Assembly code isn’t ”compiled”. To quote wikipedia:

Assembly code is converted into executable machine code by a utility program referred to as an assembler. […] The conversion process is referred to as assembly, as in assembling the source code. The computational step when an assembler is processing a program is called assembly time.

My point is that people are already trying to replace compilers with LLMs, and yes, it is stupid as shit, but still real.

Also, maybe see if an LLM can help you detect sarcasm without explicit content tagging.

Yes, but why write C when we can let the LLM write assembly directly? Compilers are unnecessary bloat we don’t have a need for with modern technology.

Ricoh GRiii

Sooo, Ricoh GRiii is actually amazing. I barely use my phone for photos anymore.

This is the one thing holding open source back, and the thing Linux users keep pissing on without understanding it.

The CTOs inept nephew can manage your fleet of windows machines and you get all the checkbox security you need for compliance (and some real security):

• Centrally managed
• Logs that are fairly hard to manipulate
• SecureBoot + Bitlocker

That same feature set on Linux will cost you a ton of money in skilled staff if you want to check the same compliance checkboxes. (As for real security, who cares, no one is doing that anyway)

Kind regards: someone who has managed Linux fleets.

Old sailfish user here, currently on an iPhone. The following things are why I switched.

• No support for Bluetooth LE, either natively or in the Android emulation. (So no smart watches, bike computers, etc.)
• Bluetooth headphones worked fine. Mostly. And not with any good codecs.
• The native app ecosystem was small. You will need to fallback to android emulation for a lot of things. In particular, good luck getting any commercial apps natively (trains, banks, and so on).
• The native apps that existed were clearly hobby projects, mostly written by one person for the use of that person. Which is fine, but none of the apps I used left enough of an impression for me to remember them.
• The web browser was fine to use, but absolutely not up to standard w.r.t. security (at that time). Plenty of known issues in the underlaying engine.
• No decent native app for maps. I never tried a native app that managed over 15fps when scrolling around. I think I also had issues with position data in android apps…
• Finally, I was constantly switching between two very different user experienced - emulated android and native sailfish.

Native sailfish was absolutely wonderful. It worked flawlessly when I stayed within the fearures that were available natively. But in modern society you simply can’t get away without all the android/ios stuff. At least not unless you have a lot of time and energy to fight the system, and I currently don’t have that.

This was a few years ago, so things may have changed for the better. I’ve been busy and haven’t kept track.

But if you want something to do - write an amazing maps app for native sailfish or get the Bluetooth LE-stuff working.

The wind energy used would eventually dissipate into heat anyway, this just puts AI in the middle.

In general: all computers turn electricity into heat. After that you get to deal with the heat, which will ironically generate even more heat (fans, pumps, etc.). This is Power Usage Effectiveness (PUE). A shitty datacenter might spend over 30% of the energy on cooling. Of course, any compute will end up heating up things, so moving the heat away efficiently is a huge deal.

I think dunkin a datacenter in water and powering it with wind is probably about as energy efficient as you can possibly get without moving into heat reuse (which we should):

• Sea water is cold, so the temperature differential is large. This means efficient and cheap heat transfer.
• Sea water moves, so you don’t need to actively pump it away or rely on convection to move the heat away.

Now, the more efficient way to deal with this is that modern liquid cooled hardware can go rather hot. Hot enough to do something with the waste heat. We can heat a campus, contribute to district cooling, etc. This is of course mostly relevant in cold climates, and often only during winter.

This design is probably a reasonably good idea from an energy efficiency standpoint. From an ecological standpoint, probably not, but I’m not qualified to answer that question.

We should do something better with our limited resources than generating AI-slop, but there are plenty of other more legitimate compute uses (like climate research) that could benefit from similar setups.

That javascript hole was probably caused by a bicycle.

To be fair, I’ve seen this done a long time before LLM slopped everything, but I think they ran emacs as PID2 in that variation.

Anyway, good luck doing it in vimscript.

Jokes aside, vim as PID 1 is just a bad idea.

Emacs on the other hand: https://github.com/emacs-os/el-init

Look - I can’t prevent my mom from being on facebook and playing candy crush. Nothing I say or do will make that happen. I can improve the situation by:

• Introducing alternatives and hope they spread (Chat with your mom on Signal)
• Reducing data harvesting during ”passive” behaviour (e.g. reduced permissions for apps. Graphene is probably the best here, but good luck getting your mom on that)
• Reducing data harvesting by the phone vendor (Samsung, Google, Apple). This is primarily done by buying an iPhone, simply due to incentives. (Again, good luck getting your mom on Graphene).

If I go too hard on my mom, she’ll just buy herself a cheap chinese android without telling me. Is that better?

Let’s stop perfect getting in the way of better.

For the threat models and data harvesting the general consumer (i.e. our moms) will face, MacOS does a far better job than Windows and iOS far better than Android (and no, your mom isn’t actually using a pixel with Graphene. Maybe she could, but she isn’t. Not really.)

If Apple can’t satisfy your threat model and privacy posturing, fine. But don’t assume everyone’s requirements are the same as yours, that’s how we scare people away.

Don’t pay for it. But use it. A lot.

GPUs are very expensive, so if you have them generating (for free) short stories of happy little kittens riding the subway in Manhattan or whatever, you are costing them a lot of money. Just don’t give them any data.

”Eventually”

As if Teams was ever a usable product.

Space ain’t happening.

I can see the point of underwater datacenters though, for some very specific use cases. Compute heavy workloads with high energy densities could possibly make sense to ”free cool” below water. DLC everything and pump the heat straight into the ocean.

Ok, while most of these don’t have companies behind them with huge revenues, most work on these projects is done by paid developers, with money coming from sponsorships, grants, donations and support deals. (Or in the case of Linux - device drivers are a prerequisite for anyone buying your product).

Developers getting paid to work on open source is a good thing. These projects may have begun their life as small hobby projects - they aren’t anymore. (And that’s probably good)

They most likely run smaller pools and have their redundancy and replication provided by the application layers on top, replicating everything globally. The larger you go in scale, the further up in the stack you can move your redundancy and the less you need to care about resilience at the lower levels of abstraction.

ZFS is fairly slow on SSDs and BTRFS will probably beat it in a drag race. But ZFS won’t loose your data. Basically, if you want between a handful TB and a few PB stored with high reliability on a single system, along with ”modest” performance requirements, ZFS is king.

As for the defaults - BTRFS isn’t licence encumbered like ZFS, so BTRFS can be more easily integrated. Additionally, ZFS performs best when it can use a fairly large hunk of RAM for caching - not ideal for most people. One GB RAM per TB usable disk is the usual recommendation here, but less usually works fine. It also doesn’t use the ”normal” page cache, so the cache doesn’t behave in a manner people are used to.

ZFS is a filesystem for when you actually care about your data, not something you use as a boot drive, so something else makes sense as a default. Most ZFS deployments I’ve seen just boot from any old ext4 drive. As I said, BTRFS plays in the same league as Ext4 and XFS - boot drives and small deployments. ZFS meanwhile will happily swallow a few enclosures of SAS-drives into a single filesystem and never loose a bit.

tl;dr If you want reasonable data resilience and want raid 1 - BTRFS should work fine. You get some checksumming and modern things. As soon as you go above two drives and want to run raid5/6 you really want to use ZFS.

Look, there is a reason everyone who actually knows this stuff use ZFS. A good reason. ZFS is really fucking good and BTRFS has absolutely nothing on it. It’s a toy in comparison. ZFS is the gold standard in this class.

You have four sane options:

• mdraid raid5 with BTRFS on top. Raid5 on BTRFS still isn’t stable as far as I know, not even in 2026.
• Mirror or triple mirror with mdraid. Have the third drive in the pool as more redundancy or outside the pool as separate unraided filesystem.
• Same as above, but BTRFS. Raid1 is stable.
• ZFS RaidZ1 (=raid5)

(Not sure about bit rot recovery when running BTRFS on mdraid. All variants should at least have bit rot detection.)

To reiterate, every storage professional I know has a ZFS-pool at home (and probably everywhere else they can have it, including production pools). They group BTRFS with Ext3, if they even know about it. When I built my home server, the distro and hardware was selected around running ZFS. Distros without good support for ZFS were disregarded right away.

I started experimenting with the spice the past week. Went ahead and tried to vibe code a small toy project in C++. It’s weird. I’ve got some experience teaching programming, this is exactly like teaching beginners - except that the syntax is almost flawless and it writes fast. The reasoning and design capabilities on the other hand - ”like a child” is actually an apt description.

I don’t really know what to think yet. The ability to automate refactoring across a project in a more ”free” way than an IDE is kinda nice. While I enjoy programming, data structures and algorithms, I kinda get bored at the ”write code”-part, so really spicy autocomplete is getting me far more progress than usual for my hobby projects so far.

On the other hand, holy spaghetti monster, the code you get if you let it run free. All the people prompting based on what feature they want the thing to add will create absolutely horrible piles of garbage. On the other hand, if I prompt with a decent specification of the code I want, I get code somewhat close to what I want, and given an iteration or two I’m usually fairly happy. I think I can get used to the spicy autocomplete.

Ahh, good old /opt/