Derbasti

How much is a media service worth? How much does it cost to produce? Can I pay a reasonable fee to the right people?

Most websites do not offer reasonable payment options. They'd earn fractions of a cent from the ads they'd show me, but the cheapest subscriptions they offer are several dollars.

On YouTube, the value of the service is provided by creators, but too little of the subscription is going towards the creators. To make matters worse, Google seems to pull every string they can to make creators as miserable as possible. Their actions are a detriment to the service, and not worth supporting. An 80/20 revenue split would seem much more reasonable.

I've done both. Let me tell you, embedding Lua into a C program is magnitudes easier than embedding Python.

The main reason is the endless fragility of Python's garbage collector. To be clear, the API is trying to be as helpful as possible, but it's still a whole bunch of complexity that's easy to mess up. Incidentally, this discussion is left out in the linked article, which makes it less than useless. In my experience with many a third party C/Py interface, data leaks are incredibly common in such code.

Lua of course also has a garbage collector, but it essentially only knows two types of values: POD, and tables, with only the latter needing much consideration. The interaction model is based on a stack-based virtual machine, which is more complex than Python's full-function abstraction, but conveniently hides most of the garbage collector complexity. So long as you're just reshuffling things on the stack (i.e. most of the time), you don't need to worry about the garbage collector at all.

Especially at work, I find existing solutions often lacking. We tend to overestimate the complexity of reinventing many things, and underestimate the cost of ill-fitting abstractions.

In particular, Google-scale frameworks are usually harmfully over-engineered for smaller companies. Their challenge is solving the Google scale, not the problem at hand. At the same time, their complexity implies that a small-scale solution will likewise require a hundred-programmer team to implement.

But I find that far from the truth. In fact, many problems benefit from a purpose-built, small-scale solution. I've seen this work out many times, and result in much simpler, easier-to-debug code. Google-scale frameworks are a very bad proxy for estimating the complexity of a task.

Could you trivially circumvent this by running Teams in the browser?

I finished my 30 day trial the other week, and went back to DDG. After a few days, I realized I didn't miss anything, so I'll happily stay with DDG. Perhaps I'm not a very discerning searcher. Most of my searches are bang-searches of Wikipedia or CPP or Python anyway.

Still, I'd be fine with supporting a sustainable search engine. $10/more is a bit too steep for my liking, though, measured against the utility I get from it.

> If not, they will have blunders and they will lose Europe since people are willing to look for alternatives as USA gets increasingly unpopular among the Europeans due to politics.

But what alternative? There is no European smartphone OS. Windows and Steam OS and XBox are US-american, too.

I suppose Linux, Playstation, and Nintendo, then?

Anecdotally, it does somewhat fine with C++ and OpenCV. Python is great, though PySide much less so. QML is almost useless. GLSL is OK, but not great. It generally can't reason well across language barriers, although it's great at converting between languages.

The deeper into a nerdy domain I go, the less likely it is to understand what's going on. And more broadly, it seems that usefulness steeply declines for larger files and projects. It just can't fit enough context into its window to make sense of complicated things.

As an extreme example, I told it to average two angles, and it just took the mean of them. Even after prompting, it thought nothing wrong of that. (The average of 1° and 359° is 0°, not 180°.) So it goes for many domains outside of webdev, UI, data science, scripting.

Personally, I've seen junior developers most negatively impacted by AI. It seems to stymie learning and reasoning skills.

I disagree on all counts.

I quit using Amazon a few years ago. For almost all goods, there are other German online retailers that are as convenient, for the same or better prices. Having to sign up for them is a small price to pay for not having to deal with Amazon's "search" function. Plus, many stores implement account-less Paypal/GPay/APay checkout.

The only thing I use Amazon for these days, is weird Alibaba imports that are not available from reputable stores. But even there, Ebay or Alibaba itself are usually more convenient.

The commercial license used to be de-facto optional, now it is explicitly so. Good for them to officially acknowledge that reality. I bet they realized that payment was never enforceable, so it just didn't make sense to keep up the fiction.

In general, the air moves with the ground, so the earth's rotation does not affect airplanes.

However, rotation of the earth imparts a coriolis force on the air, which results in jetstream winds. Aircraft routes are optimized to use/avoid jetstreams for shorter travel times.

Both of these are film terms.

Also, thoroughly abandoned for half a decade. No recent camera support whatsoever, no feature updates.

I edit with Darktable, but organize with DigiKam. Not that Darktable is bad, but I prefer to organize my rendered JPEGs, not my RAWs.

Capture One or DxO. These are the only ones as good as Lightroom in their rendering. I'm afraid they are both much more expensive.

ON1, ACDSee, Luminar, Zoner, Silkypix, Radiant, RAW Power, Photomator are other options, but IMO have rather significant flaws that make them less good.

Frankly, Ansel is outdated by now. No color equalizer, no camera styles, no Sigmoid.

Mainline Darktable has overtaken it.

If he's so hung up on the open source process of Darktable, I can't wait to hear what he has to say about enterprise software, LOL.

I always keep very sparse notes with pen and paper. Once the dust settles, I transcribe them in full sentences into my digital journal.

This act of reconstructing a coherent narrative from disparate events is an enormously useful part of writing the journal.

That's a good observation. How would you teach an LLM the processes and conduct in your company? I suppose you'd need to replace code reviews and tutoring with prompt engineering. And hope that the next software update to the LLM won't invalidate your prompts.

Not to diminish Fairphone's accomplishments here, but beginning in 2025, the EU mandates five years of software updates for all phones sold in the EU.

Awesome! This will finally let me use bluetooth devices behind a KVM!

That was a delightful blog post, thank you!

I migrated from pinboard to raindrop this year, can confirm.

I was on Pinboard's lifetime plan. No subscription, just a single payment of $15 back who knows when. A few years ago I was asked if I wouldn't rather pay $5 per month instead, and declined. I wonder if all the outages and throttling were a result of underestimating the running costs.

Regardless, service quality had declined to such levels that I switched to raindrop this year. Ironically, raindrop is free to use, and may run into the same issue eventually.

There's a reason why I took my inverter offline after making sure that it was installed correctly. A cheap power meter now serves to measure my power generation instead.

It does. The second half is written in C. Java is just used for the simplified first part.

A thought experiment:

A piece of code takes 6h to develop in C++, and 1h to run.

The same algorithm takes 3h to code in Python, but 6h to run.

If I could thread-spam that Python code on my 24 core machine, going Python would make sense. I've certainly been in such situations a few times.

And that's exactly why it's so cool to see this in the Surface tablets.

...he writes, on his Snapdragon Surface tablet.

(Also, I find it hella entertaining that a snapdragon is a cute little flower, martial naming notwithstanding)

On the other hand, the throughput requirements on image processing are often far more stringent than in audio. A full-resolution stereo audio stream is only 200kb/s = 12 Mb/min after all, whereas a video system might have to chew throw many gigabytes in the same time.

This is great fun! But it's much more prevalent in embedded DSP than desktop.

When I was working on desktop audio software, our rule of thumb was to keep CPU occupancy below 30%. More than that, and you'd be sure to get hitching in your audio stream. (This factors in thread contention, other background tasks stealing CPU cycles, file system pauses...)

A much different experience from embedded programming, where 99% occupancy is no problem at all.