.

Even if they’re not:

> They make great devices

They make great hardware. They used to make great devices - but a great device needs the whole user experience to be great. Ad-ridden UI does not make for a great device. Doesn’t matter if it’s first or 3rd party ads.

It’s a great shame. Apple used to be the design standard to aspire to. Now it seems to have embraced enshitification. All because the market needs that share price to keep growing. And lacking leadership good enough to find another way.

“To us” is pivotal there. Continuing GP’s thought experiment: what if the model that produced the output perceives that the human it was presented to offers no value in helping it learn further?

I find typst a joy to use, either in the online editor or locally. It’s the best combination I’ve yet come across of “write what you mean” - simple markdown-like syntax - with document layout and styling. Happily pay for a subscription even though most of the time I don’t use the online editor. Software worth paying for.

It’s going to happen in windows and most likely Mac too. In both cases, all your data will be shipped off to the model providers and then out into the panopticon.

The open source desktop environments don’t have the profit motivator that drives the commercial alternatives. They’re driven to deliver their interpretation of great UI. We can each debate whether we agree with the vision or execution but that’s the motive.

If they can employ AI to help remove desktop friction - and it’s local and not shipping my data off to the surveillance machine - I’d be open to it.

In fact, if it could sort out KDE wallet’s amnesia for my wifi password, I’d give it a roll now…

that's a good question. UML - at least the mainstream version pushed by e.g. Rational - failed for many reasons. A couple pertinent to this discussion:

1. Lack of closed loop between the "spec" and working code (your spec rot point). The Rational Unified Process(RUP) was a grossly inefficient, heavily manual undertaking. Mapping between artefacts - e.g. "Platform Independent Models" and "Platform Specific Models" was a manual, largely heuristic based approach. As a consequence the models were not generally kept up to date as the project evolved.

2. User experience mismatch. Developers were asked to create diagrams instead of writing code. Tool usability was poor ("write code with a mouse") and the artefacts didn't fit well with necessary tools like diffing and source code control (try diffing an xml file textually).

Coding agents have some potential for alleviating (1) in that they can read the result code and, at least to some extent, ensure spec and code are in sync.

(2) is more open. Some users - those proportionally more interested in solving the problem than designing/writing code - are more comfortable with natural-language-based specs and exploration. Those more experienced/comfortable with code will likely see those specs more akin to UML diagrams: a distraction from the real thing.

Same here:

> Colour is the namespace for values of A Distinguished Colour. Color is evidence that the source has arrived without a proper education.

Laughed unreasonably hard at that. Been a less than pleasantly civilised day at work, this was the relief I didn’t know I needed.

Exactly. Seems naive at best for an investment consultant to look narrowly at current model capability and not consider the broader market. For example:

1. The hyperscalers are in a positive reinforcement loop. Despite any suggestions to the contrary they keep getting bigger. And can, er, “influence” government policy/officials and anything else needed to keep it that way.

2. The frontier labs and their investors. Another self-fulfilling reinforcement loop. Witness the circular gymnastics among OAI/Anthropic, Microsoft/Amazon and Nvidia

3. Data. No-one believes that Zuckerberg and co are going to say “great, we can just run the models on devices we don’t own and stop the surveillance economy because, y’know, privacy matters and we really care about mental health”.

And then there’s data centre locations and “yeah but jobs” even though your power bills are going up, and “why run your own data centre Mrs CTO, let us do it for you and save all that capex and those pesky employees you need to do it”.

Don’t get me wrong: I’m rooting for local, open weight/source models. But “hey look they benchmark well” is an unhelpfully narrow basis to forecast the demise of central hyperscaler hosting.

> an environment that favors profit above everything will naturally select for individuals who are unencumbered by such hindrances.

Exactly. Dairy farms optimise for milk production so favour cows that produce the most milk.

The market economy optimises for profit so favours those most willing/able to generate it. Zuckerberg, Musk, Thiel, Andreesen and co are products of the system.

I’d hazard a guess they might be referring to an ex-British politician who went on to have a high profile role in comms at meta.

> Id honestly wager that at you and most of the people who are agreeing with you in the comments don't make the meta SDE salary, and if offered the 300k/year positions at Meta, even for a limited time, you would absolutely take it

Not OP but I can say with 100% truth and certainty that it wouldn’t matter how much money they offered - I would not work for meta. Some things matter more than money.

Exactly this. The “one UI to rule them all” paradigm has been a persistent, recurring flaw for decades. It probably hit its lowest (to date) with the exhortations to “mobile first design”. The motivation for that was reasonable: conventional desktop UIs of the time didn’t render on mobile. However the ensuing “mobile first” instead became “mobile only” - and consequently wide screen displays with buttons the size of elephants.

Phones and desktops are so radically different that your sedan/bulldozer analogy seems like shades of grey. It’s more like taking a Saturn V rocket to the local shop for a pint of milk.

Yes, I would like to see a full cost comparison. Transferring one time digitally will no doubt cost (a lot) less than physical manufacture and distribution. But it’s not one time transfer: it’s streaming on demand, every time each person listens to each track, because the economic model is rental not purchase.

I use streaming services. I like the flexibility and ubiquity of access. But my favourite music I still buy on cd or vinyl. Why? Because it means I’m not subject to the whims of a megacorp removing access and it means more goes to the artist. I’ve been buying music for 40 years and still listen to some of stuff I bought then. I hope to live long enough to do the same for the music I buy now.

I would like to see a full cost comparison. Centralised storage - replicated, distributed and maintained online as necessary - vs media that, once manufactured and distributed, essentially costs nothing to maintain. iPods/phones get replaced much more frequently than LPs/casettes/CDs. And that’s just the resource consumption comparison. There’s then the economic polarisation of wealth to the small handful of online music renters vs distributed ownership (of copies: the original work of art remains with the artist, at least in theory).