Really enjoyed David Gerard's amusing take on how programming with AI becomes like a gambling addiction for many.
"Large language models work the same way as a carnival psychic. Chatbots look smart by the Barnum Effect — which is where you read what’s actually a generic statement about people and you take it as being personally about you. The only intelligence there is yours."
"With ChatGPT, Sam Altman hit upon a way to use the Hook Model with a text generator. The unreliability and hallucinations themselves are the hook — the intermittent reward, to keep the user running prompts and hoping they’ll get a win this time."
"This is why you see previously normal techies start evangelising AI coding on LinkedIn or Hacker News like they saw a glimpse of God and they’ll keep paying for the chatbot tokens until they can just see a glimpse of Him again. And you have to as well. This is why they act like they joined a cult. Send ’em a copy of this post."
Generative AI runs on gambling addiction — just one more prompt, bro!
You’ll have noticed how previously normal people start acting like addicts to their favourite generative AI and shouting at you like you’re trying to take their cocaine away. Matthias D…Pivot to AI
reshared this
Gidi Kroon and Erotic Mythology 🔥 reshared this.
techcrunch.com/2025/06/05/bonf…
Bonfire's new software lets users build their own social communities, free from platform control | TechCrunch
Bonfire's mission is to build social software where people get to make the decisions, not big tech platform makers like Meta or Google.Sarah Perez (TechCrunch)
Gidi Kroon likes this.
Gidi Kroon reshared this.
StarkRG likes this.
Generative AI is like a restaurant with a sign like this:
"HI! And welcome! Please note that the restaurant does not take any responsibility if you get sick from eating here! Sometimes the food will look great and be fine. Sometimes it will look awful and you obviously won't want to eat it. Sometimes it will look great and taste great but make you horribly ill. It's your responsibility to run appropriate tests on the food before eating it and then deciding whether or not you want to eat it. Thanks for coming!"
Gidi Kroon likes this.
Gidi Kroon reshared this.
Accurate!
Come to think of it, if milk was a recent discovery, it would be hilarious trying to sell it too!
"Yeah so we found that if we keep a whole bunch of mammals breeding all the time, we can steal the food they make for their babies. It's a little filthy so we must boil or pasteurise it or it'll definitely kill you. Also, buy quickly and consume ASAP or it'll go bad and make you sick. But most of the time it'll be fine and you'll love it."
So just don't go to that restaurant!? No one is forcing you.
Also, the sign is on the supermarket. And the other supermarket, which has the same parent company. And the only competitor supermarket, because the country has a duopoly. And where did all these food trucks come from? I don't ever eat from food trucks but they're blocking my street and they all have the sign.
Another one of Roxy Furman's gorgeous wildlife films just came out on YouTube, this time she's filming in Costa Rica.
youtu.be/AM3Gp8Qs7Mc?si=_L2Jn5…
In this video Jenna Ortega discusses past characters she played and her motivation behind them. Very interesting!
She also has a great interview with Harper's Bazaar US, which is a great read with great photos as well.
Another good side to photography YouTube: I already liked following the channel of Justine with her mix of photography and life in the Ontario snow, but the latest video with a small photography project is so cute. She recreates old photos of her father when he lived in London, while her travelling takes her to the same city.
#Photography #YouTube #JustineFiles
Burn This Dyspho Out 🏳️⚧️ likes this.
Gidi Kroon likes this.
"Tips to take terrible photos", so fun and totally worth it to follow photography YouTube.
#Photography #YouTube
?udm=14 to Google Search addresses, and/or immediately clicking the "Web" instead of "All" tab in the results. Seems to help somewhat, but not sure yet what it exactly does.
Sneak peak of Wednesday season 2: spoiler, Jenna Ortega is still brilliant.
On ex-twitter: https://x.com/wednesdayaddams/status/1924814561418608855
On Instagram: instagram.com/reel/DJ4FRH0R4LA…
On YouTube: youtu.be/XWQlSo6eBNw
Back from watching Hurry Up Tomorrow with Jenna Ortega. Very impressed. Wrote this on letterboxd:
Very visual for a film about music... Unconventional also in other ways. Where most films are basically books with pictures, here the full possibilities of the medium are used to delve into the main character's mental breakdown. And as a viewer we are not given the comfort of distance, we are right in it.
I saw this film as a big fan of Jenna Ortega. I didn't really know The Weeknd, but he did very well.
Jenna went deep in this film and made it, fully understanding what this film required. Her character meets the main character when both their psyches are spiraling out of control. But is she the problem or the solution? Or something else?
This film is not about obsessive fans, no matter how much that's suggested (also in the aspect ratio?), it is about something else. And also there I think its treatment is unconventional.
#JennaOrtega #HurryUpTomorrow
My ★★★★★ review of Hurry Up Tomorrow on Letterboxd boxd.it/9IIX2h
Having a little look at other reviews... I'm not a film buff at all and in no way more likely to understand esoteric cinema, but how can so many people not get it? I mean even her character's name could be considered a spoiler.
And how come so many people mention things in their review that is in the short blurb description, but not mentioned in the actual film? I guess they make these to catch these people out...
Jenna Ortega will also be back in season two of Wednesday on 6 August and 3 September.
netflix.com/tudum/articles/wed…
Wednesday Season 2: Cast, Release Date, Filming Location, Plot, and More
‘Wednesday’ will return for a torturous Season 2. From the premiere date to plot details and new characters, here's everything you need to know about what comes next for Jenna Ortega at Nevermore Academy.Ariana Romero (Netflix Tudum)
JanKusanagi likes this.
netflix.com/tudum/articles/wed…
#JennaOrtega #HurryUpTomorrow
'Each man thought: one of the others is bound to say something soon, some protest and then I will murmur agreement, not actually say anything. I am not as stupid as that, but definitely murmur very firmly, so that the others will be in no doubt that I thoroughly disapprove, because at a time like this it behooves all decent men to nearly stand up and be almost heard. But no one said anything. The cowards, each man thought.'
- Terry Pratchett - Guards, Guards.
#RandomQuote #quotes #quote #bot
Gidi Kroon likes this.
reshared this
Gidi Kroon reshared this.
Looking out of the window and checking the street when you need to go to the toilet...
... the experience waiting for a delivery when living alone.
It's always Blink, isn't it? Yeah, mine too.
like this
paulfreeman, Curious Loth-cat and InarticulateQuilter like this.
Gidi Kroon likes this.
Gidi Kroon reshared this.
🧵 THREAD: A federal whistleblower just dropped one of the most disturbing cybersecurity disclosures I’ve ever read.
He's saying DOGE came in, data went out, and Russians started attempting logins with new valid DOGE passwords
Media's coverage wasn't detailed enough so I dug into his testimony:
Source for everyone in cybersecurity to read through and experience the craziness themselves whistlebloweraid.org/wp-conten…
If the world survives, this will be a required reading in many #cybersecurity courses.
Gidi Kroon likes this.
Gidi Kroon reshared this.
Soon the Germans will be making the same joke but using “American English”.
Votes for 16 year olds: the scientific argument
theneuroscienceofeverydaylife.…
A lot of the arguments against it basically boil down to "they do knife crime!"
But unless you believe nobody over 18 has ever done a crime, this claim falls apart when you think about it for even a second
Votes for 16 year olds: the scientific argument
The Prime Minister has confirmed plans to give the vote to 16 year olds, and the usual suspects are miffed. But they shouldn't be. Because scientifically, it makes more sense than it doesn'tDean Burnett (The Neuroscience of Everyday Life)
Gidi Kroon reshared this.
Roxy Furman beautifully filming beautiful wildlife on safari in Sri Lanka:
I've mentioned her before, as an actress and as a singer, but Corina Bradley is also a really good dancer. Here in a choreography of her own:
Resharing this because she also has an amazing voice:
Interview in Interview by Jenna Ortega with Natalie Portman. They talk about acting, life, being a child star, and Jenna describes how as a 5yo kid she used to think all TV was real and cameras were in everybody's house. I mean, they aren't?
interviewmagazine.com/film/nat…
New film poster for Hurry Up Tomorrow with Jenna Ortega. Out 16 May in the cinema.
Alexandria Ocasio-Cortez and Bernie Sanders have been on a Fight Oligarchy tour.
I think history is going to remember AOC's 'We are not powerless' speech. This is some watch:
youtu.be/1LQ8s8sKoXY?si=65GNYx…
Back home, Sanders made this 'Bad News / Good News' address:
youtu.be/70nHmlapu7w?si=YinCHQ…
#FightOligarchy #AlexandriaOcasioCortez #AOC #BernieSanders #Bernie #USPol
mattmaison likes this.
Matti Järvinen reshared this.
Waiting for someone to come back at me and mention that she doesn't actually say that slogan in this speech... Made you watch... But isn't misquoting historical speeches sort of the norm? The slogan is very much the theme of the speech however, it is in the caption, and she does say it in another version of this speech. Also, sorry, I misremembered.
What she does say however:
That we have to get rid of Big Money out of politics.
That we need a Democratic party that actually fights for people, that congress people who don't should be booted out, but that this means that voters should vote at every opportunity for those Democrats that will stand up for people.
That believing that a minimal wage should cover the cost to live, and that illness shouldn't cause bankruptcy, is common sense, not something out of a communist manifesto.
That she welcomes everybody who will fight for someone they don't know.
That the fabric of community is what can defeat fascism.
(Yeah, we're literally calling it that now)
Gidi Kroon likes this.
Gidi Kroon reshared this.
Gidi Kroon likes this.
Gidi Kroon likes this.
Gidi Kroon likes this.
I had similar problems when running a desktop environment (KDE, GNOME, etc). I’ve had almost no such issues since moving to a simple window manager instead. There’s just a lot less to go wrong. It’s linux, I’ve had plenty of ‘issues’ just not the basic desktop issues you describe.
WMs are *not* for everyone. I use Sway and LabWC (tiling and stacking respectively) and love them, it is a big step from a more typical desktop environment and requires much more up-front config though.
Linux for dummies is a Mac.
I'd just as soon not use that particular piece of shit.
i saw this toot a few days ago and i was like "you are exaggerating, the steam deck is pretty polished"
last night i was playing tomb raider and i put the deck down for ~5 minutes to check my phone. it went from no battery warning at all to powering off in that time and corrupted all my saves
like any normal consumer i fixed it by SSHing into it, finding the right proton prefix and copying over a save from some kind person on the internet in roughly the same position in the game
/lib/x86_64-linux-gnu/libc.so.6: version 'GLIBC_2.33' not found; can you repost this from another distro?
The bozos at Canonical gladly mess with your system, they want to be Redmond after all.
KDE respects your settings while the Red Hats developing Gnome hate you customizing "their master piece" and will gladly destroy/revert them every small update.
The linux desktop is garbage (like all desktop experiences), which is why we have so many distros. The trick is to find the right flavour of tamaguci for you. If you want a hibernating one, Debian on the stable update channel is your go to.
LB: I've noticed this with my Linux Mac. If I forget to run software updates for a few weeks, it slows to a crawl and is generally less than ideal.
Meanwhile, my Macs can go without updates for months because I forget, or can't be arsed, and they run perfectly fine.
@jk
And not having my specific settings on other machines means I need to know the standard way anyway (I don't want to switch in my head, muscle memory helps).
But some customisations live long, e.g. I have no Window top bars and that is working for years now...
I mean, I've been gaming on multi-monitor set-ups since Wolfenstein: Enemy Territory, but otherwise I wouldn't know what counts as esoteric hardware/software.
I did just read your other post about having a specific version of .exe files next to specific versions of .DLL files, and just not updating them so they continue working. It might be a bit fiddly to install and set-up the environment, but you can get a similar effect with Debian-based Linuxes using `dpkg --root=<dir>`.
Or AppImage, I guess. I don't know, I always saw the way Windows expects people to install software to be flawed. Running an arbitrary executable to install software only makes sense if every single executable comes from a floppy disk you bought from a shop.
Flatpak and Snap aren't all great either, they're mostly repeating the same mistakes as traditional package managers but with even more restrictions.
honestly, "the solution" to this problem in Linux has been... .tar.gz.
Literally, if you wanted _the_ Firefox from _the_ Firefox website and not what was in your package manager, you'd download and unpack an archive and then just... Run it. Ditto for old games.
Exactly what you want from Windows software: one exe, one set of libraries, no unexpected shifting dependencies breaking everything.
This might be a hot take, but you could argue that the package manager is one of the main things that define a Linux distro
This is a difference of Linux from Windows, a package manager where your distro maintainers compile and serve most software
Actually, I have no idea.
What you are describing is actually a "new" Lenovo laptop with "new" docking station and Lenovo firmware updates, especially if you try to run it say on a more "to the limit" (as in using 4 screens on a dock advertised to support 4 screens, I know I can be crazy like that). Then every firmware update is like Russian roulette. Will it boot afterwards? How many screens will work? Which screens will work? 1/3
Experience with my last 3 work laptop/dock combos suggest that about around the time Lenovo stops selling the things they usually get it done at a somewhat satisfactory level.
My current T16G2 OTOH, needs to for a reboot with the dock attached, a cold reboot of the dock (disconnect the power). Then usually the laptop boots, and all screens come up.
Anyway, Linux generally in my experience does not break, although that depends upon not forcing the distribution in doing unnatural things. 2/3
E.g. my work Ubuntu needs more of hammering it into shape, so sometimes things can break around the places where I forced it to do things my way. 🤷
On Fedora the only regular breakage that I have is my Python venv when I updated the distribution, and I used the system python, and suddenly the system python was a different version. Stopped the practice, install my own python with pyenv, my venv stay whole even after distribution upgrades. 3/3
#Linux package managers update packages they have installed from package repositories -- servers -- whose identities and location are maintained in files on each installation.
When users add third-party packages they run the risk that those packages will not be maintained or updated to ensure they remain compatible with changes introduced by system updates. It's the responsibility of 3rd-party packages to make those changes to keep up, not the distribution's.
Well, years ago we had an inside joke in our group of friends. Whatever the question or problem was, the answer or solution was "go Debian".
In your case maybe it's not the best answer you can get, but certainly it might be a not wrong answer. It mostly works for me, and I'm running unstable for 25-ish years.
PS. Saying that, I've the update overdue. Ask me in some time what has been broken 😅
by switching from arch to nixos, ive replaced biweekly weirdness drops with a constant reproducible weirdness flux because it doesn’t follow the FHS 👍
i do think linux ought to just move to package formats like flatpak or nix at this point. it’s not funny when you want to update discord, but discord depends on libshit.so.6.9 and you only have libshit.so.4.2.0 so you run apt update && apt upgrade and now for some reason all of your icons are different and half of them are missing
I don't know. I had everything set up the way that worked great for me and then Canonical put in the Unity desktop and Gnome2 was gone and it's taken me 17 years to get it back to almost like I wanted with AVLinux based on MX linux without using systemd and Wayland. Those last 2 have been the bane of my GNU/Linux usage since they showed up. All my issues the past few years have been because of them.
The only thing not quite there yet? You guessed it--icons I prefer to see on the desktop.
what distro are you using? I must admit I've had some issues with linux from time to time myself but most of the time it was kinda self inflicted.
My main laptop runs manjaro linux (an archlinux derivate) and I have to admit for anything archlinux like you need some technical understanding or else you can destroy your system easily. But running a Debian or something and in general running LTS versions of whatever distro...
:D possibly xD it's been some time since I used mint (about 15 years?) and I didn't like it that much back than but don't even remember why. Kinda strange you're having so much trouble on debian based distros, debian is supposed to be relatively stable.
One tip: where possible use the distros own configuration GUIs and so on. Several times I broke things was because I manually edited some config files without knowing enough or ran some command I didn't fully understand.
Oh, definitely.
Don't run package updates every 2 weeks.
I run package updates when I'm forced to, at gunpoint, by Linus Torvalds personally.
I guess I appreciate that he cares, he had to fly all that way after all.
(Seriously though, the package distribution model is to create a set of interrelated dependencies that are all supposed to work with each other in arbitrary configurations, and I'm fairly certain that's not actually a tractable problem given the sheer scale of the package repositories these days).
@nazokiyoubinbou FreeBSD with ZFS snapshots is nice for this reason. Something goes weird, you just revert to a snapshot from after your last successful boot. A number of Linux distros also offer root-on-ZFS, so can get the same results.
btrfs on Linux has similar functionality, but I personally find it annoying to actually use.
@bob_zim I mean I could use Timeshift for that, but I'm talking about actually immutable. No updates. Like find an actually 100% working setup and set it to stay.
Basically like a live CD or that sort of thing, but with a writable home directory and such.
This exists, it's just super hard to do at the level I'm talking about. It can be done, it's just not exactly a quick and easy process.
On a Debian-based system you can get a long way with a combination of the policy-rc.d trick - jpetazzo.github.io/2013/10/06/… - and by modifying apt.conf to include
Dpkg::Options {
"--force-confdef";
"--force-confold";
}
These two options combined amount to "never turn on newly-installed services by default" and "never change the configuration files of existing packages on upgrade". It's imperfect but AFAICT it's the safest way to keep running updates while minimizing user-facing change.
@nazokiyoubinbou ZFS and btrfs do a *lot* of things and have a lot of advanced features. In Linux terms, they can cover md(4), lvm(8), a normal filesystem, plus the advanced features like the caching layers, datasets/subvolumes, inline compression, checksums, and snapshots.
I *suspect* some of my distaste for btrfs is just from knowing the ZFS way of doing things like enabling compression or taking a snapshot first, and seeing the btrfs way second.
As an example, though, to take a snapshot of my user data volume, I would use this for ZFS on FreeBSD:
zfs snapshot pool/home@$(<date command>)
And for btrfs I would use:
btrfs subvolume snapshot -r /home /.snapshots/home-$(<date command>)
The actual date command is much longer on Linux, so I’ve replaced it in both with a token to avoid artificially inflating the btrfs command. It’s similar, but where ZFS only requires specifying the dataset name and snapshot identifier (after the at), btrfs requires specifying the new place to mount the snapshot.
The -r option in the btrfs command makes the snapshot read-only. You can send read-only snapshots to remote systems for backup purposes, but restoring from them leads to a read-only filesystem. You have to make a read/write snapshot from the read-only snapshot, and restore from that.
In contrast, you can send ZFS snapshots to remote systems *and* restore from them directly, and you have to take explicit action to make a writable clone (which still leaves the original snapshot read-only).
(I have not read the replies to this post before replying.)
For me, this was installing Ubuntu LTS and walking away.
Unfortunately, I later found out this only works on server, as Ubuntu LTS for *desktop* is frickin busted.
I am now hoping I can pull a similar trick with Debian LTS.
@Janeishly My Debian system, which is _directly_connected_ to the Internet (VPS) hasn't been updated in years. I limit what it runs, spent the time to build multiple layers of security, then built custom monitoring for it.
I pay attention to the news, and check to see if any new bugs actually affect my installation. (They usually don't.)
It's stable and reliable. Don't update. Or at least limit that to "security only".
It may be personal but I see it with a much different perspective.
I'm *immensely grateful* that it's possible to have a machine with so many millions of SLOC running pretty much exactly how I want, respecting my freedom, with me contributing so little in return.
I'm also grateful that I can get to learn and know what's inside, and via the stream of updates and breakages, essentially be getting free continuous learning and also staying updated.
I'm purposely running #gentoo on my main machine, playing that Tamagotchi game very much.
When I want to "get things done without surprises", I use a "stable" "supply chain", trading off my own risk against reduced functionality, increased cost.
The good thing with "Linux" is that in no case you lose your freedom.
infinite love ⴳ reshared this.
Welcome to Bluefin | Bluefin
A custom image of Fedora Silverblue, offering a familiar(ish) Ubuntu-style desktop.docs.projectbluefin.io
Hi,
Here's my experience, YMMV.
I'm also not much for doing weird things on the desktop I use everyday. I generally just want things to work.
While breakages used to be more of an issue, largely I don't have many problems when updating anymore. I have used Mint as my distro and cinnamon as my window manager for about the last 15 years and it generally just works. Generally I pair it with an older thinkpad business laptop.
Honestly, the way to have that experience is to run Debian Stable and never enable the backports repo. You'll get the same, never changing set of packages for several years.
I'm a Professional Sysadmin:tm: and I look after a bunch of RHEL desktop machines. The main issue I have with them is big corporations who should know better fucking up their software repos constantly. Intel and Nvidia, I am looking at you.
It was seen as more efficient and not wasting disk space. In hindsight, screw my disk space.
AppImages and Flatpak kind of go in the opposite direction and include absolutely everything, so that problem *should* be solved going forward (lol)
But yeah, with regards to blame, there's no John Linux or LinuxCorp to direct blame at. There's also about a billion implementations. Someone on Debian is going to have a completely different experience to an Arch user. (Source: I use both)
Android and Valve/SteamDeck manage it well enough.
I can answer definitively yes "Linux" can be configured to practically never change and never break.
But a distribution built by arbitrarily cobbling together thousands of projects moving at their own pace and keeping them up to date, sorry the answer will be likely no.
Your choice of cobblers matter insofar as the frequency of allowed breakage.
Becoming your own cobbler can help but requires much study and more frequent frustration.
"But surely they'll realise they're crashing the economy?"
"We have to raise awareness that these actions will burn the planet!"
"They'll stop once they see the impact on human lives, especially their own people."
We have to remember that the destruction is the point. They do want to reset everything, believing they are best placed to be the only survivors of the human race and the ones to start a new era. That's also why they amass so much more wealth than they need now.
Just back from watching Jenna Ortega in Death of a Unicorn. I really liked it! There's also something about the combination of her acting with detailed realism and the story being like 'well, unicorns', and that is a combination that I weirdly really liked. I guess it adds to the multiple things her character is trapped in.
My review of Death of a Unicorn on Letterboxd boxd.it/9jRbqN
Jenna Ortega's film Death of a Unicorn is out tomorrow in The Netherlands.
Hurry Up Tomorrow...
...will be released in the US next month.
"LLM did something bad, then I asked it to clarify/explain itself" is not critical analysis but just an illustration of magic thinking.
Those systems generate tokens. That is all. They don't "know" or "understand" or can "explain" anything. There is no cognitive system at work that could respond meaningfully.
That's the same dumb shit as what was found in Apple Intelligence's system prompt: "Do not hallucinate" does nothing. All the tokens you give it as input just change the part of the word space that was stored in the network. "Explain your work" just leads the network to lean towards training data that has those kinds of phrases in it (like tests and solutions). It points the system at a different part but the system does not understand the command. It can't.
Gidi Kroon likes this.
reshared this
FediThing :progress_pride:, Gidi Kroon and Trolli Schmittlauch 🦥 reshared this.
Charles Ash likes this.
Paul H reshared this.
@Gidi Kroon I hear you... News headlines of late are like political satire, only 100% true and terrifying.
It's really quite depressing.
Gidi Kroon likes this.
Gidi Kroon reshared this.
12 April, the second series of the re-re-boot-boot of Doctor Who on the BBC. Here's the second trailer:
youtu.be/_xuo_FvPI18?si=1ZR00i…
ƧƿѦςɛ♏ѦਹѤʞ reshared this.
Gidi Kroon likes this.
Jen, MPH :verified:
in reply to BrianKrebs • • •bookmarking this one! Love the cocaine analogy!
“Coincidentally, Natalie Ponte on LinkedIn posted today: try replacing “ai” with “cocaine” in all the posts you read about it. it’s pretty funny
Let’s try it!
“My cocaine skeptic friends are all nuts, they’ll be left behind.”
“Cocaine isn’t going to replace you. Someone using cocaine is going to replace you.” Checks out.
Cocaine doesn’t make you a business genius — it just makes you think you’re a business genius. Same for AI.”