Hello! Recently I ran a terminal survey and I asked people what frustrated them. One person commented:

There are so many pieces to having a modern terminal experience. I wish it all came out of the box.

My immediate reaction was “oh, getting a modern terminal experience isn’t that hard, you just need to….”, but the more I thought about it, the longer the “you just need to…” list got, and I kept thinking about more and more caveats.

So I thought I would write down some notes about what it means to me personally to have a “modern” terminal experience and what I think can make it hard for people to get there.

what is a “modern terminal experience”?

Here are a few things that are important to me, with which part of the system is responsible for them:

• multiline support for copy and paste: if you paste 3 commands in your shell, it should not immediatly run them all! That’s scary! (shell, terminal emulator)
• infinite shell history: if I run a command in my shell, it should be saved forever, not deleted after 500 history entries or whatever. Also I want commands to be saved to the history immediately when I run them, not only when I exit the shell session (shell)
• a useful prompt: I can’t live without having my current directory and current git branch in my prompt (shell)
• 24-bit colour: this is important to me because I find it MUCH easier to theme neovim with 24-bit colour support than in a terminal with only 256 colours (terminal emulator)
• clipboard integration between vim and my operating system so that when I copy in Firefox, I can just press p in vim to paste (text editor, maybe the OS/terminal emulator too)
• good autocomplete: for example commands like git should have command-specific autocomplete (shell)
• having colours in ls (shell config)
• a terminal theme I like: I spend a lot of time in my terminal, I want it to look nice and I want its theme to match my terminal editor’s theme. (terminal emulator, text editor)
• automatic terminal fixing: If a programs prints out some weird escape codes that mess up my terminal, I want that to automatically get reset so that my terminal doesn’t get messed up (shell)
• keybindings: I want Ctrl+left arrow to work (shell or application)
• being able to use the scroll wheel in programs like less: (terminal emulator and applications)

There are a million other terminal conveniences out there and different people value different things, but those are the ones that I would be really unhappy without.

how I achieve a “modern experience”

My basic approach is:

1. use the fish shell. Mostly don’t configure it, except to:
   • set the EDITOR environment variable to my favourite terminal editor
   • alias ls to ls --color=auto
2. use any terminal emulator with 24-bit colour support. In the past I’ve used GNOME Terminal, Terminator, and iTerm, but I’m not picky about this. I don’t really configure it other than to choose a font.
3. use neovim, with a configuration that I’ve been very slowly building over the last 9 years or so (the last time I deleted my vim config and started from scratch was 9 years ago)
4. use the base16 framework to theme everything

some “out of the box” options for a “modern” experience

What if you want a nice experience, but don’t want to spend a lot of time on configuration? Figuring out how to configure vim in a way that I was satisfied with really did take me like ten years, which is a long time!

My best ideas for how to get a reasonable terminal experience with minimal config are:

• shell: either fish or zsh with oh-my-zsh
• terminal emulator: almost anything with 24-bit colour support, for example all of these are popular:
  • linux: GNOME Terminal, Konsole, Terminator, xfce4-terminal
  • mac: iTerm (Terminal.app doesn’t have 256-colour support)
  • cross-platform: kitty, alacritty, wezterm, or ghostty
• shell config:
  • set the EDITOR environment variable to your favourite terminal text editor
  • maybe alias ls to ls --color=auto
• text editor: this is a tough one, maybe micro or helix? I haven’t used either of them seriously but they both seem like very cool projects and I think it’s amazing that you can just use all the usual GUI editor commands (Ctrl-C to copy, Ctrl-V to paste, Ctrl-A to select all) in micro and they do what you’d expect. I would probably try switching to helix except that retraining my vim muscle memory seems way too hard. Also helix doesn’t have a GUI or plugin system yet.

Personally I wouldn’t use xterm, rxvt, or Terminal.app as a terminal emulator, because I’ve found in the past that they’re missing core features (like 24-bit colour in Terminal.app’s case) that make the terminal harder to use for me.

I don’t want to pretend that getting a “modern” terminal experience is easier than it is though – I think there are two issues that make it hard. Let’s talk about them!

issue 1 with getting to a “modern” experience: the shell

bash and zsh are by far the two most popular shells, and neither of them provide a default experience that I would be happy using out of the box, for example:

• you need to customize your prompt
• they don’t come with git completions by default, you have to set them up
• by default, bash only stores 500 (!) lines of history and (at least on Mac OS) zsh is only configured to store 2000 lines, which is still not a lot
• I find bash’s tab completion very frustrating, if there’s more than one match then you can’t tab through them

And even though I love fish, the fact that it isn’t POSIX does make it hard for a lot of folks to make the switch.

Of course it’s totally possible to learn how to customize your prompt in bash or whatever, and it doesn’t even need to be that complicated (in bash I’d probably start with something like export PS1='[\u@\h \W$(__git_ps1 " (%s)")]\$ ', or maybe use starship). But each of these “not complicated” things really does add up and it’s especially tough if you need to keep your config in sync across several systems.

An extremely popular solution to getting a “modern” shell experience is oh-my-zsh. It seems like a great project and I know a lot of people use it very happily, but I’ve struggled with configuration systems like that in the past – it looks like right now the base oh-my-zsh adds about 3000 lines of config, and often I find that having an extra configuration system makes it harder to debug what’s happening when things go wrong. I personally have a tendency to use the system to add a lot of extra plugins, make my system slow, get frustrated that it’s slow, and then delete it completely and write a new config from scratch.

issue 2 with getting to a “modern” experience: the text editor

In the terminal survey I ran recently, the most popular terminal text editors by far were vim, emacs, and nano.

I think the main options for terminal text editors are:

• use vim or emacs and configure it to your liking, you can probably have any feature you want if you put in the work
• use nano and accept that you’re going to have a pretty limited experience (for example I don’t think you can select text with the mouse and then “cut” it in nano)
• use micro or helix which seem to offer a pretty good out-of-the-box experience, potentially occasionally run into issues with using a less mainstream text editor
• just avoid using a terminal text editor as much as possible, maybe use VSCode, use VSCode’s terminal for all your terminal needs, and mostly never edit files in the terminal

issue 3: individual applications

The last issue is that sometimes individual programs that I use are kind of annoying. For example on my Mac OS machine, /usr/bin/sqlite3 doesn’t support the Ctrl+Left Arrow keyboard shortcut. Fixing this to get a reasonable terminal experience in SQLite was a little complicated, I had to:

• realize why this is happening (Mac OS won’t ship GNU tools, and “Ctrl-Left arrow” support comes from GNU readline)
• find a workaround (install sqlite from homebrew, which does have readline support)
• adjust my environment (put Homebrew’s sqlite3 in my PATH)

I find that debugging application-specific issues like this is really not easy and often it doesn’t feel “worth it” – often I’ll end up just dealing with various minor inconveniences because I don’t want to spend hours investigating them. The only reason I was even able to figure this one out at all is that I’ve been spending a huge amount of time thinking about the terminal recently.

A big part of having a “modern” experience using terminal programs is just using newer terminal programs, for example I can’t be bothered to learn a keyboard shortcut to sort the columns in top, but in htop I can just click on a column heading with my mouse to sort it. So I use htop instead! But discovering new more “modern” command line tools isn’t easy (though I made a list here), finding ones that I actually like using in practice takes time, and if you’re SSHed into another machine, they won’t always be there.

everything affects everything else

Something I find tricky about configuring my terminal to make everything “nice” is that changing one seemingly small thing about my workflow can really affect everything else. For example right now I don’t use tmux. But if I needed to use tmux again (for example because I was doing a lot of work SSHed into another machine), I’d need to think about a few things, like:

• if I wanted tmux’s copy to synchronize with my system clipboard over SSH, I’d need to make sure that my terminal emulator has OSC 52 support
• if I wanted to use iTerm’s tmux integration (which makes tmux tabs into iTerm tabs), I’d need to change how I configure colours – right now I set them with a shell script that I run when my shell starts, but that means the colours get lost when restoring a tmux session.

and probably more things I haven’t thought of. “Using tmux means that I have to change how I manage my colours” sounds unlikely, but that really did happen to me and I decided “well, I don’t want to change how I manage colours right now, so I guess I’m not using that feature!”.

It’s also hard to remember which features I’m relying on – for example maybe my current terminal does have OSC 52 support and because copying from tmux over SSH has always Just Worked I don’t even realize that that’s something I need, and then it mysteriously stops working when I switch terminals.

change things slowly

Personally even though I think my setup is not that complicated, it’s taken me 20 years to get to this point! Because terminal config changes are so likely to have unexpected and hard-to-understand consequences, I’ve found that if I change a lot of terminal configuration all at once it makes it much harder to understand what went wrong if there’s a problem, which can be really disorienting.

So I usually prefer to make pretty small changes, and accept that changes can might take me a REALLY long time to get used to. For example I switched from using ls to eza a year or two ago and while I like it (because eza -l prints human-readable file sizes by default) I’m still not quite sure about it. But also sometimes it’s worth it to make a big change, like I made the switch to fish (from bash) 10 years ago and I’m very happy I did.

getting a “modern” terminal is not that easy

Trying to explain how “easy” it is to configure your terminal really just made me think that it’s kind of hard and that I still sometimes get confused.

I’ve found that there’s never one perfect way to configure things in the terminal that will be compatible with every single other thing. I just need to try stuff, figure out some kind of locally stable state that works for me, and accept that if I start using a new tool it might disrupt the system and I might need to rethink things.

The long-anticipated "year of the Linux desktop" could see renewed interest in 2025 as Microsoft's planned end of support for Windows 10 approaches, potentially driving users to explore alternatives. With Windows 10 reaching end of support in October 2025, many users will face decisions about upgrading hardware for Windows 11 or considering different operating systems entirely. Linux distributions have evolved to offer increasingly polished desktop experiences, with improving hardware compatibility and familiar user interfaces. 2024 saw Linux adoption grow thanks to the Steam Deck's success, reaching a 4.04% market share in December, up from 3.85% during the same time last year. More Linux laptops, improved gaming compatibility, and growing awareness of its benefits also contributed to its steady rise.

Read more of this story at Slashdot.

Liz Pelly, Harpers:

According to a source close to the company, Spotify’s own internal research showed that many users were not coming to the platform to listen to specific artists or albums; they just needed something to serve as a soundtrack for their days, like a study playlist or maybe a dinner soundtrack. In the lean-back listening environment that streaming had helped champion, listeners often weren’t even aware of what song or artist they were hearing. As a result, the thinking seemed to be: Why pay full-price royalties if users were only half listening? It was likely from this reasoning that the Perfect Fit Content program was created.

After at least a year of piloting, PFC was presented to Spotify editors in 2017 as one of the company’s new bets to achieve profitability. According to a former employee, just a few months later, a new column appeared on the dashboard editors used to monitor internal playlists. The dashboard was where editors could view various stats: plays, likes, skip rates, saves. And now, right at the top of the page, editors could see how successfully each playlist embraced “music commissioned to fit a certain playlist/mood with improved margins,” as PFC was described internally.

Reading this article made me feel quite sad. The musicians responsible for tracks like these seem quite talented. Browsing the roster of Epidemic Sound — one of the production companies cited in Pelly’s story — indicates many of them are working in multiple genres. I did not hear anything particularly interesting or notable, but I do not want to suggest this is bad; there is a valid place for stock tracks. And, though this article focuses on Spotify, there is no shortage of Epidemic’s music and that of similar production companies in Apple Music playlists, either.

Despite their skill, none of these musicians make very much money from Epidemic Sound, according to Pelly, or in payouts from streaming companies. That is not a unique situation — streaming notoriously pays nearly all artists poorly — though it seems like an ongoing part of the devaluation of music. Filler tracks have long been commonplace, but they used to be confined to production libraries, not placed side-by-side with recognizable artists. For many musicians to earn enough, they become contributors to their own downfall.

(Via Jason Tate.)

⌥ Permalink

After spending more than $10 billion to try to develop an autonomous robotaxi, General Motors is now calling time on the endeavor. On Tuesday afternoon, the automaker announced that it is done investing in Cruise's robotaxi development and will instead combine the startup's technical team with its internal efforts at GM.

After several years, GM has accepted the inevitable: Given the costs, there's no way to build a profitable robotaxi business. This year, GM will have spent around $2 billion on Cruise.

"GM is committed to delivering the best driving experiences to our customers in a disciplined and capital efficient manner," said Mary Barra, chair and CEO of GM. "Cruise has been an early innovator in autonomy, and the deeper integration of our teams, paired with GM’s strong brands, scale, and manufacturing strength, will help advance our vision for the future of transportation."

On a call with investors, Barra said that GM still "believe[s] in the importance of driver-assistance and autonomous driving technology in our vehicles."

GM bought into Cruise in 2016, when self-driving fever was at its peak. It had lofty goals of deploying robotaxis in a number of major US cities, and as early as 2019, it was lobbying the National Highway Traffic Safety Administration to be allowed to leave the steering wheel out of its test vehicles.

Cruise built its tech using a fleet of Chevrolet Bolts but had bigger plans for the Origin, a driverless pod purpose-designed for robotaxi work. But the Origin was laid to rest earlier this summer when GM killed the program in favor of Cruise continuing to use the Bolt.

Harder than it looked

The early timelines began to slide even before COVID, as the reality of trying to develop an autonomous car capable of handling a near-infinite number of edge cases began to bite. In 2017, one of the startup's robotaxi test cars hit a motorcyclist, but much more seriously, in 2023, a Cruise robotaxi dragged a pedestrian more than 20 feet after she was knocked into the autonomous vehicle's path by another vehicle.

That was sufficient for California to suspend Cruise's license to test driverless vehicles just three months after the state had approved its plan to charge customers for rides.

The prospects for the robotaxi industry seem bleaker than they did a few short years ago. In 2022, Ford and Volkswagen, which had each invested more than a billion dollars in Argo, realized there was no path to profitability and decided to reallocate their resources.

GM's decision will save it around a billion dollars a year, and instead of developing an entirely autonomous product, it will build on its Super Cruise technology to create a so-called "level 3" partially automated driving system similar to Mercedes-Benz's Drive Pilot. A level 3 assist allows drivers to take their eyes off the road when active and is intended for lower-speed conditions such as traffic jams.

But do customers actually want this tech? While beloved by investors, giddy with dollar signs in their eyes each time Tesla increases the price for its highly controversial Full Self Driving package, there is very little evidence the car-buying public has any such appetite. Not only is there no evidence that partial automation makes driving any safer, but hands-free driver assists also failed to even crack the top 10 desired features among new car buyers versus stuff they actually want, like wireless device charging or a panoramic glass roof.

Read full article

Comments

A really crucial factor in upholding the utterly absurd “health care” system in this country is that the vast majority of employed people probably have little or no sense of how much they’re actually paying for medical insurance.

Using myself as an example, I have very mid-tier medical insurance through my employer, and this is how much I’m paying for medical/dental family coverage (three dependents):

$30,846.

Sounds like kind of a lot doesn’t it? Except that 85% of that cost is in the form of an employer contribution, sent straight to the giant insurance company that the University of Colorado contracts with to provide this benefit. That money — $26,112 to be exact — is pretty much a straight pass through of what would otherwise be part of my salary, but is being sent to the good folks at Anthem instead.

What I get for this is an assurance that I won’t have to pay more than $18,200 out of pocket for medical expenses in the calendar year, meaning that there’s no way I’m going to have to pay more than let’s see here . . . $49,064 for medical care in 2025. And I’m someone with “good” insurance options!

BTW the administrative assistant who I try not to burden with too many dumbass boomer who still can’t figure out how to make a PDF with the copy thingee requests also has this wonderful insurance available to her, although in her case essentially a third of her real salary is being siphoned by an insurance company before it ever hits her paycheck. (I’m not even going to throw Medicare/Medicaid and Social Security taxes into these calculations.)

So if you think people are mad now, imagine how much angrier they would be if they could see this little shell game being pulled off in front of them, as opposed to being hidden in the fine print of their pay advices. Commenter Dr. SteveA:

Regarding the assassination of the Financial Investment Company executive, Brian Thompson: Assassination and other murder is wrong, bad. Period. This guy no doubt was loving father and husband and my condolences to his loved ones. However he was NOT — as much of the press insists on describing him as — a “Healthcare Leader”.

I’ve seen that term used in general media reports and in medical newsletters such as STAT. It’s probably scripted from United’s corporate media team press releases. And it’s being used to then falsely conflate this murder, with attacks on actual healthcare workers… shootings in emergency rooms, assaults while doing their job at vaccination sites, clinics or hospitals; the threats to Dr. Fauci and that sort of thing. The latter are all actually healthcare.

Insurance companies are at best politely referred to as part of the financial services industry and have NOTHING to with providing healthcare (just the opposite) and their CEOs are not healthcare executives, they are financial investment company executives.

By comparison, while I certainly object to many of the policies of drug companies like Pfizer or for-profit hospital chains like HCA, they are in the healthcare business. Insurance companies are not in the healthcare sector; they are in the financial investment sector.

United, Aetna, etc are literally in the business of denying healthcare; full-stop. Their business model is to take money — premiums from individuals or productive businesses providing employee benefits and taxpayer money diverted from Medicare and Medicaid — for the purposes of having a for-profit investment pool. That ius the actual business they are in. It has nothing to do with health or healthcare. Selling the product on the basis that they provide “healthcare coverage” is simply their loss leader. The money they are forced to spend on paying for healthcare is called, by them, medical loss ratio. It remains a fact that the business model is to deny and delay care, and then deny, delay and reduce payments to actual healthcare providers.

Compared to the universal coverage systems that every other rich capitalist democracy around the world have, the uninsurance and under-insurance kill tens of thousands of Americans unnecessarily every year. And unnecessarily impoverishes tens of millions more.

Americans loathe our corporate health insurance system. But most are not fully aware that it does not have to be this way. Universal, not-for-profit, publicly financed healthcare coverage is the baseline norm in every other rich capitalist democracy. It is not a lefty fringe idea that can’t happen.

It has already happened among ever other peer country. Call it single payer, expanded and improved original Medical for All, the New York Health Act, it remains what we strive for in America.

Again my condolences his loved ones. And also to the millions of loved one left behind by the hundreds of thousands killed by the existence of American health insurance companies.

The post Rage against the insurance machine appeared first on Lawyers, Guns & Money.