jamil7

Kind of the same feeling, I don't use all of services they offer either and when I looked at self-hosting, it all seemed kind of heavy and fragile to self-host. I ended up replicating the parts I used with a small API layer connected to a managed postgres db for a tenth of the cost or something. I'd say it's pretty handy for prototyping but not sure I'd want to build a business on the back of it.

I’m not an expert but isn’t this not really comparable since traders are typically trying to generate cashflow?

I think there’s a huge amount of inefficiency all the way through the software stack due to decades of cheap energy and rapidly improving hardware. I would expect with hardware and energy constraints that we will need to look for deeper optimisations in software.

> You need to get private healthcare

No you don't, you can stay in the public system.

> you need to save up for retirement and so on

You can still contribute voluntarily to the public pension if you want to but the majority of self-employed people here don't as this is seen as a feature not a bug.

Part of the concept of being self-employed in Germany is that you're opting out of the welfare state and are responsible for taking care of yourself. With the potential upside of better raw earning potential.

Oh nice, I once built a very simple version of something like this to be able to run a whole lot of existing JS code on mobile.

I’ve landed on a few similar techniques and have been using unit tests quite a bit as a guardrail for LLMs. One thing that’s useful when using aider is alternating between the /add and /read-only contexts so that it can only edit the tests or the code but “see” both.

I generally agree it’s decreased steadily but I also remember macOS 9 and early X versions especially being pretty buggy and having awful performance.

They’re maybe more convenient for the shop but the lines seem equally long where I am as people aren’t as fast as a cashier. I don’t notice any difference in efficiency other than I now need to scan and look everything up myself. So a net loss for me.

I feel very differently about the article. What I took away from it was the authors long-held critique of how high school literature and writing is taught and evaluated. Or how teaching in general has become increasingly outcome focused rather than process driven. ChatGPT is used to illustrate this but I feel like the article could have been written using another example like plagiarism or paying someone to write for you or any current LLM.

> All this recent hype about sync engines and local first applications

Kind of but only really in the web world, it was the default on desktop for a long time and is pretty common on mobile.

Thanks, super helpful. I tried it out and had similar observations, in my case, I also work on a Swift codebase but with a few internal dependencies that are in separate repos, Aider can handle this by pointing it to files in other repos with the read-only command. I wasn't able to get that working with Claude code. I also found the "magic" dependency resolution part of Claude code to be a little wasteful/poor, at least in my codebase in which it chewed through time and tokens searching for the correct files.

How does it compare to Aider with the same model?

I agree with you, just my obsevations from using SwiftUI "in anger" for the last few years, performance is atrocious on macOS and acceptable on iOS, especially with some tuning.

Yeah it's in a really bad spot on macOS still, iOS performance is a lot better. I'm sure all of this can be overcome but I'm not certain it's a priority for them.

What sort of citation would you be looking for here? It's generally accepted that it's not a priority for them, the discrepencies in performance and features between UIKit vs AppKit have always existed and are even more pronounced in SwiftUI. Due to how it's used they can't apply the same business model to the mac as they can iOS and therefore can't extract as much value from it, this is reflected in revenue.

There's always been a lot of room for them to optimise performance on macOS to a level that they do iOS, but modern Apple basically wants macOS to go away and sees it as more of an annoyance, so they don't spend any time tuning it. Then the elephant in the room is SwiftUI itself, there's a performance cliff there, wherein if you've optimised everything else, you might just hit the brick wall that's the layout engine with no recourse or ability to even peak under the hood. We're at a point now where building a fully native macOS app, with the first-party toolchain, will give you far worse performance (in terms of responsiveness not resource usage) than something like Electron. I have a suspicion teams inside of Apple are also running up against these issues as they start to actually adopt the framework.

Yeah, it's the reverse of what's often pushed in the media. There have even been initiatives for autonomous EU security projects, with the US (I think even Trump at some point, despite what he says now about NATO) being against anything that would undermine NATO.

The first part of this comment is a bit of a straw man, but I'll bite.

The reason is that the discussion is nuanced, a smartphone plays a very different role in someone's life than a Nintendo Switch. For better or worse, a smartphone has become a requirement for critical services that we all rely on.

You can keep using an older version of Xcode if you like. I mean, every other tool chain that I can think of does more or less the same thing. There are plenty of reasons to criticise Apple's developer tooling and relations, but I don't see this as being especially different to other platforms

This isn’t true you can set the target multiple versions back. The main problem right now is a huge amount of churn in the language, APIs and multiple UI frameworks means everything is a moving target. SwiftUI has only really become useable in the last coupe of versions.

> It's a sad state of affairs that AI is taking over arts/music/creativity stranding people with boring, routine, meaningless jobs"

So far it’s not though.

> The problem is AI current best use case is creative work, art, music, programming

This is where it’s being pushed and marketed but I’m not actually sure it’s the best use case.

> Why? Either they're idiots who think they need control or they are tyrants

Many politicians are individuals without any talent who desire power and control, politics is the only avenue open to people like that.

> the electricity is usually produced by coal plants or rare earth mined renewables that have an even greater environmental burden

Renewables don't typically rely on rare earths, maybe small amounts in some wind turbine magnets.

Yeah this is a fancier version of what we do at work, although we just poll the event table.

Surely the landfill employees would have recovered it and divvied it up if it was ever recoverable or even there.

I agree but have been on teams where this slows down PR review or arguments break out. Delegating to the linter and format on save can get the team past this.

> a worse Llama?

As someone who lives here, I'd actually be surprised if we even got that. I expect lots of taxpayer funded websites, manifestos, PowerPoints and numerous discussions and ultimately nothing.

> What did delaying this cost them, though?

It potentially cost the whole field in terms of innovation. For OpenAI specifically, they now need to scramble to come up with a differentiated business model that makes sense in the new landscape and can justify their valuation. OpenAI’s valuation is based on being the dominant AI company.

I think you misread my comment if you think my feelings are somehow hurt here.

Agree 100%, this was also bound to happen eventually, OpenAI could have just remained more "open" from the beginning and embraced the inevitable commoditization of these models. What did delaying this buy them?