Bazzite uses BTRFS, but not snapshots I think.

Opensuse microOS flavors go all-in on full system snapshots but that means they also have a bad grub encryption unlock interface (instead of Plymouth). I have had some funky things with it like it missing keystrokes and if you get it wrong once, you have to reboot your entire system instead of just retrying.

Other OS’s use home folder snapshots only or something like that?

The different variants are not quite clear to me.

Placebo is a hell of a phenomenon though lol

That is completely false about the states, but might be true about the specific citizens

https://www.axios.com/2025/06/03/immigration-enforcement-ice-deportations-states-trump

Houston literally had the highest amount of ice raids lol.

https://www.newsweek.com/map-shows-states-ice-arrest-immigration-2107629

Texas is one of the places where ice makes the most arrests lol.

But there has rarely been instances in recent history where gun nuts have used their guns against agents of the government instead of unarmed civilians. They don’t like it when they shoot back.

It’s also false

https://www.axios.com/2025/06/03/immigration-enforcement-ice-deportations-states-trump

Hey, something I can maybe help with.

Flatpak IDEs on the main system are not very useful for development. I got rid of mine entirely. I am developing firmware so it might be a bit different from your case, but what I did in have a single arch distrobox where I could install everything embedded-dev-related that had to work together (JLink, nordic tools, code-oss, etc…) on that. Then a few standalone debugging tools like STLink and Saelae logic2 could be installed to the home folder by default and Code could still find them from the distrobox (but they could be installed in the distrobox also). It doesn’t even need to have an init system, but I ran into a few problems like having to manually chmod usb devices to give STLink access. Udev rules are also hit or miss in /etc/udev/rules.d, e.g. the STM udev rules just don’t work, but nordic does.

High storage consumption is likely negligible (or at least nitpicky) since storage is so cheap nowadays. Your SSD doesn’t care if it has 15GB or 20GB of system programs, especially when development codebases and SDKs, games, and media will likely make up 90% of space and almost never share libraries even on traditional systems.

But actual results and bugs have very little to do with corporate firings or open positions, as 30 years of history show us.

If corporations “think” they can fire people, with AI as an excuse, and put that cost in their pockets, they will do it. We are already seeing it in the US tech-bro sphere.

Companies will tank themselves in the medium-long term to make short term profits. Which I think is the “dev market” that OP is talking about. It shouldn’t affect the market, but it will because you have MBAs making technical decisions. I could be wrong, but the tech market is very predictable as far as behavior. They will hire a skeleton crew and work them to burnout to fix the AI slop. (Tech industry needs unions now)

It is funny because electric motors have nearly unlimited* torque depending on the kind. If you have thick enough power cables and winding conductors, you can just keep pushing it harder to get more torque.

It is like the thing they are very good at, besides sound levels, double or triple the efficiency, low/no maintenance, simpler with less parts, no emissions, etc…

Literally the only good thing about combustion engines are their fuel source energy density.

I think the problem is that motorheads see the enshittification of the auto industry as a whole and just say it’s because of electric motors because it happened right about the same time as EVs started coming out and try to push back on the wrong thing.

Yep. I have posted on stack overflow exactly 3 times. One time it was marked as duplicate and referenced to something that was not even the same topic. One time I had too much detail and debugging done for the classic knowitalls to come make a smartass remark and was completely ignored. The final time I got one comment, addressed it, and that person was never heard from again lol.

Nice! Always good to convert something that is no longer supported.

For future projects, I would also suggest using the SEEED studio Xiao series https://www.tinytronics.nl/index.php?route=product%2Fsearch&search=seeed+studio+xiao if you don’t need a lot of IO, which is often the case with IOT stuff. I use it for everything HomeAssistant related.

The esp32-c3 version for the cheapest with a nice antenna if you have to put the module near metal (my homemade doorbell uses this)

The esp32-S3 model for more processing power (my VoiceAssistant satellite uses this)

Esp32-c6 variant to replace the esp32-c3 for everything that you want to convert to Thread in the future (esphome is starting to have Thread support https://esphome.io/components/openthread/)

Esp32 board are great, but the xiao series is so tiny that you never have to worry about them not fitting!

I wish I could use unattended-upgrade.

It literally restarts my server even when I disable the option, leaving it hung if the USB boot key isn’t in there.

I had to stop using it, so now I just manually upgrade because that doesn’t auto-restart without my permission…

Yes, but I am also of the opinion that not one single acronym should be used without at least once in the section saying what the acronym is. Many many programing docs with say what am acronym is exactly once, somewhere in the docs, and then never again.

Also, if there are more complex concepts that they use that they don’t explain, a link to a good explanation of what it is (so one doesn’t have to sift through mountains of crap to find out what the hell it does). Archwiki docs do this very well. Every page is literally full of links so you can almost always brush up on the concepts if you are unfamiliar.

There seem to be 10 extremely low quality, badly written, low effort docs for every 1 good documentation center out there. It is hard to RTFM when the manual skips 90% of the library and gives an auto-generated api reference with no or cryptic explanations on parameters, for example.

But on this threat model? Why would it not be good?

It has to physically accessed on the PCB itself from what I gather.

There are 2 “threats” from what I see:

• someone at the distribution facility pops it open and has the know how to install malware on it (very very unlikely)

• someone breaks into your home unnoticed and has the time to carefully take apart your vacuum and upload pre-prepared malware instead of just sticking an IP camera somewhere. If this actually happens, the owner has much much bigger problems and the vacuum is the least of their worries.

The homeowner is the other person that can access it and it is a big feature in that case.

I miss the Compose key so much on windows for my work computer

Hell, a 12TB WD red Plus in the EU is 300€. $160 for a 14TB is absolute dirt cheap

They sell tons of stuff that isn’t food.

Clothes and work & running gear(even Aldi merch once a year), budget electronics and appliances (their toaster ovens are very often used for DIY solder reflow ovens), garden tools and supplies like slug pellets and shears, and even lower quality tools that work fine for ocassional light use.

That is the largest overstatement of the year.

It rivals aliexpress/amazon brands in quality, sure, but it still generally isn’t even at the quality of Ryobi, which is a lower-medium tier brand together with Bosch green, much less makita, dewalt, or Bosch blue.

Though for many non-renovating homeowners just needing the occasional drill work, it works just fine!

I am still relatively inexperienced and only embedded. (Electronics by trade) I am working on an embedded project with Zephyr now.

If I run into a problem I kind of do this method (e.g. trying to figure out when to use mutexes vs semaphores vs library header file booleans for checking ):

• first look in the zephyr docs at mutexes and see if that clears it up

• second search ecosia/ddg for things like “Zephyr when to use global boolean vs mutex in thread syncing”

• if none of those work, I will ask AI, and then it often gives enough context that I can see if it is logical or not (in this case, it was better to use a semi-global boolean to check if a specific thread had seen the next message in the queue, and protect the boolean with a mutex to know if that thread was currently busy processing the data), but then it also gave options like using a gate check instead of a mutex, which is dumb because it doesn’t exist in zephyr.

For new topics if I can’t find a video or application note that doesn’t assume too much knowledge or use jargon I am not yet familiar with, I will use AI to become familiar with the basic concept in the terms so that I can then go on to other, better resources.

In engineering and programming, jargon is constant and makes topic introduction quite difficult if they don’t explain it in the beginning.

I never use it for code with the exception of codebases that are ingested but with no documentation on all of the keys available, or like in zephyr where macro magic is very difficult to navigate to what it actually does and isn’t often documented at all.

There have been some big changes at brother in the past couple of years. I am holding out hope that they don’t enshittify and go the way of the others with firmware updates.

On the bottles website, it says that the bottles are sandboxes. It has a full subsystem container for each program that is isolated from the main system (according to them I guess).

If you run it through something like bottles offer a bit of protection in that respect?