I haven't read it, but I find it hard to believe that good social advice would go out of date. Our brains haven't changed on a fundamental level since then.
If anything, modern social interaction has diverged from what is good for us and what we really want.
I've seen this a lot with travel youtubers where they want to show genuine interaction with the population rather than generic tourist stuff. So they set arbitrary restrictions like not using maps, or avoiding PT and trying to hitchhike to destinations. Pretty much everyone everywhere is willing to help how they can and have a conversation in the meantime.
Having some obvious goal like "I'm a tourist and I'm lost" immediately cuts past the "Is this person a scammer/beggar?" you normally think when a stranger walks up to you.
There's a video on YT that basically proposes the idea that Ewan McGregor's & Charlie Boorman's "Long Way Round" series has been the last "true" travel documentary made for TV. I thought that was a bit silly initially, but after watching it(the YT video) I can't shake the feeling that I agree with the premise.
Basically - in "Long Way Round" Ewan and Charlie take two motorcycles and they ride them from the UK east, going through Russia and finishing in the States. But even though they have support crew of two cars following them with extra equipment, it doesn't change the fact that back when it was filmed, they didn't have internet connected phones with them. The maps of Syberia were innacurate, they didn't speak any Russian, they were driving through really difficult terrain and you could "feel" that they were far from civilization - contact with family was infrequent,they relied on kindness of strangers a lot because there was no other way to continue. There was a lot of communicating with hand signs and pointing.
Compare that to their later "Long way up" where they rode up through South America - at that point they both had mobile phones with maps and instant translators - they were never lost, and even in the middle of absolute nowhere they would walk up to a local and just use their phone to translate. They had video calls with their families practically every night. Even though they were in some places no less remote than Syberian roads, it didn't feel remote in the same way.
And yes, I know there are YT travellers who try to artificially follow the same kind of philosophy - no translators, no maps, just going in raw so to speak. But the world has changed. Even if they don't use phones, the locals do even in the middle of the amazonian jungle. And it is amazing how connected humanity is, but there is certainly an era that has ended for all of us.
I really recommend this video, whether you have seen Charlie's and Ewan's adventures or not:
Long Way Round felt a lot more "natural" than the follow-on movies. They were all good, but Long Way Round was by far the best. IMO, followed by Long Way Home, which had no real pretense about being remote or "adventurous" - just two buddies doing what is likely their last moto tour together.
For some basic stuff like vim it works fine. But for almost everything else I'd rather a regular CLI tool or a web interface. I suspect a lot of the popularity comes from people who want to feel like a hacker using 10 terminal windows, but actually want a GUI like experience.
For me, TUIs compensate for the fact that I can't get good remote GUI rendering on Linux. Yes, X11 tunneling exists, but the experience has always been abysmal for me for anything not hosted on a machine that sits on the same LAN as the client. For Wayland I don't even know if such a thing is possible since I don't think the architecture supports it.
But the terminal is just fundamentally the wrong basic abstraction on which to build a structured GUI, it just happens to require few enough bits to be sent over the wire that it actually works reasonably well over SSH as opposed to pushing graphics.
You can use Sunshine streaming on Wayland just fine. I have a headless sever with Sunshine and a Wayland desktop and can stream it remotely with great quality and minimal latency.
That said for most tasks I would still rather SSH in and use CLIs (not TUIs).
I have never tried it until now, and I hadn't looked into it. But I just tried `waypipe ssh` to a remote server I have for doing asynchronous Claude work in VMs, and it actually works pretty great! Maybe I'll switch to that for my emacs/magit setup, the lack of clipboard integration when running emacs in a terminal over ssh is enough of an argument for me.
Edit: yikes, pressing M-w caused emacs-pgtk to crash with a Wayland protocol error, so it isn't trivial and requires some configuring I guess.
Edit 2: Apparently I have to install wl-clipboard and write a bunch of emacs lisp to work around this. I don't think I have the patience for that, and I fear that such problems will be even harder to solve for applications which are not as flexible and programmable as emacs. So far I'll conclude that remote Wayland is not ready and stick to TUI.
Edit 3: No, the problem is probably mismatched waypipe versions on client and server. Still not fun.
> Yes, X11 tunneling exists, but the experience has always been abysmal for me for anything not hosted on a machine that sits on the same LAN as the client.
I have used X11 tunnelling to machines on the other side of Europe and it was OK. I did prefer ssh for responsiveness. What happened to NX? What about other remote desktops?
Did you just invent a non-existent core Wayland developer and attack them for things they didn't say? I am all up for shitting on modern software stacks and praise the Unix philosophy, but I am really turned off by posts trying to make this a question about gender identity and generation, as if that has anything to do with it. Old cis-dudes come up with bad software architectures too.
Sorry, but it is kind of a generational thing. If you grew up in an environment where one set of assumptions always held, you are more likely to think that these assumptions are universals and just not even bother with the possibility that they might not be the case. This goes double if you're young; even bright young devs often have not had the breadth or depth of experience to consider that that crusty old protocol they're trying to get rid of due to "legacy cruft" is the way it is for a reason, one which is still reflected in the use cases of real users working today. This has direct bearing on the Wayland situation: one of the reasons why Wayland is having trouble getting traction against X11 is because the Wayland devs have lost sight of the fact that Wayland is designed to solve one set of problems and X11 another.
And real Wayland developers have made the exact argument I attributed to my fictional one, just with fewer anime emoticons (and "watch this talk by Daniel Stone" as a final mic drop).
Obviously people want GUIs. That's why TUIs should be compared to GUIs, not to CLIs. TUIs are nice since you get a lot of the benefits of a GUI, without having to leave the context of the terminal.
I feel like the better solution here (than trying to shoehorn a GUI into an interface meant for text) is to make terminal windows graphically-aware, like how things work in Plan 9.
I already do use tiling window managers and they don't really accomplish the “if you launch a graphical app in a terminal window it takes over that terminal window” flow. Closest I've found is Niri's support for tabbed windows, but even that's just sticking the graphical app window on top of the terminal window instead of the terminal window itself becoming the app window.
Depends, I'm building a markdown down editor that just previews in the web ui. However, I can use the web ui to do tasks like uploading files, view git commits etc. Different interfaces for different purposes. The CLI gives me focused mode, and then visual stuff that steal my time goes in the web
I dunno, pre-LLM TUI's at least tended to be okay, and keyboard navigation was a first class citizen. Besides, if you were using a TUI instead of a GUI then you basically always ended up saving memory/battery life, and TUI programs are generally more portable than trying to run some ancient GUI program.
I typically prefer CLI myself but having a TUI to manage torrents for instance was much more ergonomic.
A lot of the complaints in this thread seem like they're aimed more at recent vibecoded UIs than the concept of a TUI.
Like, okay, they are a big step back with accessibility, but they're flickering garbage because they were vibecoded in a weekend and the TS or Python library they're built on was similarly forced upon this world.
For almost every tui, a webui works better imo. Most torrent clients offer a web management ui and it's always going to be easier and more feature filled using a platform that was actually designed for it rather than hacking a gui in to the terminal.
Well, "better" is relative. Web UIs definitely have the advantage you are mentioning. However, they are also much slower to use, so it is a tradeoff.
I am blind and do rely on accessiblity. But I would choose a TUI over a Web UI every day simply because the web is slow as hell. Feels like running away from a "monster" in a dream. You have rich details, but you are being drowned in unnecessary verbosity and an inherently slow stack of tools working together.
The web might be accessible, but it is barelyuseable in practice.
Its more or less the same with every screen reader and browser I know. If there was a simple fix for this inherent issue, I wouldn't have to talk about it. You might get a bit of stuff out of the way if you configure less verbosity of the screen reader, but it doesn't really help with slow interactions. Problems include:
* Virtual buffer: In most systems, the DOM is rendered to a "virtual buffer" and the screen reader lets you navigate that, because cursor (caret) support was on the map for accessibility, but was never really sufficiently implemented on the browser side. So screen readers had to solve it in a separate step.
* Using the keyboard to do screen reading and navigation on a site conflicts with keyboard shortcuts from that site. So most screen reader have two modes: You switch between entering text in a field and navigating/reading the site. Its basically like vi insert mode, but less simple.
Those are the major issues from my POV. The rest of the slowness likely comes from having to go through the accessibility API of your OS.
Vim is special because 99% of what we do is editing text, and it is the text editor—the importance of that task overcomes the poor discoverability of a TUI. Most other programs should be CLI, so they can fit in the conventional command line toolbox.
Unless I actually know the author I assume everything here is vibeslop and full of mistakes.
Maybe I need to switch to some news publication that actually does real research and writing still. Because public forums like this have been completely destroyed by LLMs.
Long term all of the major LLM platforms will have invisible ads, influences, and propaganda woven into the content. The temptation will be irresistible for these companies.
The difference imo is removing the information from the source. Previously you'd use the source of the information to gauge how much you trust it. If it's a reddit post or a no name website you'd likely be skeptical if it doesn't seem backed up by better sources. But now the info is coming from an LLM that you generally trust to be knowledgeable. And the language it uses backs up this feeling.
The OP post is highlighting how incredibly easy it is for a very small amount of information on the web to completely dictate the output of the LLM in to saying whatever you want.
> But now the info is coming from an LLM that you generally trust
But it's not from the LLM, the LLM clearly cites the wikipedia article as its source. This is just performing an internet search with extra steps, and ending up with misinformation because somebody vandalized wikipedia.
Is there a library of good tools for LLMs to call? I have to imagine the bot-detection avoidance mechanisms are a major engineering effort and not likely to work out of the box with a simple harness and random local LLM.
Kagi also has an API. People who hate ads are probably the same folk that should be paying for Kagi. That's the sane alternative world where companies respect their users.
Oh, you got me so excited. I've had a Kagi sub for 3 years, but their API is still in closed beta. I guess I could (and should reach out and ask for access).
firecrawl: "if you post content or intellectual property within the Services or give us Feedback about the Services, you hereby grant to us a worldwide, irrevocable, non-exclusive, royalty-free license to use, reproduce, modify, publish, translate and distribute any content that you submit in any form [...] You also grant to us the right to sub-license these rights"
exa: "Query Data is used to improve our products and technology, including by training and fine-tuning models that power our Services"
perplexity: "Perplexity may retain, copy, distribute and otherwise use Search Data for its lawful business purposes, including the improvement and development of products and services."
linkup: "Client grants Linkup a worldwide right to use, reproduce and modify the Client Data, including prompts, for the purposes of providing, maintaining, developing, training"
tavily: "we may use certain portions of your query data to improve our responses to future queries"..."We may share your query data with third-party search index providers (e.g., Google)"
That's not how it works. Whether local or hosted, every modern model has a cutoff date for its training data, and can be leveraged by agents / harnesses / tools to fetch context from the internet or wherever.
Reddit has always been fake, but it used to be a real person performing creative writing pretending to be a true story. Now it's spammed out slop at scale.
You need massively expensive hardware to run them, and they aren't as good. It's pretty clear the base price of AI tools is way higher than we are being charged right now.
I wouldn't call my $2k Strix Halo computer "massively expensive", and it runs e.g. Qwen 3.6 27b brilliantly, with tons of memory to spare and is a full x86 powerhouse pulling 120w at absolute max.
IMO the programming world is far too myopic about / insistent on using laptops, especially macbooks. Just because a crappy deal exists doesn't mean everyone is forced to take it. Local AI is a high performance computing problem and laptops are fundamentally a crappy form factor for it; buy an efficient desktop computer and be surprised at what's possible even with today's crazy prices.
If anything, modern social interaction has diverged from what is good for us and what we really want.
reply