You can use it all day and stay well below the quota. Small context, with the right model for the job. Surgical precision.

But… At some point you shut off your brain, use the most expensive model on the highest reasoning level with your whole codebase as context and just wait for tens of minutes while it burns all the tokens. To speed this up you then send six agents to tackle the same problem from all angles.

I’m surprised to see comments of people actually using it. I know it’s been topping distrowatch forever by inflation numbers.

Also hepatitis A.

There’s an infamous usenet-thread with Torvalds and Tanenbaum arguing over monolithic vs microkernel design. I microkernel is cleaner from a separation standpoint, but that also introduces hurdles and overhead.

Linux is popular because the hardware support is pretty great. There’s few laptop/BSD combinations that work well with sleep/suspend/wifi, while just about any laptop will have everything working with a recent Linux kernel.

I think it’s spot on. Even if you use power tools you need to know what kind of screw goes where. But it sure goes a whole lot faster to use power tools.

Sounds to me from the article that this is a seasoned engineer who just doesn’t want to use the tools she’s being handed.

Remember back in the day when you’d see these little badges on websites saying Best viewed with Internet Explorer? And some sites just wouldn’t work right on other browsers?

Soon you went be using any of those shitty sites, either

No, I’d hire the one using power tools and PEX pipes. Not the one stuck in the 19th century.

If I was hiring for my hypothetical construction company I’d have a real easy time picking between an employee who gets the job done and one who refuses to use power tools.

Linux has extended quite a few system calls. Not really a problem as they mostly support the POSIX ways.

But there are a few corner cases around threading and file locks that do break on mainline Linux.

Not too big to overcome, as there are exceptions like EulerOS that are both compliant and certified.

Pros:

• I have the source. I don’t have to wait for fixes or features. I just do it myself and send a patch or PR upstream.
• I can run it on just about anything, and well.
• Sane defaults and handling of user permissions - by design
• Modern filesystems that don’t silently rot your data
• Full control
• No forced updates
• No telemetry

Cons:

• Not a priority for pro applications
• Not fully POSIX compliant

I haven’t used windows in almost 30 years, but… I probably missed some games at first that DOSBox couldn’t run well (yet). Not a problem any more.

I got my first car 24 years ago. A hand-me-down that my grandpa bought with 87kkm on it, that my brother crashed. It took a lot of welding to get that car to pass inspection.

Four cars later I still have never owned a new car. I aim to buy 10yo without a loan. My current one no longer gets updated maps for its built-in infotainment system.

If I buy a 10yo EV it’ll definitely need a new battery pack. That changes the economy completely. I guess it’s cheaper to drive so I’d shell out for 10yo of driving in advance.

How about not rounding up hundreds of highly intelligent beings and their families to murder them for no reason.

If they’re so intelligent, why do they let themselves be rounded up to be murdered year after year?

Remember, Mario has a brother who wears green.

The fuck is Starmer, prime minister of the UK supposed to do about it?!

I read it as he hadn’t issued a statement about the British citizens on the flotilla who were also abducted. That seems like a PM thing to do.

zram makes sense if you do not have swap.

zswap is probably enabled by default in most distros. It compresses cold pages on the fly so that they’re ready to quickly get swapped in and out.

I do hit the swap partition occasionally on my 32GB systems.

It doesn’t really kick in until you have proper pressure. I want my swap partition for hibernation, anyway.

I wouldn’t it mind it if I was on a free tier. I’d think some of my subscription moneys would go to the creators.

Well, spotify. I did a run with spot-dl before ending my subscription.

I self-host with navidrome and stream/sync to my phone with Tempus.

But 97% of my listening nowadays is just streaming HYPR demoscene radio.

Millennial here.

I grew up with sneakernet through irc-napster-kazaa-limewire-directconnect-bittorrent-oneswarm. I gave all that up when netflix and spotify.

Those subscriptions have been ended a couple of years back and the eye patch is back on.

Netflix’ catalogue has just diminished, as everyone who owns rights to the good stuff want to do their own streaming service.

I wasn’t really listening that much ro spotify, but when they started injecting ads into podcasts I bid adieu. (Yes, injected - I’d listen to an English podcast and get very local ads between segments).

Only if it’s double-layered.

Single-layered are 4.7GB.

My first distro shipped with pico, so I often choose nano for an edit. Micro is fine, too, but I won’t go out of my way to install it.

I challenged myself to write a couple of projects in neovim over a month or so. I finished my projects, but it still felt like the tool was getting in my way. Muscle memory is to break. Current job requires a bigger IDE, but I still do my commit messages with neovim.