Forest @forestjohnson@pixie.town

@notplants It's based on XMPP, so do with that information what you will.

@notplants This is something that I also deeply want but I kind of suspect that the best you might be able to get is going to look something like "burner phone as a service" and even then it might be hard to keep that going without those SIM cards being deactivated.

@notplants I'm pretty sure this doesn't / cant exist. Even Twilio is a freaking nightmare to try to use these days, buried in compliance paperwork and kyc

And a lot of places will even block or ban phone numbers that come from "sketchy" providers.

Based on your comment about wanting it to be easy and actually work, for example, work in the browser, I suspect you may have already run into jmp.chat 😥

@reese 192.168 or 10.0.3

@reese Poor tate suffered from a GPS blackout and got slightly lost during this time ! The Sun had something to say on the RF spectrum.....

@nolan @simon

Even if we can get past the prompt injection issue, which as of today, it doesn't really seem to be possible...

A lot of things in this talk are wildly optimistic. The talk does not acknowledge that technology is designed and built only for the benefit of the large and wealthy organizations building it, all of the benefit for users is incidental.

I think the entire talk kind of hinges on this one central point which she made that "People wanting information to be free" are to blame for a lot of the issues with the web today (lack of micropayments). But wait. I thought it was the other way around where "the information wants to be free."

It doesn't really matter what I want. I get reality every time. And the reality is that there's no cost to copy information on a computer. As a matter of fact, there's no way to access information on a computer without copying it.

So I think this idea that we're going to replace all the tarpit crawler blockers with micropayments is incredibly optimistic.

Edit: oops hit the post button by accident before I was done typing.

Any system that can tell the difference between a web crawler that's going to copy your content and repeat it back without paying you, And a personal agent that's going to read your content once and present it to a human, It's going to fall to the same power dynamics of platforms that she rails against. (It sounded like she was talking about different prices for human users vs crawlers)

I like my proof of work bot deterrent because it seems like the price is acceptable for individuals but not acceptable for crawlers, even though it's always the same price, because I have no way to differentiate.

And this whole idea that you can own your data by having a local agent running on your phone... Don't make me laugh. Who on this planet can really say that they own their phone? That it does their bidding and no one else's? I don't really think of my phone as something that I "own", that I can store my sensitive data on, even though I use third-party Android ROMs to try to de-Google it.

I think she's missing the massive erosion of trust in technology and changing relationship that young people have with the internet.

Maybe the reason web traffic is declining is because young people just don't like the internet, Because it's become so toxic in ways that are practically hegemonic and unavoidable.

@handle

> how much value is there in password managers if you give no advice about passwords? selfhosted vaultwarden behind 7 proxies doesn't help if all the passwords stored in it are "password"

Mostly the value is that it forces the user to collect and inventory thier passwords, and to back them up. IMO the primary most significant risk involved in using passwords is that you lose the password and you get locked out of your account forever, I believe this accounts for over half of all password related problems, and often has the most severe consequences.

Also, I don't know if vaultwarden will warn you about weak or already leaked passwords, but I know things like 1password and bitwarden will.

@handle

> what is the difference between an encryption key seed and a password to you?

Well, they have completely different security models. If you're going to try to guess a password, you always have to send your guesses one at a time over the network to someone else's computer that you don't control.

With encryption key seeds, you can have an entire rack of GPUs all guessing at once without bothering anyone, and then you can buy another rack.

@decentral1se @notplants Oh yeah, I remember now. I remember being slightly frustrated by how much screen real estate it took up.

I don't think it needs the borders like that. Maybe there's a way to turn it off? don't know.

But in general, I like the idea of a Tmuxish program that has the nano style keys labeled at the bottom, and shows the mode.

@decentral1se @notplants I used it one time and I was surprised by how long it took to start up. I don't think I have any complaints about the software besides that, But also, I never really used it again, idk what that means

@notplants I remember having a lot of technical issues with TUI's like 10 to 15 years ago. They would never work correctly on the shell that I had access to. For example, a some sort of weird Unix shell for windows that I may have been using at the time. Or maybe the TUI was part of a installer for Linux that I had to click through, but it wasn't working w/ the virtualbox display or something. Or maybe I'm trying to set up a automated build of a piece of software. And I have to somehow configure my automation to click through the TUI, but it's really brittle and hard to manage.

These days I feel like TUIs have less technical issues where it's completely broken... The standards for terminal emulators have increased. But they still have all the same usability issues typically. They seem to thrive on hidden modes and missing affordances. Even just basic foundational things, like pagers being inserted by default on certain Linux commands, I find it to be very detrimental and tech-elitist.

The first time a user gets put into a pager, that will end their shell session completely because they will not know how to get out of it.

So every single cheat sheet for basic things like looking at journald logs, listing git commits, docker build etc etc also has to add in special syntax to prevent the pager from coming up or it has to explain what the fuck the pager is and how to get rid of it. Except none of them actually do that.

So, I think if we want TUIs to be a thing, we should probably figure out how to give them affordances and have certain standards they must meet (no hidden modes, must contain instructions showing how to exit and how to enter help if there's more than like 3-5 commands) for their inclusion into package repositories.