Zanfa

You’re right! They’re all using the same API, there’s no other better “opt-in” API. Some developers just want to make the UX worse for their own nefarious purposes. Nothing to do with Apple.

One of the things that blew my mind when I moved to the US from Europe were the enormous portions and the amount of grease in every single dish when eating out. Even simple salads were shiny and drenched in oil. It only takes a small percentage of excessive calories over long periods to explain the obesity epidemic.

> Instead of just presenting you with all of your photos so that you can upload one, you first have to click "Manage" -> "Select more photos" -> "Add the photo".

That's not a OS limitation, this is a UX dark pattern from WhatsApp that they've purposefully added to make the UX terrible to push people into granting "Full Access".

I just tested it with both "Add Photos Only" and "Limited Access" modes with Signal and iOS does exactly what you described as the perfect UX. It's literally the following:

1) Tap Add Photo in a chat

2) System photo picker appears

3) Select which photo you want in your entire gallery (not limited to photos previously granted to Signal)

4) Photo is sent to chat

Again, this is with both non-Full Access modes. I think your beef is with Meta, not Apple.

But you don’t have to do it ahead of time. When you click add photo, you get the system picker to choose the photo and once you’ve selected what you wanna grant access to, that’s it. Literally not a single menu needs to be opened, nothing needs to be configured.

Any UX other than this is something the app developer has implemented on top. iOS works exactly like you described.

> I click “add photo”, the system dialog opens, I select a photo, and then that gets sent to the app. Somehow, Apple managed to screw that up.

That’s exactly how it works for me in iOS at the moment.

In addition, I can see the list of photos each app has been granted access to in Settings > Privacy & Security > Photos.

Extraordinary claims require extraordinary evidence and if there's one thing we've learned is that progress in AI is not linear nor predictable. We've been a few years away from fully self-driving cars for a really long time now.

> And to be fair, the problems that Facebook had when they introduced React are not common problems at all.

That’s one of my favorites. The exact bug they described during React launch presentation, that React was supposed to help fix with the unidirectional dataflow. You know the one where unread message badges were showing up inconsistently in the UI in different places. They never managed to fix that bug in the 10 years since React was announced and I eventually left Facebook for good.

Not once have I seen a site go from SSR to SPA and been pleasantly surprised. It always trends towards worse in responsiveness and overall UX.

I’m sure you could make something work better as a SPA, but nobody does.

I’m pretty sure they used to do syntax highlighting on the server before and it was fast. Now they send down unhighlighted text that seems to choke the browser with anything but the smallest diffs.

Nobody will give you a regular business loan if you need the cash to R&D your (especially non-software) product in the first place. Even more so without personal liability or in the amounts to compete for engineers in VC-funded companies.

In my experience it's the other way around. It's typically the bottom of the barrel contractors who sabotage their own (and others) safety for a quick buck.

The same way a senior engineer is more likely to use version control, backups and automated testing rather than YOLOing vibe coded applications to prod.

Most manufacturers just don't give a shit. Had the exact same experience with a well-reviewed Acer laptop a while back, ended up getting rid of it a few months in because of constant annoyances, replaced with a MacBook Air that lasted for many years. A few years back, I got one of the popular Asus NUCs that came without networking drivers installed. I'm guessing those were on the CD that came with it, but not particularly helpful on a PC without a CD drive. The same SKU came with a variety of networking hardware from different manufacturers, without any indication of which combination I had, so trial and error it was. Zero chance non-techy people would get either working on their own.

And not based on income alone, but including their entire net worth.

It was. And it was glorious. Fast and snappy, loaded instantly and felt responsive. Then they did the full SPA redesign…

> Now they rewrite their frontend in react and it feels sluggish and slow.

And decided to drop legacy features such as <a> tags and broke browser navigation in their new code viewer. Right click on a file to open in a new tab doesn’t work.

The typical setup I’ve seen for FPV drones is RPG warheads or small mortar shells for drops. I’d love to see one drop 152/155mm shells though.

They stick with the nominative case. That’s the only safe way not to butcher somebody’s name in a language like Estonian that has 14 cases. It’s infinitely easier to update copy to use only nominative than try to apply the cases automatically.

> I don’t know how much clearer it could be.

Maybe not give it a misleading name that implies full self-driving capabilities. Also not have the CEO publicly make grandiose claims of the performance over 8 years.

> If a system were more error prone, people would not trust it so much.

Unfortunately not. Youtube is full of videos of FSD trying to crash into oncoming traffic, parked cars etc, but then at the end of the video the driver goes “well that was pretty impressive” and just ignores all the suicide attempts.

LLMs would also need to use historic commits as context, rather than just the current state of the codebase in isolation. Most codebases I've worked with go through migrations from a legacy pattern A to a newer and better pattern B, used across different parts of the codebase. Rarely can these migrations be done in a single go, so both patterns tend to stick around for a while as old code is revisited. Like the HTTP example, even if LLMs pick up a pattern to follow (which they often don't), it's a coin flip whether they pick the right one or not.

We’ve had camera-triggered intersections for years now in Estonia. They seem to work equally well in daylight as well as darkness. I don’t think you need any ML for it either, you basically only need to detect motion within a predefined region which can be done with traditional CV algos quite reliably.

Please no. Whenever I see an online store as a SPA catalogue I shudder, because it usually breaks after browsing a bit in a weird state. And it resets to somewhere random should you hit back, refresh or try to send a link to somebody.

Does a real-world example of bandwidth saving even exist for SPAs? It’s always the other way around where what could’ve been a single page load, ends up being 6 separate asynchronous calls to different APIs to fetch random bits and pieces while the user stares at spinners.

I’m sorry, but calling anything but above-replacement birth rates “slow-motion extinction” is ridiculous. This is the equivalent to expecting babies to be 15 meters tall by the time they’re adults from extrapolating their rate of growth in their infant years.

Below-replacement rate might be an economic issue around retirement, but as far as the human species goes, it’s a nothingburger at this scale. We’re not passenger pigeons.

For inspiration, STM32Cube is otherwise PoS software, but it has a pretty decent utility for exactly this for most of their STM32 MCU lineup. Why they didn’t just make it a website is beyond me, but it is what it is.

Not the GP, but there are some regulations about windows, not sure if local or EU-wide. Windows at floor level above ground level must not be fully openable or must have an outside barrier. But thats a pretty sane restriction, given those windows are basically just glass doors to nowhere.

> In the game example having the weights optimized for game A doesn't help with game B. It would be interesting to see if training for both game A and B help it understand concepts in both.

Supposedly it does both A and B worse. That's their problem statement essentially. Current SOTA models don't behave like humans would. If you took a human that's really good at A and B, chances are they're gonna pick up C much quicker than a random person off the street that hasn't even seen Atari before. With SOTA models, the random "person" does better at C than the A/B master.

According to Carmack's recent talk [0], SOTA models that have been trained on game A don't perform better or train faster on game B. Even worse, training on game B negatively affects performance in game A when returning to it.

[0] https://www.youtube.com/watch?v=3pdlTMdo7pY

> JSX really isn’t that great. Sometimes it feels like most React devs don’t know how to write a for loop or if statement because they mostly use map(), ternaries, and short-circuiting, which are not very ergonomic compared with markup-based approaches.

I'm the last person to vouch for React and the horrors it has inflicted upon web development, but the one thing it definitely got right compared to Angular (and Vue) is the use of JS for control flow, rather than fake attributes and whole new language.

Out of all the possible failure modes of fiber optic drones, scissors are pretty much the least likeliest issue you’ll encounter.

This also happens with iPhones. And under the right circumstances you can feel the same tingling sensation on skin contact as well if the other person is holding a plugged in iPhone.