Pulumi code ends up looking like a DSL anyway with all the stuff you end up using from the top-level pulumi package to do anything vaguely complicated.

Only now, compared with Terraform, you need to worry about resource ordering and program flow, because when you have a dependency between resources, the resource object you depend on has to be instantiated (within the program flow, I mean - Pulumi handles calculating the ordering of actual cloud resource creation) before the dependent resource. This gets old really quickly if you’re iterating on a module that creates more than a few interdependent resources. So much cut, paste, reorder. FWIW CDK has the same issue, and for the same reason - because it’s using a general-purpose programming language to model a domain which it doesn’t fit all that well.

I like Pulumi and it’s got a lot going for it, especially if you have complex infrastructure requirements. You get a bunch of little quality of life enhancements that I wish Terraform would adopt, like cloud state management by default, and a built-in mechanism for managing secrets in a sane way. Python/TypeScript etc. modules are much more flexible than Terraform modules, and really help with building large chunks of reusable infrastructure. The extra programmability can be useful, though you need to be extra-careful of side-effects. You get more power, but you also get some extra work.

But for most people deploying a bit - or even quite a lot - of cloud infrastructure, Terraform is honestly just easier. It’s usually some fairly simple declarative config with some values passed from one resource to another, and a small amount of variation that might require some limited programmability. Which is exactly what Terraform targets with HCL. It’s clear to me that Pulumi sees this too, since they introduced the YAML syntax later on. But IMO HCL > YAML for declarative config.

Do we have to bring this up again? It’s just boring.

systemd is here and it isn’t going anywhere soon. It’s an improvement over SysV, but the core init system is arguably less well-designed than some of the other options that were on the table 10 years ago when its adoption started. The systemd userspace ecosystem has significantly stifled development of alternatives that provide equivalent functionality, which has led to less experimentation and innovation in those areas. In many cases those systemd add-on services provide less functionality than what they have replaced, but are adopted simply because they are part of the systemd ecosystem. The core unit file format is verbose and somewhat awkward, and the *ctl utilities are messy and sometimes unfriendly.

Like most Red Hat-originated software written in the last 15 years, it valiantly attempts to solve real problems with Linux, and mostly achieves that, but there are enough corner cases and short-sighted design decisions that it ends up being mediocre and somewhat annoying.

Personally I hope that someone comes along and takes the lessons learned and rewrites it, much like Pulseaudio has been replaced by Pipewire. Perhaps if someone decides it needs rewriting in Rust?

The WiFi card is probably a Realtek 8852AE, which has become very common in laptops since 2021. Unfortunately Realtek driver support tends to lag quite a bit.

If you want to run Ubuntu Desktop 22.04, then you’re probably best off waiting a few weeks for the Ubuntu Desktop 22.04.4 point release. It’s due sometime this month. It will boot and install an “HWE” (Hardware Enablement) kernel and drivers, that are based on the kernel from Ubuntu 23.04, and therefore should work out of the box with your WiFi card.

While it’s possible to upgrade an existing Ubuntu 22.04 installation with the latest HWE kernel, doing it by downloading the relevant packages on another machine and moving them across using a USB stick is going to be somewhat frustrating if you’ve not done it before. You’ll certainly learn a few things, but it may not be an enjoyable experience. I’m a grizzled Linux veteran, and I’m pretty sure I’d end up forgetting to download one or more packages and having to swap back and forth between machines.

In the meantime, I would just continue to use Ubuntu 23.04. In fact, if it was me, I would probably just stick with 23.04, upgrade to 23.10 and then subsequently 24.04 when they become available. What you do once you’re on the 24.04 LTS release is up to you. By that time, other distros will probably also work out of the box too.

At least for me, there is a big difference between naming things at home and naming things for work.

Work “pet” machines get systematic names based on function, location, ownership and/or serial/asset numbers. There aren’t very many of them these days. If they are “cattle” then they get random names, and their build is ephemeral. If they go wrong or need an upgrade, they get rebuilt and their replacement build gets a new random name. Whether they are pets or cattle, the hostnames are secondary to tags and other metadata, and in most cases the tags are used to identify the machines in the first instance, because tags are far more flexible and descriptive than a hostname.

At home, where the number of machines is limited, I know all of them like the back of my hand, and it’s mostly just me touching them, whimsical names are where it’s at.

Ungulates. Because who doesn’t like a hoofed animal?

My client machines are even-toed ungulates (order Artiodactyla) and my servers/IoT machines are odd-toed (order Perissodactyla). I’m typing this on Gazelle. My router is called Quagga, both after the extinct zebra subspecies and the routing protocol software (I don’t use it any more but hey, it’s a router).

Biological taxonomy is a great source of a huge number of systematic (and colloquial) names.

I could well be wrong about the AAC passthrough, and I should have hedged that statement with “allegedly” as I’ve not tested it myself.

To your other point though, I disagree - there are plenty of ways you could pass through an unchanged AAC bitstream, but still mix in other sounds when required. For example, having the sender duck the original bitstream out temporarily and send a mixed replacement bitstream while the other sound is playing. Or (and this would only work if you control the firmware on the receiver, but if you’re using Apple headphones with an Apple device, that’s not a problem) sending multiple bitstreams to the receiver and letting the receiver mix them.

I can only comment on my experience with my own equipment and ears, but in my experience, 990Kbps LDAC is noticeably more transparent than 256Kbps AAC for Bluetooth audio.

I can fairly reliably guess whether or not I remembered to switch my Sony XM4s out of multipoint mode the last time I used them (when in multipoint pairing mode LDAC is not supported and 256Kbps AAC is usually what gets negotiated). The difference is small, but over a few minutes of listening, the sonic signature when it’s using AAC is just a little bit “off” and my ears don’t like it as much.

Could I ABX the difference using the usual ABX setup with short samples of music I’m not familiar with? Probably not. Can I tell the difference over an extended period using music I know well, and that I often listen to uncompressed? Yes, pretty easily.

LDAC is not a particularly sophisticated codec, but it doesn’t have to be when it has a 990Kbps bitrate. It’s also possible that the FDK-AAC codec that I think both Pipewire and Android use for real-time AAC encoding is not the best tuned for 256Kbps CBR. AIUI in 256Kbps CBR mode, FDK-AAC has a hard low-pass filter at 17KHz, and I can still hear above 17KHz.

Yeah, I agree.

I bought them for their noise cancelling primarily, and they’re excellent at that, but otherwise they’re not great. The un-EQed frequency response is terrible for headphones in their price range: flabby, wildly over-exaggerated bass and no mids at all. Running without EQ I can barely hear lyrics - every singer sounds like they’re mumbling underwater. I’ve had $20 IEMs with better tonal balance. They respond well to EQ but the on-board EQ doesn’t have enough frequency bands to even come close to fixing them. Wavelet on Android doing EQ duty makes them listenable. Even when you do EQ them properly, they still sound a bit dull and lifeless.

No idea how they got so much praise when they were launched. The power of marketing budgets I guess. For a while I was gaslighting myself thinking I had a faulty pair or maybe there was something going wrong with my hearing, but having heard another pair, and doing comparisons with my other headphones - most of which are far cheaper - I realised that no, they’re just not very good as headphones.

It is worse than uncompressed, but 990Kbps LDAC is the closest codec to totally transparent I’ve heard for Bluetooth audio. AptX HD is nearly as good to my ears, and is better than 660Kbps LDAC. The differences are very small though, especially when compared with the differences on the analog side, e.g. the amp, and particularly the headphone design.

Apple side-steps the problem by, at least when you’re listening to Apple Music, simply sending the AAC stream as-is to the headphones and has them decode the audio. I don’t know why that isn’t a more common approach.

I’m still somewhat bemused that we’re talking about Bluetooth codecs at all. It surely can’t be that difficult technically to get 1.5Mbps actual throughput on Bluetooth and simply send raw 16-bit/44.1Khz PCM. 2.4Ghz WiFi is capable of hundreds of times that speed. Bluetooth has been stuck at the same speeds for decades.

I have a Radsone ES100 Bluetooth DAC/headphone amp, and that supports LDAC, multipoint, and doesn’t compromise the LDAC bitrate when you have multipoint enabled. You can even leave it plugged in as a USB DAC and still use multipoint BT with LDAC, and it switches smoothly between the sources depending on which device started playing a stream most recently.

I was distinctly underwhelmed by the BT implementation when I got my Sony XM4s, it’s kinda weak by comparison.

The UI font in MacOS is called SF Pro. If you have access to a Mac you can simply copy the .otf font files over to Linux (they are in /System/Library/Fonts on MacOS) and install and use them there.

If you don’t have access to a Mac, Google Roboto Sans is a very similar design (it was the default Android UI font for several years) and if it’s not already installed by your Linux distro, it’s freely downloadable.

MacOS font rendering is dreadful on non-Retina/HiDPI displays. If you want similar rendering on Linux, turn font hinting off, and set antialiasing to greyscale only, no subpixel rendering. It will look very similar, if not identical, to modern MacOS.

For non-Retina displays I vastly prefer FreeType’s subpixel antialiasing and “slight” hinting to what MacOS does.

That’s only true if China gets no further than attempting an amphibious landing on Taiwan. If China succeeds in creating a bridgehead on the island, then many of the same land-based weapons and systems that the US is currently supplying to Ukraine, or that Ukraine would like to have, come into play, including 155mm artillery, rocket artillery, tanks, air defence missiles, and land-based multirole aircraft like the F-16.

From a war planning point of view, unfortunately you can’t assume that China’s amphibious landing would fail. In fact, I think it’s more likely that China would succeed in establishing some kind of foothold on the island in the early stages of a future Taiwan war than not. If the amphibious force is large enough, it would be very difficult to eliminate all the landing craft, especially if there is a successful misdirection.

This is without considering that North Korea could also simultaneously launch a land-based attack on South Korea to dilute any US response in either theatre.

The real meat shields in the war in Ukraine are Russian conscripts. At least Ukrainian conscripts have the conviction that they are defending their country’s internationally recognised borders.

The US (and the rest of NATO) is being cautious for a reason, and it’s not because they’re using Ukrainians as “meat shields.”

NATO stocks of war materiel were at historically low levels before February 2022, and it’s difficult for the US to commit fully when China is sabre-rattling over Taiwan. That’s Xi’s (and Kim Jong-Un’s, to a lesser extent) gift to Putin. Sabre-rattling keeps the US from engaging fully in Ukraine, even though China won’t be ready to invade Taiwan for several years yet.

Unfortunately for Ukraine, it’ll be several years before NATO materiel stocks start to grow above 2022 levels, but they will grow.

The question is, will they grow fast enough?

Personally I’m predicting world war in 2027-28 unless the West pulls its finger out.

I pulled them from the table at https://en.wikipedia.org/wiki/Casualties_of_the_Russo-Ukrainian_War#Total_casualties which seems to be kept fairly up-to-date with the reasonably credible claims, including the Meduza estimates. Original references are in the Wikipedia article.

I added the 30000 upper bound as a fudge for the Ukrainian military deaths because the one somewhat credible estimate from the US that table lists only covers up to May 2023. While Ukraine has been on the offensive since then, I don’t think increasing the KIA numbers by more than 50% in 2 months is credible even under the circumstances.

Most Fediverse users are Western. The Western world has plenty of media diversity, and you can find virtually every viewpoint you can imagine represented there. Open criticism of government, all the way to the top, is a normal part of everyday life, and media outlets regularly criticise each other, and themselves, for bad takes and poor journalism.

Because of the diversity of media opinion, it is harder to push an agenda, so mainstream Western media does it, by and large, with substantial subtlety, building trust first, and seeding ideas over long periods of time.

Russian and Chinese media aimed at a Western audience seems brash and full of bad takes by comparison. It is rarely, if ever, critical of itself or of its own government, and also rarely provides any independently verifiable evidence for its claims. To a Western audience used to Western media, it appears so one-sided that it is laughable. That is why it is easy for people in the West to dismiss it as propaganda.

You could probably write a PhD thesis on why media outlets in China and Russia find it difficult to play the Western media game, but I think the main issue is this: If you live in a society that doesn’t itself value diversity of opinion and thought, it is difficult to produce media for a society that does value that without it seeming off-kilter. It’s a bit like the difference between being fluent in another language and “feeling” the language. To a native speaker listening to it, the difference is really obvious.

If pro-Russian media come up with a decent methodology to independently estimate Ukrainian losses, then I’m sure some discussion can be had around that.

Don’t expect the opposition to do the work for you.

Somewhere between 9000 and 42000 civilians depending on estimates, and probably 20000-30000 military, again depending on estimates.