# I am the Watcher. I am your guide through this vast new twtiverse.
#
# Usage:
# https://watcher.sour.is/api/plain/users View list of users and latest twt date.
# https://watcher.sour.is/api/plain/twt View all twts.
# https://watcher.sour.is/api/plain/mentions?uri=:uri View all mentions for uri.
# https://watcher.sour.is/api/plain/conv/:hash View all twts for a conversation subject.
#
# Options:
# uri Filter to show a specific users twts.
# offset Start index for quey.
# limit Count of items to return (going back in time).
#
# twt range = 1 2032
# self = https://watcher.sour.is?uri=https://anthony.buc.ci/user/abucci/twtxt.txt&offset=1532
# next = https://watcher.sour.is?uri=https://anthony.buc.ci/user/abucci/twtxt.txt&offset=1632
# prev = https://watcher.sour.is?uri=https://anthony.buc.ci/user/abucci/twtxt.txt&offset=1432
@prologic Oof, a long story. One disk went bad, and I replaced the disk. As that disk was resilvering, a second disk started dropping off the bus and then reappearing. This array is 6 disks arranged in 3 mirrors (so, 2 disks for each mirror), and that meant I had *two* mirrors with only one disk each supporting them for awhile 😱 Anyway, something about that disk disappearing and reappearing threw the entire array into.....disarray (pardon the pun). I can't even explain what happened but it was really in a bad state and the resilvering just wouldn't complete.
I bought some new disks, made a new array, and used zfs send to get as many filesystems as possible from the old array to the newly-built one. One filsesystem was cranky, so I used an older snapshot of that one instead and was lucky that it worked fine. Finally, I rsynced a few directories that seemed like they might have been out of date in that old snapshot, and that worked too.
Having a tool that automatically takes snapshots on a regular interval saved my ass.
Also, I have to say that using a tool that automatically takes ZFS snapshots on some kind of schedule is a must-have. I have a filesystem that was badly corrupted but one of the snapshots survived intact. I was lucky that it was mostly a slowly-changing filesystem and the snapshot was up to date. I use sanoid for this purpose but I suppose there are many.
On the bright side, I have two shiny new pools with shiny new disks, and a bunch of old hardware that needs a new use.
Welp, after a few harrowing days, I managed to recover all my data from a failing ZFS pool. It had been serving me well, with no significant issue since 2013, about 10 years at this point. But the time has come to destroy it, which I did. Goodbye old friend. 😢
At my work email address I've started receiving spams with titles like "Resolving Wifi issues' or "troubleshooting wifi issues". It's weird to suddenly be getting so many spams with such similar titles, so I wonder if that's a new technique spammers are using to avoid filters.
Years ago, after Marc Andreesen declared that software was going to eat everything, there was a lot of talk about fields like law/legal being mostly automated by computers. I think it's interesting that, in the end, these same Silicon Valley people turned on their own--with Github Copilot and GPT purportedly being used to "write code"--long before they made any significant inroads into law.
That looks like eating your seed corn to me, and it's another signal that we're in the endgame of this madness.
@prologic Speaking as a relative newbie, I think it'd be helpful. It was confusing at first and I only figured out what was going on by "overhearing" y'all talking about it.
@prologic @marado it solves the problem of yarn becoming a surveillance system like virtually every other social network (aside from the technical reasons lol)
@prologic Old OAuth/app connections can be an attack vector. A Microsoft account connected with your LinkedIn account could also be an attack vector--the Microsoft account is a distinct thing, and could be compromised and then used to access your LinkedIn account.
Unfortunately these things tend to be very leaky and have big attack surfaces. I wouldn't jump to the conclusion that your password was what was breached unless you have evidence that's what happened. I'd change your password and set up 2fa just to be safe, but I wouldn't stop there--I'd audit anything that can access the LinkedIn account in any way, and cut those off if you can.
@prologic How do you know the password itself was leaked, as opposed to some other type of compromise?
Oof, I've been super busy with work and life, and have not been twting much. I don't expect I'll be able to come up for air much for several more months. All good stuff going on, just lots of it at once.
I have to admit I kind of miss the shizes 😢
@bender "tharbon" is autocorrect for "that on". It might also be how you say it in Esperanto, I can't be sure...
@prologic definitely. I'll do tharbon IRC just to avoid alerting the affected parties.
@lyse 🤦♂ I don't know why I didn't think of that before! Thank you for this.
Turns out the shizes all hail from the same part of the world, which I will not mention, and in fact in a couple very tidy IP address ranges that were easy to block. So perhaps that does the trick.
@lyse 🤦♂ I don't know why I didn't think of that before! Thank you for this.
Turns out the shiz's all hail from the same part of the world, which I will not mention, and in fact in a couple very tidy IP address ranges that were easy to block. So perhaps that does the trick.
@prologic That's one kind of tarpit, but there are many ways to slow down the registration process. I don't know if anyone who uses yarn/twtxt objects to email verification flows (some people don't like them).
@prologic the registration flow does not require an email address. You can just enter a username and password (anything at all for the password) and blammo, you have an account.
There needs to be some kind of tarpit). Even a minimal one would probably stop this (davi)shiz. Right now there's none! What stops someone from writing a script that mass-registers thousands of accounts per second?
@prologic, what would you advise about dealing with this person? This is like the 6th time they've registered and I've deleted them. I would like to prevent them from registering without turning off registrations altogether. Options?
@prologic no flipping way! You mean if i update to this I get a free mastodon account? Nice I don't want ti register there!
My hands smell like balloons 🎈
@cel Hello! I'm just poking around following people...
I have a big gnarly scala file that I've been refactoring for days. But for some reason, it keeps getting longer even though I'm pulling code into other files or deleting it outright?
I know some people swear by them, but I just don't like tiling window managers. They usually work very badly with my visual defect. I'm often zoomed way in on a window, and so whenever a window auto-snaps somewhere I'm not expecting, or auto-resizes, it effectively disappears from view and I have to go find it.
@prologic ha, exactly! If you could do the kind of stuff that ChatGPT does on a Raspberry Pi of a few years ago, that's be an amazing accomplishment.
As it stands, they're just throwing raw compute power and vast volumes of data at the problem. Of course interesting shit is going to pop out. You know you're getting somewhere when the capabilities increase while the power consumption stays the same or decreases.
@prologic I tend to agree with you, and it's one of the reasons why I did evolutionary computation in my PhD. My PhD advisor was big on the idea that even though we don't know what life or intelligence is well enough to make it from scratch, maybe we can set up an artificial world in which (simulated) life can "emerge" from the primordial soup, so to speak. I thought that idea was pretty compelling and I worked on it for awhile. It's why I, too, and frustrated by the term "AI" and how it's slapped onto anything these days. Some of the stuff that people call AI right now would have been called "an algorithm" or "a computer program" not so long ago 😆
@bender @prologic The "smarter people than I am are doing this" argument is a way of giving up.
AI is not in its infancy. Alan Turing wrote about what we now call AI in the 1940s/1950s, almost three quarters of a century ago. Some of what he wrote is how it still works today, in spite of all the "smart people working on it".
Back then, the first prototype transistor was being created. It was the size of your fist, more or less. Now, we can cram tens of billions of them into a square centimeter of silicon. If AI had "progressed" similarly, we'd have had walking talking robots around us long ago. Instead we have toys and marketing hype.
It's important to see it for what it is and not accept the marketing putched. We've all been inundated with technoutopian visions that never come to fruition, over and over and over again. It's time to be skeptical, and to demand better.
@prologic @adi Remember that OpenAI exists to hype itself. That's it's real business. When I was toying with my startup back in 2016, OpenAI was hyping their language models as something that would revolutionize life as we know it. They even went so far as to claim it was so revolutionary they were hesitating to release it because they were afraid of the dangers it posed. Which was horseshit, and they released GPT soon after.
ChatGPT is the same game, played with the same toys. Hype. You and your brain will be better off if you tune it out, ignore it, block it, filter it out of your feeds, etc, just like you would with any other type of marketing hype about any other product.
@prologic mmk-de @adi It's a case of all kinds of things. ChatGPT is being pushed very hard by OpenAI right now. You're seeing it in the news so much right now because they are spending endless sums of money to get it in front of your face. It's marketing. Nothing more. You should ignore it just like you ignore other advertisements. It's not even GIGO; it's just G.
As far as artificial general intelligence goes, we (humans) don't know what it'll take to make one of those. We don't even know what the word "intelligence" means.
The field of AI has been filled top to bottom, since its inception in the 1950s or so, with the strange belief that you can just break "intelligence" down into small parts--a module that can do math, a module that can play chess, a module that can emit language--and that somehow magically those parts, working together, will exhibit what we consider to be intelligent behavior. Minsky's book/theory *The Society Of Mind* articulated that view clearly in the 1980s. I, personally, think it's transparently a bunch of horseshit, but who am I 🤷 Anyway, that's the "dream" these people are chasing, I think.
There's a much longer history here of humanity trying to breathe life into inanimate matter. The notion of a golem from Jewish folklore has this quality, for instance, as does the story of Galatea, an ivory statue made by Pygmalion that comes to life.
So dig a bit deeper, and what I think you see is that these Silicon Valley types are so ignorant of the humanities that they are re-creating, in ignorance, myths and legends and stories from long ago, except with computers as the main characters.
@prologic mmk-de @adi ChatGPT is being pushed very hard by OpenAI right now. You're seeing it in the news so much right now because they are spending endless sums of money to get it in front of your face. It's marketing. Nothing more. You should ignore it just like you ignore other advertisements. It's not even GIGO; it's just G.
As far as artificial general intelligence goes, we (humans) don't know what it'll take to make one of those. We don't even know what the word "intelligence" means.
The field of AI has been filled top to bottom, since its inception in the 1950s or so, with the strange belief that you can just break "intelligence" down into small parts--a module that can do math, a module that can play chess, a module that can emit language--and that somehow magically those parts, working together, will exhibit what we consider to be intelligent behavior. Minsky's book/theory *The Society Of Mind* articulated that view clearly in the 1980s. I, personally, think it's transparently a bunch of horseshit, but who am I 🤷 Anyway, that's the "dream" these people are chasing, I think.
There's a much longer history here of humanity trying to breathe life into inanimate matter. The notion of a golem from Jewish folklore has this quality, for instance, as does the story of Galatea, an ivory statue made by Pygmalion that comes to life.
So dig a bit deeper, and what I think you see is that these Silicon Valley types are so ignorant of the humanities that they are re-creating, in ignorance, myths and legends and stories from long ago, except with computers as the main characters.
@screem I switched from screen to tmux years ago and don't regret it, though I have to say that at this point I don't remember much about screen or what motivated me to switch!
@lyse yeah I think it's dangerous too. I should probably set these variables since I compile yarnd from source.
@prologic the compiler you have installed is unaffected. The compiler that's downloaded functions as a library that your installed build system and interpreter use to compile other code.
@jlj Yeah, it's true. It's a shame, what a loss.
@support OK, not only am I getting spam account that don't post anything, I'm get REPEATS.
I've been enjoying Metafilter for the most part, but there are way too many posts about ChatGPT. I finally figured out how My Me-Fi works so that I can have a view of the site without all the "omg ChatGPT did X" posts.
One benefit you get for this pain is that you can ship scala source code that almost surely compile and run. You don't get into the mess of "oh well this only compiles for scala 2.13" and the person downloading it's like "shit I have scala 2.12 and I don't want to upgrade to compile your code." That issue leads to the world of virtual environments or containers or nix-shells or....which is messy no matter how streamlined people try to make it.
Instead, the person doesn't have to change their scala install or environment in any way, and doesn't have to set up a weird local or temporary environment. They can just compile your code as-is with the tooling they already have, and it'll just work. Part of the build process will grab a shim that lets the scala 2.12 compiler they have compile the scala 2.13 code that your source code requires.
Even with yarn this issue comes up. I had to upgrade the go install on my VPS in order to compile and run yarn, because yarn required a newer version of go than the server came with. In my case that didn't matter because I don't write go and I hadn't set up a special go environment on that server. But if I had set up such an environment, it probably would have been a pain to get yarn to compile. I'm sure there are ways, but the thing is with scala I wouldn't even have to think about it.
Whether go actually needs this though? 🤷
@prologic 0°C here. It was -23°C not so long ago.
@support Considering I already deleted you once before, @estebanplona, the answer is no.
Should we put telemetry in yarnd?
@eldersnake I don't know enough to say for sure but it wouldn't surprise me at all if that happened.
But more than that, resist letting them cross more and more lines of intrusiveness. Instrumenting build tools with telemetry that gets sent to external servers while you work crosses a line. Other build tools don't do that.
Why are you so quick to cede that ground? You really want to have *yet another* aspect of your life be a site where you can be spied on? Because you best believe that once the line is crossed and people get used to this, the data collection will increase in scope. You want to be part of letting that happen?
It's so weird to me that after all the horrifying shit they've done, people *still* default to trusting big tech companies.
Stop that! Resist everything they propose! They don't deserve your trust outright. Make them earn it. if Google wants you to trust that they'll be good stewards of the data they collect from golang, they can start by being good stewards of any other data they collect about people (which they're not). They can, oh I don't know, stop doing things like collecting location data even when you literally click a switch labelled "don't collect location data".
It's so weird to me that after all the horrifying shit they've done, people *still* default to trusting big tech companies.
Stop that! Resist everything they propose! They don't deserve your trust outright. Make them earn it. if Google wants you to trust that they'll be good stewards of the data they collect from golang, they can start by being good stewards of literally any other data they collect about people (which they're not). They can, oh I don't know, stop doing things like collecting location data even when you literally click a switch labelled "don't collect location data".
@prologic @xuu they're setting up the infrastructure and developer buy-in to stick more and more insidious garbage in there. Before you know it they'll propose "training AI" on this data or you'll find they've been doing it all along.
Always say no to more data collection. Always, no matter how good it may seem. These companies are institutionally unable to control themselves. Whether this Russ guy personally has bad intentions is irrelevant. The company and culture he's embedded in has repeatedly demonstrated they are untrustworthy. *Especially* when it comes to data like this.
@prologic @xuu they're setting uo thr infrastructure and developer buy-in to stick more and more insidious garbage in there. Before you know it they'll propose "training AI" on this data or you'll find they've been doing it all along.
Always say no to more data collection. Always, no matter how good it may seem. These companies are institutionally unable to control themselves. Whether this Russ guy personally has bad intentions is irrelevant. The company and culture he's embedded in has repeatedly demonstrated they are untrustworthy. *Especially* when it cones to data like this.
@prologic @xuu they're setting up the infrastructure and developer buy-in to stick more and more insidious garbage in there. Before you know it they'll propose "training AI" on this data or you'll find they've been doing it all along.
Always say no to more data collection. Always, no matter how good it may seem. These companies are institutionally unable to control themselves. Whether this Russ guy personally has bad intentions is irrelevant. The company and culture he's embedded in has repeatedly demonstrated they are untrustworthy. *Especially* when it cones to data like this.
*sniff sniff* I smell a spammer.
@bender
Looking forward to the year two thousand
@prologic thanks! He should definitely respond to your messages, you're on the same team! Team small!
@prologic should I have received an email? Nothing so far.
@prologic I haven't thought it through super carefully but I think so. Mainly I'm still reacting to the notion that "humor" shouldn't be restricted, which was expressed upthread.
@prologic yeah, I think it's vary difficult. I think different groups of people have different needs and tolerances. Giving them tools to shape their own space is important, but that includes blocking stuff they don't want to see.
@prologic no worries. I think it's hard to read to tone in text base media generally
@prologic I think circulating block lists, and especially blocking "dark senses of humor", is 100% necessary for a heathy environment. There is no reasonable justification for forcing people to see "humor" or whatever you call it, when they don't want to see it. Extremely bad actors say exactly this kind of stuff, and Nazis and fascists and shifty people of all stripes, for at least 100 years, have tried to hide their shiftiness by pretending it is "free speech" and "humor". It is unacceptable in 2023 to use the same rhetoric, period, end of story.
And please don't try to tone police me. I say what I say the way I say it with knowledge of what that means. If I'm not "gentle" there's a reason for that.A better thing to do is ask why I wasn't "gentle" in that particular case.
An online friend I knew years ago used the phrase "freeze peach" to refer to what "free speech absolutists" always demand online. Now I use it too. Consider this an invitation to use the phrase in your own world.
@movq I hope you get a chance to rest.
My eyesight degraded to the point that I couldn't drive a car legally anymore by 2009, so I was forced to change my life around. I'm very fortunate that I work in tech and in a part of the country where I can walk to almost everything I need, and even then it was a difficult adjustment. Nowadays, though, driving every day is a distant memory and seems quite an odd thing to have to do.
@thecanine @movq
> I also think it’s important to add, that no group, event or action, should be considered “protected from humor”.
"Humor" is punching up. "Abuse" is punching down. Yes, groups "below" you in the pecking order should be free from your "humor", because that "humor" amounts to abuse much of the time. What you're saying here is what lousy people say to defend their abuse as "just a joke; why don't people have a sense of humor anymore" etc.
> to now almost fully sanitized. It ruined the fun.
Are you joking? You are not looking at the same internet I am, then, and I'm old enough to have used USENET. This is nonsense.
> The Fediverse fractured, mostly because some percentage of users, had to find a way to silence those offensive to them, for the entirety of their instance. I really don’t want this place, taking the same path.
This is 100% horseshit and wrong. There is literally no sense in which the Fediverse "fractured." One look at connectivity graphs among instances proves that, so where is this coming from? Where are you getting this nonsense and why are you repeating it? Again, you are repeating the talking points/bullshit of bad actors. If black people (for instance) don't want instances they use to federate with instances that allow and promote racism, that is their right and they should be protected in doing what they need, not called "fractured". Just like you don't have to invite a racist into your house and listen to their garbage for the sake of avoiding "fracturing", whatever you mean by that.
It's so bizarre, but unfortunately predictable, how people who don't even know what happened in a given event feel the need to jump up and defend "humor" or "free speech" or some such empty ideal. I'd urge you to ask yourself why you felt the need to do that.
@movq thanks for saying that. I'm glad there are people who care enough to speak up.
@prologic "outside of the protocol for now". that's web3/cryptospeak
@bender cool I poked my nose into that
@bender oops try this link instead (heads up: clicking initiates a PDF download)
Would it be horrible to add PDF as a file upload type to yarn.social? Or, can it already do that and I'm just being a lunkhead about it? Every time I've tried to upload a PDF I've gotten an error.
@prologic @dfaria “Computers enable fantasies, many of them wonderful, but also those of people whose compulsion to play God overwhelms their ability to fathom the consequences of their attempt to turn their nightmares into reality.”
-- Joseph Weizenbaum, creator of ELIZA
I called it out in channel and took screenshots. Happy to discuss in a less public space with anyone--not going to share that stuff and propagate the harm.
What looks to me to be pretty nasty transphobia was posted in the yarn.social IRC channel.
I'll go on record unequivocally: transphobia is not OK ever. Transwomen are women, transmen are men, and it's not in any way funny to pretend otherwise even as a "joke". It's hurtful and harmful.
It's not a surprise that Musk is destroying everything good about twitter (including, now, closing down free API access on Feb 9 which will kill thousands of cool bots). It's part of the game plan. It's a flex, and he's a dick.
@prologic OK, I submitted two feature requests. Do with them what you will!
@prologic I think the point is to create friction. A person registering an account has to have an email address, has to be able to receive an email, and has to click a link in the email they receive. This is meant to put off (some) people who are not serious about joining. We don't need to store anything involved in that flow for longer than the time it takes to complete.
@prologic Would it make sense to file two feature requests? One for "deleting multiple users at once", and a second for "find inactive users"? They are different problems, and the 1st is more useful than the 2nd for the time being.
@prologic I'm not sure how I feel about email verification. I keep going in circles about it. I think it'd definitely slow things down, and it'd also give me access to an email address to potentially gripe at an administrator about their spammy users. But then again it does add a barrier to entry that might be undesirable (I'm not a big fan of having to have an email address to do stuff online, and I know there are people who are reluctant to share their email address, with good reason).
@prologic I definitely do not think Amazon Mechanical Turk is the solution, that's for sure.
In all seriousness, this would do it for me for now:
- Some kind of slide-to-submit captcha at registration time (which I think can be made accessible if that's a concern),
- A simple interface to delete more than one user at a time and to find inactive users
The 2nd one could be as simple as letting me enter a comma-separated list of usernames into the delete user field. It'd be even better if there were a way to search for users who've registered but never posted (obviously more filters would make it more useful, but that one alone is MVP for me). I have no trouble doing this search with a command-line tool to save the time/effort of making a nice UI for it.
Regarding the 1st one, I don't think the users registering on my pod are bots. There are too few of them, they don't post anything, and there's already a barrier to entry slowing down bots. So I'd down-prioritize that one for the time being, frankly. It's probably something to have in the works in case there ever is a flood of bots, obv.
