pocket size ssh terminal
⁃ six-inch display
⁃ one week of battery life
⁃ cellular
⁃ wifi
⁃ usb-c display port for keyboard and external display
⁃ operating system has ssh only
⁃ local storage just to keep ssh keys
⁃ minimum ram and cpu just enough for zero-delay ssh
⁃ price below 500 eur
is it achievable?
@hi ADV cardputer? edit: under £50 with wifi card, BT support, LoRa support available, USB-C interface although battery life isn't very food. Display is barely an inch.
https://github.com/m5stack/M5Cardputer/blob/master/examples/Advanced/SSHClient/SSHClient.ino
@hi this already existed in 2000s, minus the usb-c. It was called the HTC Merge. Loved the slide out keyboard. Nothing like it really exists anymore to my knowledge.
@hi Hmm? OpenSourceSDRLab UConsole with ClockworkPi https://www.aliexpress.us/item/3256811687027709.html?channel=twinner
Should be foldable with built-in keyboard.
And doable with 200 eur smartphones, so except for the battery life it should be achievable for maybe 100 bucks.
@hi - I think so. The Gemini PDA has a 6 inch display. Runs linux which you can configure to be headless and only have SSH on it. It's wifi and cellular. Not sure on battery life, but its robust. Price is right. USB-C.
It does have its own keyboard on it as well.
@hi Saw this thing just this morning and it's on my potential pocket computer list especially due to the LoRa support. Obviously I don't care about the "AI" accelerator but a stripped down RISC-V Linux port with just the console and busybox would be rad.
I am a bit confused about one thing though; it seems you have to have the keyboard attachment to get the LoRa/mesh radio?
@hi Clockwork's PiCalc gets close. Cellular is missing but I've run SSH firmware on it. It's a microcontroller so a bit limited but it worked. Battery life was good but might need some extra cells ... easy enough to expand or swap since you only need one battery in at a time so you can hotswap.
Alas, the OS did not have SSH-only; it was some java morass. Basically Android was the HipTop 2.0 (and got worse and worse, particularly after Alphabet Inc./Google ditched physical keyboards).
@hi I think the battery life will be mostly limited by supporting WiFi and Cellular, also these will, like USB-C, consume a bit of CPU time in most implementations, so a too small CPU won't work.
Otherwise, no problem, an even lower price should be possible. Not sure if such a minimalist OS exists (Linux might be too good) for a SoC, would need to research it. But you could always take small SoC and all protocols as application level software.
@hi Isn't openBSD at least 32 bit as well? I wondered if a 8 or 16 bit CPU could also do it, cryptography might be a PITA on 8 bit though.
@hi https://pine64.org/documentation/Phone_Accessories/Keyboard/ pinephone+keyboard , but even with the extra battery in the keyboard, no where near a week of battery.
Nura (formerly postmarketOS) can be installed with no window manager.
I wanted the pinephone to be this so much, but either I got a bad device or it was a big miss.
@hi
I've had a similar idea!
Mine wants to be a head-mounted display with a few lines of text, paired with a chord keyboard like the Loop, or maybe just one of those teensy media keyboards, and just enough processor to shell into a grownup computer.
@hi Hmm. Something like a $69 SBC like the zimablade would do it, but has no screen. It has a HDMI output, and is small and light enough you could glue its case to the back of the screen to make an integrated unit??? I see 10" HDMI screens for $70. Now, the"OS"... 🤔
@hi That seems completely doable to me if it uses an e-ink screen. Battery life depends most on how you plan to use it, though.
@hi The old GPD Micro comes pretty close and adds a real serial port but misses on battery life.
It would not be hard to build something like that with an ESP32 to get the battery life up excluding the cellular bit.
There are a few vimprovements that I miss when using vi(1)
• text-objects like
ci"
or
da(
are incredibly powerful, especially when combined with the dot-operator to repeat the change
• the far more powerful regular-expressions (things like positive/negative look-ahead/-behind)
• the ability to evaluate expressions in replacements like :%s/\d\+/\=submatch(1)/g lets me do some amazingly-powerful edits on a regular basis
• insert-mode completion (being able to do things like type control+p in Insert mode to search backwards for a word matching what I've started to type)
For more info, the corresponding vim help targets are
:help text-objects
:help pattern
:help sub-replace-\=
:help ins-completion
(I regularly use both vi(1) and vim depending on what the editing calls for)
lil computer with transflective lcd (40x4), and solar? wdyt?
@mntmn I love that the keyboard display is comparable to the external one. The variety of hardware we have these days is amazing, and yes, I love good constraints for hardware.
@cararemixed haha that's true. but it has more horizontal res, officialy 40 characters (the oled has like 21 @ 6 pixels (edited))
@mntmn love solar and transflective lcd, i think 40x4 would feel too limiting though
160x160 like the palm m500 or 240x160 like the game boy advance would be still feel a bit cramped, but full of possibility
how does e-ink compare in price and power to that 40x4 display?
@mntmn I would love nothing more than a device with a transflective LCD I can use in the sun. I know Pixel Qi are only new old stock, but at least that specific LVDS connector saw ubiquitous use on displays that are likely still available today
@mntmn I'm so getting the standalone keyboard when I can afford it. I have a fancy Moonlander but the Pocket keyboard is the most comfortable to use
@mntmn lol but also: color epaper? https://www.waveshare.com/product/displays/e-paper/4.26inch-e-paper-hat-plus-g.htm
@mntmn Did you not have a banana handy? :)
I love this thing, could there be a slightly wider keyboard layout?
@mntmn I'd take one. Currently I'm using a https://byok.io, but this device has the clear issue that the battery will be dead in a few years and I'm not even sure if I can open the casing.
A writing device with open hard- and software would be great; solar powering would be magical.
Want to convert temperature from Celsius⟷Fahrenheit and you have units(1) on your machine?
$ units
751 units, 62 prefixes
You have: 17.6 degC
You want: degF
63.68
You have: 74 degF
You want: degC
23.333333
No web temperature-conversion website or online connection required.
Works for other units too if they're defined in your units file: /usr/share/misc/{units.lib,definitions.units} depending on my machine here
(post instigated by @stefano's recent post about it being 17.6°C in the house)
yeah, I know there are other add-on packages to do unit conversions, but units(1) is available in the base install for FreeBSD & OpenBSD (might be in NetBSD too)
Does it know less-common units such as football-fields per dog-year?
https://dianne.skoll.ca/writings/totally-true-news/2026-05-21-metric.php
I did a bare-metal install of #NetBSD today on my 2017 HP Pavilion (Intel/Nvidia) laptop using a USB-connected 500 GB HDD.
It took a couple passes before I figured out that I had to manually remove the MBR on the target drive for the installation to boot.
Then it was easy because I've done a bunch of practice installs.
On this laptop, everything works:
WiFi
X11
Sound (and volume buttons)
Brightness buttons
YouTube video plays great
I'm sticking with CTWM for now. Nice to see that most GUI apps are automatically added to the menu (and show after a CTWM restart). I added some of the stragglers to a local .ctwmrc.
Browsers Firefox and Chromium did appear in the menu automatically. So did htop and Geany. Gvim and ROX did not.
So far, everything is very responsive. This COULD be a daily driver. More testing needed.
Next thing: Power management. #OpenBSD is great at this. Hope #NetBSD will bring it.
I guess powerd is already running .... Fans are quiet, so it must be working
#NetBSD WiFi networking issue:
Same hardware (2017 HP Pavilion with Intel/Nvidia and Intel Dual Band Wireless AC 7265), same location relative to router:
Debian: 526 Mb down
NetBSD: 15.5 Mb down
Debian is 3,400% faster
I need to test with the Edimax N150 ... but this is disappointing.
@passthejoe NetBSD wiki (or was it website?) has notes on optimising performance. I got at least 3x network speed up from tweaking sysctls
Update:
I did an #OpenBSD install, and after getting the firmware for iwm, I was able to get 58.7 Mb down.
Without any tuning, OpenBSD is 380% faster than NetBSD. And Debian is 900% faster than OpenBSD.
@passthejoe Sometimes it also helps to specify "mode 11n" to make sure it's not limited to 802.11g speeds.
@brynet Great tip. I will try
@passthejoe .. or "mode 11ac" if your AP supports it, ofc.
AC 7265 should be capable of far more than ~60Mb/s on OpenBSD iwm.
@brynet I'm getting 5x speed with `mode 11n`
quite a difference. Thanks.
@passthejoe 11ac?
@brynet I'll have to try that one. I'm sure my router can do it.
@brynet Where the signal is stronger, mode 11n is 35% faster than without it
mode ac works but isn't any faster than not specifying a mode.
For a spotty connection, mode 11n is much better
@passthejoe @brynet and yeah, freebsd's iwm port happened early in openbsd's dev cycle and I haven't finished updating it to what's needed to start doing 11n (and then 11ac.)
(and then do the same for netbsd's iwm, most likely.)
source: me, freebsd wifi
@passthejoe @brynet yeah! i was hoping someone would pick up the iwm reigns and run with it since it's mostly just porting openbsd bits to freebsd bits but for some weird reason people don't like touching wifi stuff. but everyone HAS wifi stuff.
So yup, guess I'm gonna do it. :-)
I'm going through the five stages of grief on this and will be back to #NetBSD in the near future.
I spent some time on my VM today and will do another bare-metal install when I can.
Going to try for a wired Ethernet connection (to a wireless bridge) and see how that performs.
"𝙳𝚎𝚎 𝚒𝚜 𝚊 𝚝𝚎𝚡𝚝 𝚎𝚍𝚒𝚝𝚘𝚛 𝚏𝚘𝚛 𝙰𝚙𝚙𝚕𝚎 𝙸𝙸 𝚒𝚗 𝟺𝟸𝟷 𝚋𝚢𝚝𝚎𝚜. 𝙸𝚝𝚜 𝚝𝚘𝚝𝚊𝚕 𝚍𝚘𝚌𝚞𝚖𝚎𝚗𝚝 𝚜𝚒𝚣𝚎 𝚒𝚜 𝚓𝚞𝚜𝚝 𝚘𝚗𝚎 𝚌𝚑𝚊𝚛𝚊𝚌𝚝𝚎𝚛. 𝚃𝚑𝚒𝚜 𝚒𝚜 𝚟𝚎𝚛𝚜𝚒𝚘𝚗 𝟷.𝟶 𝚘𝚏 𝙳𝚎𝚎, 𝚊 𝚠𝚘𝚛𝚔𝚒𝚗𝚐 𝚝𝚎𝚡𝚝 𝚎𝚍𝚒𝚝𝚘𝚛 𝚍𝚎𝚜𝚒𝚐𝚗𝚎𝚍 𝚊𝚛𝚘𝚞𝚗𝚍 𝚊 𝚛𝚒𝚍𝚒𝚌𝚞𝚕𝚘𝚞𝚜𝚕𝚢 𝚜𝚖𝚊𝚕𝚕 𝚋𝚞𝚏𝚏𝚎𝚛 𝚠𝚒𝚝𝚑 𝚜𝚘𝚖𝚎 𝚖𝚒𝚗𝚘𝚛 𝚜𝚒𝚣𝚎-𝚌𝚘𝚍𝚒𝚗𝚐 𝚐𝚘𝚒𝚗𝚐 𝚘𝚗 𝚊𝚜 𝚠𝚎𝚕𝚕. 𝚂𝚖𝚊𝚕𝚕𝚎𝚜𝚝 𝚙𝚘𝚜𝚜𝚒𝚋𝚕𝚎 𝚎𝚡𝚎 𝚜𝚒𝚣𝚎 𝚠𝚊𝚜𝚗'𝚝 𝚛𝚎𝚊𝚕𝚕𝚢 𝚝𝚑𝚎 𝚖𝚊𝚒𝚗 𝚐𝚘𝚊𝚕, 𝚋𝚞𝚝 𝚜𝚘𝚖𝚎𝚝𝚒𝚖𝚎𝚜 𝚒𝚝 𝚠𝚊𝚜.
𝙳𝚎𝚎 𝚠𝚘𝚛𝚔𝚜 𝚒𝚗 𝟺𝟶 𝚌𝚘𝚕𝚞𝚖𝚗 𝚖𝚘𝚍𝚎 𝚘𝚗𝚕𝚢."
https://github.com/mpower-codeworks/Dee-Text-Editor-Apple-II
Good progress today on miware (my DOS for #ilo). I have a working, though very spartan, line editor now, and adapted my #uxnmin for ilo to run under it.
I'm going to try to get most of the remaining utilities for this completed over the next couple of weeks, but at this point I should be able to do so as long as the RSI pain doesn't get any worse.
also rsi sucks. takes forever to heal ❤️🩹
@hi I've largely given up on actually healing from the RSI issues. I'm happy as long as I can manage to keep working on my projects, even if it does mean that I sometimes have to take much longer breaks than are ideal.
i wrote down my current thoughts on AI, first draft, titled "AI is Fake People": https://minute.social/index.html
what do you think?
i also should say that sections from Jordan S. Carroll's book "Speculative Whiteness—Science Fiction and the Alt-Right", which @holo_memory showed me a while ago, influenced my perspective on tech moguls' too literal embraces of sci-fi ideas https://www.upress.umn.edu/9781517917081/speculative-whiteness/
funny, after reading the draft, @dngrs sent me this video by Another Roof, which is concerned with AI in mathematics but hits on very similar fundamental themes, which is that we enjoy art made by humans, and this also applies to mathematics https://www.youtube.com/watch?v=o0MjPe7wXs8
@mntmn
> In the same way in which I want to experience art made by other humans as a form of emotional and mental connection, I enjoy using software, hardware, and other products that are made by caring, considerate human beings, and not by unfeeling robots. All of these artifacts are essentially fragments of human communication
So much this.
@mntmn What do I think? I think that I wholeheartedly agree. I think that you write with a beautiful clarity. I think that I'm more happy than ever that I decided to order my next laptop from you. :-) Please keep creating--with words, with code, and with circuit diagrams.
@mntmn I was speaking at a conference recently and many of the others there would do this charade about their business not heing possible without AI. I'd ask how and they'd say "I couldn't possibly do all the work myself" with no hint of irony when I talk about nobody hiring when workloads get tough and instead throw money at AI... which they agree hurts the community.
@mntmn These are people who want to be one person bubbles. They reject the importance of others. Is it main character syndrome? Is it greed? Is it classism?
There is an oversized component of this which is all about labor. It's not just jobs we need but the insistence that it's always better to throw away old ways of working. It's an obsession that this is part of the cost of progress and progress is more important than humanity itself.
@cararemixed @mntmn for me it’s closely linked to commodity fetichism. Code, like other arts, are crafts that are intimately linked to the worker who produced them. They are less perfect commodities because, even if you pay for them, you can see the fingerprint of the human who made them. On the other hand, AI products fullfil the capitalist deception of a pure commodity and are more in line with the dominant ideology.
@spell_castro @cararemixed "pure commodity", that's a useful term, thanks
@mntmn this is without a doubt one of the best pieces of writing on the topic. Thank you for sharing your thoughts <3
@mntmn Lucie, that article is spot on.
The kind of people who push this are power hungry, their empathy is damaged.
We can tell the care and empathy you have for the mnt community.
@mntmn I'd love to boost this, very well thought out and written, but is this a permalink? Seems like if you blog more in the future, this would shift out from the main index.html, but correct me.
@mntmn I'm reading Magnifica humanitas by the Pope. It's vey similar yet it falls short of consequences. Doesn't makle it less valuable because any reader should come to the right conclusions.
Life is art, programming is art and letting a machine take over is cutting short on what's beautiful in life.
> The people who try this stuff are always of the same kind
I think this is wrong. It might feel that way, but it's weakening the argument. There is more to explore there about what drives people into all this.
> I had a realization that *pain* is the main anchor of reality
I think you're right on the money with this one. And that is also why the idea of AI alignment and trying to teach it ethics is not going to ever fully work.
Ethics are an emergent property of shared vulnerability. Something that can't die cannot truly feel empathy for something that can.
> because they only see the results
Again right on the money I'd say.
The obsession with "efficiency" and believing that it's about the destination, when in reality it is about the journey.
@mntmn
> they lack empathy and are, above all, power hungry. They see others as playthings, or "toys"
I guess it’s a common “just business, nothing personal” thing. Yes, everything/everyone is a tool in this POV
I think you put up a thesis and then wrote about 3-5 other things that are on your mind, but aren't actually related to the thesis.
Whether or not I value genAI "Art" or human "Art" has nothing to do with making fake people or causing intentional suffering.
I do agree that the remixed stuff isn't as helpful as the marketing says and that "not helpful" is on the bad side of a "feels good to use to feels bad to use" spectrum.
But I think that about human output too, so...
I've been carrying my Assumed Intelligence and Unsupervised Text Extruder dread around for a similar amount of time and today I started writing down my own concerns.
After sending them to a friend I read your thoughts and discovered a whole other angle that was hidden to me, even though as a software developer with over 40 years of experience, I've regularly drawn parallels between art and the process of software development in all its frustrating glory.
Thank you for bridging my gap!
@mntmn like it very much and <3 "please don't sacrifice human connection on the altar of productivity."
@mntmn
Excellent piece!
Also quick note: You're linking to the homepage of your blog, i don't see a way to link to JUST this piece but that will be needed for sharing, as once you write something else it will get pushed off the front page.
@mntmn i'm 100% with you on the second half (discussing programming as an art form and as an empathetic endeavour), but not really on the first one (transhumanism being exclusively a billionaire's fantasy). to the extent i was given a human mind i'm not happy with it and i won't be until i shape it to my own ends. do i get to be put into the same category as the likes of Musk and Bezos, now?
@mntmn the other, not really immediately relevant but still important thing is... i can conceive of a machine mind being a peer (though not an equal), in theory; in the same way i can conceive of a cat's mind or a rabbit's mind being a peer (but not an equal)
LLMs are obviously not it, and it's obvious that the people arguing for this being the case right now are doing this in bad faith. but i don't subscribe to what i will (perhaps uncharitably) describe as "human chauvinism". haven't for a long time before the current discourse has been relevant, if anything, the tipping point for me was how people treat pets
this sort of thing is why i haven't, personally, affiliated myself with any of the number of "by humans, for humans" movements in computing despite my work being de facto part of a shared effort. i just don't feel good about accepting that kind of worldview. i don't really talk about it because it would be needlessly divisive in almost any context
@whitequark the thing about pets really resonates with me. i've thought about animals multiple times while writing this, but didn't know how to include this topic. another self-critique that i had when i went to bed is like "i hope it's clear i don't mean prosthetics", and assistive technology. which is connected to your point about transhumanism. i'd call myself ex-transhumanist maybe, and recently learned about posthumanism. i'm not against augmentation and cyborgism, and am transgender.
@mntmn i still don't actually know how we (as a society) should deal with cats. for example, TNR programs are probably worse from a genetic perspective than outright killing problem animals (though not as effective). on a less morbid note, cat colonies clearly have culture and approximately not a single person, no matter how cat-minded, would consider destruction of this kind of culture... consequential
(not isolated to cats, but cats are a good benchmark in that most humans are obsessed with cats in a positive way. so other animals get it worse in many ways.)
@mntmn and yeah, I make no drastic difference between "posthuman" and "transhuman". the ends of the axis are (this is mostly a joke) "catholic" and "everything else". i'm on the "everything else" end
@mntmn re: assistive technology: I posit that there is no fundamental difference between "assistive" tech (as in prosthetics for disability) and "enhancing" tech besides, perhaps, intent. i find it far more concerning when someone is more bothered with enforcing the natural order of things than actually making sure another's life is good
birth control, for example. most cis women can safely opt out of menstruation, or at least reduce the frequency to maybe once or twice a year. this is something that is between "utterly medically necessary" and "incredibly convenient". but it's only really getting recognized now in medicine
@whitequark I think you're onto something here. As a disabled (legally blind) person myself, I'm suspicious of policies and positions that try to hold onto the natural order of things or old ways of doing things, while allowing a narrow exception for recognized disabilities. One example is going back to handwritten assignments and tests, while making an exception for blind people or people with mobility impairments. Yes, *they* can use a computer, a restricted one.
@matt @whitequark yes, i get that. i'm often asking myself, "am i just becoming conservative?" but i don't think it can be simplified like that. my focus, for now, is really quite narrow and concerned with making "works", or "art", and i'm trying to find a frame for my own works in that category rather than a commodity/automation/effiency category. i don't really subscribe to anything as being (un)natural, but rather trying to find useful axioms that can anchor my experience in reality.
@matt @whitequark to overstretch this point a bit, what i mean is deciding on reasons for accepting this reality as a meaningful one to inhabit, despite all the suffering. mutual sharing of this condition and communicating it through "works" is such a reason for me. if this is taken away, there is only the absurd left, or "the nausea" of depersonalization.
@mntmn @matt yeah, i respect that. i don't share the desire to avoid depersonalization or to find meaning in pain—dunno if you have chronic pain yet but that sorts that part out real quick—i think personhood is at least as commonly oppressive as not—but it's a good reason, at least
i don't think i have a framework like this built for myself. i don't think i need a framework like this built for myself; i would only need a framework to convince someone, and i needn't convince myself that the way i want the world to be is important
@mntmn @matt one thing i come back to is: efficiency for whom? like, the EU border is very efficient, if you're white and speak English, for me it is and has always been the opposite. efficiency at achieving what outcome? we both know what it does.
i don't really abandon the notion of efficiency. i could let my body fall into such level of disrepair that i do nothing of meaning again. that would be inefficient, and that's the strongest argument not to do it. but it has little to do with the commodity production
@whitequark @matt yes, my main takeaway is that i need to be a bit more verbose to avoid misunderstandings. what i mean is the dogma/altar of efficiency/productivity gains as the highest goal in commodity production, not e.g. in private life.
@mntmn
Important perspective. Thanks for the write-up.
I just wonder: what should we do as a socitity? We will not get it back in the bottle. Do you have any ideas how to handle the situation?
@mntmn I think you really judge a bunch of people you don't know rather harshly.
I actually used to have interest in recreating actual human thinking in a computer and it had nothing to do with a lack of empathy or need for power but all about trying to understand how I think.
So I made it about 3 paragraphs in and got super offended and stopped.
Now everyone can call me some AI techbro or whatever.
@crazyeddie i'm sorry, but what i didn't say is that this topic is or was of course also interesting to me, but not in the way that AI tech bosses are doing it. do you want to tell me about your approach?
@mntmn Well, then it's odd for you to say anyone who is or was interested did so because of lack of empathy and a desire for control and power. Maybe that was your agenda but mine was understanding.
I don't have an approach. I read some Dennett and Hoffstadter, made a few rudimentary AI before LLMs, had some manic ideas, and did some BCI stuff that turned out to be fairly rudimentary. I do believe a computer sentience is possible since that's what we are, but they're pretty far from it.
@mntmn good article! good code is also a joy to read. I don't think we should valorize suffering (although I don't think that's your intent), but I do see the current rash of AI marketing is callously indifferent to it. Like, if a person believes that their computer system is able to suffer the well adjusted thing to do is stop developing it, then accept your personal moral responsibility for that system and then never do it again. But no, suffering at scale is the frustrated customer's problem
@mntmn I disagree with the parts about pain and suffering being essential for art (I have found them both to be severely detrimental for me and to the numerous artists I know), but I broadly agree with the observation that the desire to create a fake human servant comes from a fucked up place, and with the necessity of empathy.
@mntmn it is hard to pin down exactly what art is and why we do it. I think the philosopher who has gotten the closest is Lynda Barry. Her theory is it serves an important biological function that can be observed most clearly in the way children make art, it is something stays with us our entire lives, but it is usually repressed by the time we reach adulthood. I think pain and suffering are more often the means of that repression.
@mntmn I think what motivated ancient people to fill caves with hand prints and drawings of animals was motivated by neither pain nor suffering but how beautiful and alive the world is. A hand print of a child on a wall much to high for a child to reach is a story about something else.
@mntmn I think the precise motivations differ from person to person. I find that the act of creating art is a euphoric process, and that the euphoria concludes shortly after the working session concludes. This can be positive for my mood, as long as I was not too depressed to experience it. I would say the primary motivator for me is I see things in my mind's eye and I want to record them in a capacity that allows me to share the experience. I also draw as a form of meditation.
@mntmn Yeah and AI has done some harms that are affecting us globally, however it has helped a lot in some other ways
@mntmn I think the "pusher" terminology is apt. Many marginalized people have been sold a dubious vision that LLMs will level the playing field, and the blame lies with the pushers. Your writing also makes me think "who decided software engineers needed to be *less* empathetic?"
@mntmn You are are touching on an angle that I haven't seen clearly stated before: "consider Claude, Muse, Grok etc. themselves as works made by humans who are clearly deeply unwell"
Seeing AI as the expression of 'fake people' - people completely disconnected from their core humanity, inflicting the resulting suffering on the rest of us. The actual process of getting and staying in touch with one's true self requires work and is painful. The process makes us truely human and can't be faked.
@mntmn Enjoyed the read very much: both the analysis and the beautiful language, thank you for sharing! Also, low key Mark Fisher vibes.
@vadimkotov thank you, the mark fisher reference really sent me down a rabbithole now, hard to believe but i didn't know him yet!
@mntmn there is the argument that LLMs provide access to people for whom computers are opaque machines. Using computers was like cryptic wizardry, but now they can get things done without relying on others (and their patience / judgment).
@nintegge i don't believe that this is true.
@mntmn which part of the argument?
@nintegge the access/opaque part. lets take macintosh or ipad for example (i'm an ex-apple user from like 10 years ago). these were designed to be easy and accessible computers, and successfully so. or if you mean application development, there is a long history of "easy" tools like hypercard, scala (on the amiga), excel, visual basic, access, filemaker. google docs. or remember people messing with myspace and tumblr. website builders and no-code tools. why are these suddenly no good?
"now they can get things done without relying on others"
llms inherently rely on others, the humans on the internet and elsewhere that generated the training data. this unpaid labor is obfuscated, digested, and sold back to users regurgitated.
incredibly weak argument when, as lucie stated, the bar to making so many things was already incredibly low in many key areas.
@mntmn i earnestly think it is because those systems take effort to learn to use and are limited in their capabilities in ways that LLMs are perceived not to be
plus, it is difficult to get a job programming in scratch or hypercard
meanwhile, with an LLM, you can make drive-by open source contributions to your favorite projects without even learning to code, and the way things are going, odds on it'll be accepted
@mntmn In many parts I agree, but not fully on the art thing. Yeah, there can be code that is art. But most isn’t. It’s like with decoration for me,, where there are unique art pieces, but most isn’t industrial produces „functional“ decoration. Or furniture, where you can have a bed well made by hand that can count as art, but a majority of beds just are mass produced.
For you as programmer, code can be art. But for most people it isn’t even there, the application is and has to serve a purpose.
@mntmn Too much to read and digest here, and so many good and valid opinions, but my take on the headline "Fake Humans" misses the mark for me. It's humanity stolen from each and every one of us vomited back in our faces with zero regard for the value of actual human work, contact, community. It's just a data processing machine on an unprecidented scale, harvesting our quintessential existences for junk food for unmindful consumers.
@mntmn CLASS CLASS CLASS
i LOVED reading this
the LOVE you put into this piece, oh my God, i'm actually feeling emotional
thank you for writing it
❤️
$ git clone https://github.com/netbsd/src#runbsd #solarpunk #netbsd
Cloning into 'src'...
remote: Enumerating objects: 6112442, done.
remote: Counting objects: 100% (405/405), done.
remote: Compressing objects: 100% (157/157), done.
remote: Total 6112442 (delta 314), reused 248 (delta 248), pack-reused 6112037 (from 2)
Receiving objects: 100% (6112442/6112442), 3.33 GiB | 30.09 MiB/s, done.
Resolving deltas: 100% (4439958/4439958), done.
Updating files: 100% (206518/206518), done.
$ ./build.sh -O ../obj -U -u -m amd64 -N1 -j4 release
...
are there any interesting new(ish) very-high-level "easy" programming languages that simplify generic application development for non-developers? i.e. what people would use now (aside from excel) to make their own apps/tools if LLMs hadn't happened
i'm asking this because this might be an important avenue if you want to offer an alternative for folks that are using LLM/chat bot coding services now but rather wouldn't, but traditional programming languages have relatively high hurdles
@mntmn I’ve been on the lookout for this type of stuff for ages. I had hoped something atop web tech would appear, but the 2010s only made everything a billion times more complicated.
There are a few products, some with FOSS-distributions like Budibase, that make things very easy, but I don’t feel like we have this as “a community stack” anywhere.
@mntmn It might be worth checking out some of the options in
https://en.wikipedia.org/wiki/Visual_programming_language
The slant of that article is "education", but I think there are (still) people trying to do production work with them.
@mntmn Something that rings a bell is a few years ago, when companies like OpenAI were still incubating, I helped a client identify a number of different software vendors who provided drag-&-drop tools where people in the business line could define their process and the compiler translated the graphic objects to high-level code like Java or Python. The code was usually messy and often designed to run in the vendor's "server software" but it did provide business value.
@mntmn I saw a few people mentionning no-code in replies, it was gaining some traction around 2018~2020 if my memory is correct. Like you did mention, some/most did an AI shift, like zapier.
if I remember correctly, it was mostly stuff similar to ittt (if this then that) connecting input to database/document/key:val storage
@mntmn I can’t help but feel that the problem is starting at generic programming languages. Jumping straight to Turing completeness when most people just want like, “do x to this list of things” but have to deal with the sort of risks that the eBPF bytecode compiler wouldn’t let you put in the kernel. Same as if you just want a message processor but you still have to deal with a filesystem and json deseriatlization and and and… none of this has anything to do with what you’re trying to do.
@mntmn had LLMs not become cheap, programmers would have continued in the languages they still use.
Non programmers would have continued not to program.
There is no cheap way around this.
I expect once the cost goes up most of the new llm users who aren't programmers will return to hiring professional programmers (and artists and writers).
Any no code tool will be targeted to non developers and will try to extract payment.
Most useful desktop apps are best done in python. iphone is pro only
@mntmn
To answer your question to be honest , I would lean into python ,
Its straight forward and provides a robust library thats well maintained ,
and while your putting those Raspberry Pi RP2040 controller's on those computers there ( Which is very nice and impressive by the way ! ) ,
the language used there for that controller , would be microPython , which is similar in some context and structure.
@mntmn in the tradition of the good old hypercard this young and fragile plant comes to my mind: https://www.hyperxtalk.com
@stoltenberg ah, that looks like applescript
@mntmn @stoltenberg Looks very interesting; however the book from the project was “written by Claude, proof-read by the project lead”.
@mntmn
If you mean low-code or now-code tools I really like Rei3. https://rei3.de/home_en
I really like their app signing an versioning workflow.
@Ekpu thanks, that seems very interesting, although the UI looks a bit long in the tooth. gonna check it out
@mntmn There's a whole genre of tools called "low-code" or "no-code" the predate LLMs. Though they are still probably an order of magnitude more complicated for non-programmers than LLM coding. Examples: Budibase, Appsmith, Tooljet, Retool. (I would bet that many of these pivoted to being AI-driven.)
@sarahstreza right. i checked some of these out a while ago and indeed there are many ai pivots. i was wondering if someone already did something post-ai in that area, so to speak.
@mntmn @sarahstreza I can vouch for the underpinnings of Budibase to be relatively solid (disclaimer, they use my tech and are a customer of ours), but I can’t speak to how much the AI marketing messes with that or if that’s just a conveniently ignorable layer atop.
But yeah, still complicated and def “requires people to think”, for better or worse :)
i guess this is not what you are asking for, but i regurarly use max / puredata / plugdata for sort of these tasks. like automatic control over a video switcher for example.
and very simple bash scripts using things like pandoc etc. i use a lot. but still sometimes a bit of a sudoku for my brain tbh.
@lislegaard right, i think the whole music and performance software landscape maybe doesn't have this accessibility problem in the first place (cool) because there have been so many creative UI approaches and modular systems for audio.
yes i think you are right. which kinda prompts the questions, what are the different use-cases where people use llm's because things are not as good as in music / performance?
if i REALLY think about what i see around, there are not many ai-coded products i care for. i guess many people augment and automate with ai, but it is not like they woulddn't have made it without ai.
so.. most ai is back to slop? or what?
maybe debugging and small tools? i remember someone lost their bluetooth device and used claude to make a small app that would show bluetooth strength in a way what they could use it find the device. pretty handy and i don't think they would have done it without ai. BUT i feel like this is a thing we should be able to do with computers without ai even as non-programmers anyways.
but yeah. i think i would be interested to hear what kind of stuff these ai-people are making that you are
asking for the alternative to.
edit: asking for alternative workflow for MAKING. i know you are not asking for the ai-created apps themself.
@lislegaard what i've heard so far are things like:
- website and booking system for some small business/group/event
- agent that runs ghidra to modernize some old software (more advanced case)
- agent that reverse engineers/figures out how to drive some piece of hardware
@mntmn I love this question, in a world where we have things like Gnome Builder (where you can point and click an interface but then you're faced with wiring the backend C code correctly) there's gotta be something that makes the backend simpler... All the languages I can think of are education focused so while you can do cool stuff with them it ends up complex
@mntmn non-developers become developers when they code though? I find this line a bit to fuzzy to make a recommendation. If you look back at what "non-developers" did with VB6 it was impressive. I think we lost that with the web. It doesn't have to be Basic but it's been there since early personal computing and it's not ideal but its approachable.
@cararemixed yes, i agree, and trying to think of ways to very gently introduce programming syntax in something that i.e. just text/images first. for example, if you just write the word "hello" and run it, it would just show that word in the middle of the screen. and then you can gradually add stuff to refine it. html kind of works like that, it just has not-good-looking defaults and too much boilerplate. i guess that's why we ended up with markdown. but markdown lacks interactivity and forms.
@mntmn The right bit of extensions around HTML could do some of this. Avoiding JavaScript would probably be required as it'd be a very leaky abstraction. HTMX style decorations could go some distance. Maybe wrap it around CouchDB so apps can be served, data held locally, and application installed as replicated data. I personally feel like the web has great ideas but the temptation to break out of a restricted subset ends up ruining most of these.
@cararemixed yeah, i mean with normal HTML i think it already loses at <!DOCTYPE html> and <meta charset="UTF-8">. that's just an example of bizarre things we take for granted but that don't need to be there
@cararemixed i remember learning C as a kid and i was bewildered about having to write #include <stdio.h> and having no idea what that meant or how to memorize it
@mntmn The idea of forking the web has a lot of appeal but I don't think we're in a world where we can build momentum without community to grow it inside. We can try to hide details but it seems like we're setting ourselves up as just an easy-mode over "the real thing" which is not the point IMO. The simple and hopefully easier thing should also be the real thing.
For general-purpose high-level programming, #Python is still very approachable and capable. Very well documented with plenty of learning avenues for all experience levels.
For actually starting a newbie programmer, a general-purpose language would probably be too much, given the implied learning curve and delayed payoff. It may be more fruitful to know exactly what they want to do and pick some more focussed automation / hacking that will give rapid results for their initial efforts. @mntmn
@mntmn Not that I know of. It runs opposite to what I have observed in this industry over the last ~40 years.
Pascal was great. (Delphi was amazing.)
C proved faster. But, C is dangerous. The internet fostered C because geeks who avoid general social contact liked it, because C is dangerous, so using C demonstrated geek machismo.
So C grew huge and took over both the Windows market and the Mac market (via NeXTstep, which is a Unix).
That has led to 2 things:
1. The rise of "better C replacements" which are _harder_ -- C++, and then after that rose and fell, Rust.
2. A steep decline in easy entry-level tools. Hypercard dead, the main clones dead, Visual Basic dead, Excel deeply uncool, Java banned from the WWW and now a dull enterprise tool, etc.
@mntmn Several variations have been possible to use with the micro:bit - https://tech.microbit.org/software/
@mntmn lua and many others were already listed.
I have a language I wrote, that is super simple.
I've been thinking about turning it into a type 2 lisp, at the moment it's only syntax is white space.
It uses yaml for data. It is stackless, built manually.
It auto fills function calls.
It's written in python so it can transparently use any python library.
I created it solve problems I was having with a contract to write a flashing and testing app for a little iot board.
It's super easy to write apps with a process flow.
I've thought about writing a wumpus game leveraging the manual stack. I need to revisit SPR.
@mntmn Visual BASIC/Gambas comes to mind. Scratch is also a building blocks style build environment. Both require minimal code to perform basic GUI tasks.
@mntmn Decker is the first thing that came to mind personally: https://internet-janitor.itch.io/decker
Sort of a spiritual successor to HyperCard
I desperately want to see more things like this though, and I'm especially hoping that some of the work @spritely is doing can eventually be incorporated into an environment like this
@mntmn
Weirdly, I’ve seen people use Godot game engine to make apps. Build you own UI. Export to Linux, Apple, windows, android, etc etc etc
@mntmn A bit out of the blue, but Elixir is an approachable language with Erlang/OTP behind it.
Ash framework (disclaimer, it is partially ensloppified) lets you skip a lot of incidental complexity that modern development always struggles with. You can get your entire functionality done before even thinking about how it should look or what transport protocol to choose.
@mntmn hyperscript is a language that has a lot of its philosophy based on hypercard. I'm not sure if you could create a whole app with it or not (and it'd probably be a website if you did) but it's one of the languages most close to English that I've seen in a while
@mntmn i would recommend signing up to scrimba.com it's the most engaging/interactive learning site and it covers everything - from css/html/javascript up to reactjs and even more.
there are lots of pathways that guide your learning and I think you'll find something you enjoy learning, and plenty of places to apply that learning and you'll find a path that suits how you learn and what you like doing with code.
@mntmn Filemaker sits in here. more a program builder than a programming language. my dad built his own inventory and tracking system with it.
@mntmn unfortunately, every time we've created a programming language for this purpose, it's never turned out easy enough that anyone can just pick it up without adjustment. COBOL was supposed to easy enough that business people could write computer programs for their needs without hiring a computer scientist and today it is one of the most niche and difficult languages in existence. Python was designed to read like English but it is a machine learning language today 1/2
@mntmn Learned about the history of VisiCalc recently (precursor to Excel). It was the "killer app" for the Apple II. Businesses choose the Apple computer as a platform, with the sole basis being to run VisiCalc.
The more I learn about the world, the more I question the origins and organic success of things that are heavily promoted as the canonical chosen standard all humanity must adopt for reasons, i.e. excel.
@mntmn I think the language itself won’t matter much, but the human community around it will. A community that isn’t grumpy and has a lot of people willing to help train newcomers is what is needed most. After community the next most important thing is documentation and clear tutorials so the newcomers don’t turn to LLMs for advice.
The onboarding problem is a community/social thing.
@mntmn uhhmmmm, I don’t know if this will be helpful… but Julia is something my husband uses for data analysis…. Again I’m not a programmer, but I’ve heard that this program has a nice language something something similar to ease-of-use like jsons… 😹 (I’m only suggesting what I know, which isn’t much 🥴)
@mntmn In the 90's that was the premise behind MS VBA and VB6 (Visual Basic for Applications). Then in the 2000's VB.Net. They made it easy to automate MS Office apps and also a plain GUI app. I still support apps with VB, though I usually use C#.net for PC apps.
These days isn't that what Python is supposed to be?
@mntmn "Briefcase" converts Python projects into standalone native applications, for Lin, Mac, Win, iOS, Android and web. Writing the code of your app is just raw dogging Python though. https://briefcase.beeware.org
@mntmn if its about logic flows, node-red has been somewhat big in the industrial space.
n8n is an open source IFTT/Zapier kind of thing, the latter I would say are the biggest tools atm for writing software of sorts. The whole low-code/no-code trend from many years ago sparked a couple tools in this space.
Many more proprietary platforms, AirTable for anything data heavy f.e.
Idk if Odoo maybe offers something in their open source portions
Please don't scroll past. This is important.
(edit: Note that the FreeBSD subreddit is in no way affiliated with the official FreeBSD project.)
I am being pressured by FreeBSD's reddit moderation team to censor the Omarchy #content_review or to remove it completely. They have began a harrasment campaign with the goal of removing me from this world.
the review: https://merveilles.town/@prahou/117501303459299869
The issue at hand is that a picture of a partial swastika appears in close proximity to the word FreeBSD, this they perceive as an attack on the FreeBSD project - if you've seen a newspaper before, you might understand the principle of two separate unrelated articles appearing next to each other. You might also understand that I am reporting on common associations with the Topic of the review. These are not my inventions, and I am not calling anyone anything. I am reporting on the news. Nevertheless, as much as I am a journalist, I am an artistè, and if my audience sees in my work something I didn't put there, I am not going to take that experience away from them. This probably says more about them than it does about me.
While I am a man of principle, there is one thing I like more - money.
I will delete it for $2,000. For $5,000 I will apologize. For $10,000 I will use daemonarchy exclusively for 6 months on all my computers. For $25,000 I will not do anything. For $75,000 I will not post anything for a month.
https://liberapay.com/analognowhere
https://ko-fi.com/analognowhere
To everyone who likes my work, this message is for you (from said freebsd moderator) [cw ableist slur]: https://beige.party/@fuzzy/117510032933612034
To everyone else who feels personally attacked by my art: send money, you can end this here and now.
Thank you for your attention.
pls boost
#runbsd #omarchy #fediverse #fediart #unix_surrealism #swastika #freebsd #news
@prahou damn, people online gotta offend on anything at this point. I'm with you on this one, and I think they should really go get a newspaper from down the street for once 😅
@prahou this person is probably upset the original review didn't have enough slop. maybe it's time for a bonus freebsd issue? extra up to date coverage.
@cararemixed Thank you for your support. I'm pretty happy with the redacted FreeBSD issue. I think this is a great look for their project.
@prahou Hmm, I don't see this issue of Funhole Content Review in a negative light...I see some satire, directed at a particular group of people, sure, but this is a reoccurring theme of your art. Eh, they can pay you if it so deeply offends them. I hope they do.
@WanderingInDigitalWorlds Thank you for your support. I don't see it either, that's why this whole drama took me by surprise.
In fact, some of my best friends are FreeBSD users!
@prahou thank you for explaining the controversy for those of us who have not seen a newspaper before.
@prahou
Power tripping mods are going to power trip. You make fun art on the internet that's multi-layered and interesting. They moderate a subreddit and ban people for saying AI sucks.
They just want to play with the new fascist shaped distro and are glad someone ported it to their favorite os. They got a look in the mirror and didn't like that the mirror was there in the first place. They made the connection.
I hope they pay up lol
@prahou FreeBSD!
...wait no not that free!
@Cat Thank you for your support, sib.
@Cat @prahou I'm still running FreeBSD because:
A) I'm history's greatest monster,
B) I'm a lazy bastard,
C) I love the smell of napalm in the morning!
D) All of the above.
I already put my Slytherin banner away, I dunno what else performative product choices I could make for sanctimonious people to like me.
@prahou
I don't have a cent on me to support properly but my thoughts and prayers...
wait wrong social network, it ain't fascistbook :-)
I guess your art touched him in places he didn't want to be touched ehehe.
@prahou too much politics in this issue, sorry, I can't support anyone with diacritic in the name. Please rename yourself into something bland and generically anglo-saxon. Until then these €17 are staying with me (it will be spent on beer)
@prahou I am fully supporting you but I am also a bit confused as if I should send money or not to keep the art up sooooo.... will a follow do as well ?
@prahou well that's certainly one obvious self-report by the freebsd people
@Ember I don't want this to be the take-away. I'm certain this is the doing of a single confused man, but the fact he wears the cape of FreeBSD while he does it is probably not a good look for the project.
Thank you for your support.
@prahou
Does the small animal have to be an adult, or do babies count as small?
You know what, I should have never asked.
I feel for you in these trying times.
@prahou Why would anyone care about Reddit?. Didn't they sign over its content to train AI models 3 years ago? 🤷🏻
@prahou wow, there are so many comments on that thread. Who knew there was so many people in the world?
@harib_murshidi @Netraven This is what I get for trying to help people make better art. Thank you for your support.
@prahou If there's anything I took away from loosely following this situation, it is that THIS IS EXACTLY why art like this needs to be made and matters. Also highlights the importance of art literacy...
I just wanted to voice my support. Thank you for making art! God knows I am inspired to work on some myself now!
@prahou I sent you money to keep posting. Please consider it a negative amount towards the sums requested.
@prahou I thought you were joking. Daemonarchy is actually a thing. I hope you get the 10K.
@prahou this complaint is utterly absurd.
Please note that the sub reddit has no official linkage with the FreeBSD project.
I am very happy to support and enjoy your work. Please let the trolling continue!
@dch It took me by surprise, you know. But it's been happening quite often lately - people telling me what I think. I'll continue because I can't stop.
Thank you for your support.
@prahou wow that was a influencal text, i am influenced, you influenced me.
@prahou
What counts as a small animal to you?
@ygor What do you have in mind?
@prahou
Sorry for the delay, I was in a meeting (which was in no way related to my question).
Anyway, would an octopus that can fly at hypersonic speeds qualify as a small animal?
@prahou “I’m offended by the swastika in close proximity to the thing I like”
The same guy: “as long as its ’purely technical’ and ‘apolitical’, I’m ok with actual nazis working on projects I like”
That’s fucking rich!
P.S.:
I'm looking forward to more of your art :)
/\* I'm unable to donate at this moment :( \*/
@prahou Oh wow, that escalated quickly. So the Reddit mod has a pull on the fedi server as well? Wild.
People with larger reach than me should talk about this more to the fedi servers users, I wonder how they will they feel about it once they're made aware.
@prahou youre in good company, infosexchange banned me for calling myself a tranny (im a tranny and dont think theres anything wrong with being a tranny)
@prahou $30k to kill FOSS Angel and promote Grendel to Godhood.
@Officeplant Deal. Paypal?
@claushoumann @jerry Trying to figure that out! I will do literally anything to revert that decision! Almost anything.
Well. A fair amount of things.
@prahou @claushoumann no longer banned. Please refrain from offering to kill small animals for cash.
@jerry @prahou @claushoumann
Probably don't do it for any remuneration.
@prahou i thought your comic was great and could only be misunderstood willingly. wish i had $5000 lying around, i'd love to read your entirely unwarranted apology :)
@prahou wow my home instance sucks
@cyberquixote wait! I AM BACK!
@prahou PSA stands for Panda software announcement
@prahou i'm genuinely confused, is all this about a bunch of nazi bootlickers being offended by a nazi symbol? or did i miss some "nuance"
@prahou I re-read the post in question and it was clear to me it was a satire/critic of Omarchy and its slop "clones"
But, I guess that trying to implement support for software created by a guy who posted on his personal blog awful, shitty ideas is okay (spoiler : it's not, even if you are what you call "neutral").
You have my support !
@prahou Mad about swastika, then uses the R word. Someone has serious cognitive dissonance to deal with.
@prahou Uh, if that guy actually had a stress-induced heart attack a little while ago, maybe he shouldn't go around picking fights on the Internet? You can't just start slinging shit around you and then go "NO YOU HAVE TO BE NICE TO ME OR I'LL HAVE A HEART ATTACK AND IT'LL BE YOUR FAULT".
@prahou the majority of people I ban are from infosec exchange. I hate their community/users. I guess I will ban their server today
@prahou pointing out here for anyone else on beige.party who has a problem with fuzzy's behavior here, he's probably broken two of beige.party's server rules. Beige shitposters can report them.
2: No racism, sexism, homophobia, transphobia, xenophobia, ableism, casteism, or any other form of bigotry.
4: No harassment, dogpiling, doxxing or otherwise abusive behavior.
> While I am a man of principle, there is one thing I like more - money.
... because I'm an artist :)
that said, the accusation is bonkers.
@prahou ok but how much do we have to pay you to eat a bug? #ThisWholeSituationIsLudicrous
@arclight that ship has sailed, can't do those kind of rewards anymore..
Thank you for your support.
@prahou Thank you for your service.
For the record I was just checking what the current "I will eat a bug for <X> dollar" value was. Something something, prediction markets, gotta unload these NFTs, etc.
@prahou I can offer you $20 but you have to use NixOS for 7 hours.
@mica In a single or over many sessions?
@prahou your choice.
@mica how would you verify i did it? do you want me to stream it?
@prahou I'd expect a slew of comics with references that only a NixOS user would understand. A snowflake smoking crack? Nvidia driver tanks? I dunno, you're the artist!
@prahou@merveilles.town fuzzy already prematurely blocked me. lol. lmao, even.
@prahou i have no idea what freeBSD or omarchy are. or computers for that matter. i'm just glad and thankful i get to see really cool Content beyond linuxmint comprehension.
I know you took that reward tier off but can I pay you to swat that fly that got into my bedroom and wakes me up by buzzing near my ear at 2am?
@prahou@merveilles.town holy shit this is some of the most biting satire i’ve ever seen. keep it up!
@prahou so hang on, dk if I got it right. a fascist, namely DHH, made a nazi thing, aka omarchy, and put free BSD in it, or made free BSD a part of it, then the free BSD project didn't object or say we're not associated with any of this but we can't stop him from using our code? More so, some people from there actually endorsed it, and then the same people are offended by you rightly putting a nazi symbol next to their project's name because that's how they act? that right so far? If so, they deserved it for sure, and no, you definitely don't have to subject yourself to the torture of having to use omarchy, even if they give you the money you mentioned in the post, because screw that thing and may the whole thing be washed away by the metaphorical sea.
@prahou the founder, the fucking founder? and they are still offended by you calling truth to power? what the holy hell universe is this? I know we ran out of timeline resets a while ago, but darn it, I wish I kept one :p
@esoteric_programmer @prahou Calm down, he's not the founder, though he was a key participant in its formative years. He went to Apple, left, joined one of the vendors, eventually migrated to NVidia; I do not recall if he is still employed by them, but it may explain his apparent proselytism of something I do not condone. I don't personally have anything to do with him, and he's not been directly involved in the community or its development workflows in decades, which bears pointing out.
@bms @esoteric_programmer @prahou never in the history of "calm down" has asking anyone to calm down ever worked.
@prahou I read what I could find. It can be scary being targeted by people who make a big misunderstanding. But I think you're handling it perfectly, and you're in good company. ❤️🩹
@prahou This is a hilarious coincidence, as I was just remembering the amusing and 100% effective trick of saying "Don't be a nazi" into the void, and not having to wait long for someone you never knew existed to shout back "Don't call me a nazi". https://mastodon.social/@jripley/117513645874985052
@prahou oh, its p•rrin who got upset at it. that tracks, they’re an AI bro, so calling out omarchy may have struck a nerve
Let it be known that I love consuming Content and providing you with engagement.
What the scallop
I've heard of some Fedi people having attitudes akin to reddit moderators but I wasn't expecting the literal example to show up 💀
@prahou It seems like furry.engineer has that guy suspended, so he must have a history of this behavior. I wonder why beige.party doesn't do anything about it. Do they allow this?
@prahou I will stand by your side. When I feel sad I seek and find comfort in the world of technomage. Your art is very bunsome, plus: Everybunny loves Grendel! :3
Please consider my previous purchase of your stickers as license to continue being a gadfly in perpetuity. 🤣 🫶
@prahou I once saw a swastika on my computer screen, and sometimes those pixels are also FreeBSD and kittens. How could this happen?
@prahou I've always been a fan of this tactic. Except in fascist instances, the money I'd receive for bought silence would enable me to do recurring monthly donations to things fascists dislike (Mermaids, Trevor Project, etc.)
@prahou No, will not boost. If you're a journalist, you know about positioning, especially if you're the one doing the positioning.
And, as soon as you said "While I am a man of principle, there is one thing I like more - money." you hit my "go fuck yourself trigger. This is not a free speech issue - it's a naked money grab.
@nzlemming Thank you for your support!
@prahou either you're being sarcastic AF or you misunderstand what support means.
@nzlemming Thank you for your support!
@prahou Or you've got a bot running.
@nzlemming It's my AI girlfriend, thank you very much! Thank you for your support!
@prahou @nzlemming has she used her sephora coupon yet
@ozzelot @nzlemming What do you think?
@prahou @nzlemming i think she is savoring it for the perfect product drop.
@ozzelot @nzlemming You're probably not wrong. She made me print 400 of the latest content reviews just for the coupons.
@prahou This is why you should be using a laugh track. Also, congratulations on the journalism degree.
@nzlemming @prahou I’m got to check the Lost and Found box to see how many senses of humour have been handed in. I’m guessing there are loads. Not everyone gets satire …
prahou time to step up your game and do a dressed money grabs!
@bsandro @nzlemming i need money to buy clothes first! Thank you for your support!
@prahou
I figured out the misunderstanding - your 'money grab' is dressed with thoughts and prayers provided by yours truly, that's why it looks like naked! While in reality it is not.
@nzlemming
@prahou I sympathise with you, and since I share your values, I want to offer you my support:
- full support for 10% of your donations,
- strong support for 7%,
- standard support for 5%,
- thoughts and prayers for 2%,
- either thoughts or prayers for 1%.
Cash only. The sympathy is free!
@prahou had to draw something that entire concept is just... The internet is a crazy place. And best response this one has seen in awhile. If I had money I'd pay you to keep making art
@wavejumper3 @prahou it took me a while to figure out the computer user is not speaking spanish/latin/etc
for fun is pain and pain is fun and computers are both and neither and real and fake
@ozzelot @wavejumper3 Slow down, voltaire!
@prahou @wavejumper3 okay i'm gonna undervolt it a little
that's a computo thing
@prahou So let me get this straight.... The FreeBSD team is upset about.... An unrelated article? Are their brains smooth or some shit?
@DavidsCreation Turns out it's not the freebsd team at all! just a reddit moderator of the freebsd subreddit, who has nothing to do with the team! Thank you for your support!
@prahou Reddit... Of course.... What else is there to say? Truly a platform of brain dead fellows.
@prahou hey have you reached out to anyone in the project about harrassment? I can go run it up the flagpole if you'd like.
@erikarn oh, they reached out themselves to me, and explained the perpetrator (and the subreddit) was not affiliated with them whatsover.
Thank you for your support!
@prahou yeah, it's controversial, but neither are part of the project. that doesn't mean their action don't reflect on the community, but we don't have a hammer/carrot to use here.
@erikarn i added an edit to make it obvious. i can't delete the post unless i'm paid $2,000, i'm sure you understand.
@prahou you continue to be OpenBSD's funniest fighter, thank you very much. I have recently enjoyed setting it up on an old thinkpad and it's as much fun as ever :)
@somebody Thank you for your support! I'm glad to hear you enjoy it, it is still my favorite system.
@prahou
I saw what you did. Couldn"t go through with it, huh? There may be a spark of good inside you yet!
Did I just write almost 5000 words about why generative "AI" should be resisted, going through every single one of my reasons?
Yes. Because I need this article to exist so I can point people to it.
Enjoy and share widely.
https://estherweidauer.com/articles/why-resist-generative-ai.html
I'd like to offer some constructive criticism.
Under "It's just a tool". I immediatelly thought of a counter-example to your argument. Self-hosted llms, like deepseek/gpt-oss. I think you should address these to sound less ignorant and therefore more likely to convince people.
Under the "AI will solve" section, I think some informed people, will point out recent developments, specifically OpenAI's model claiming to solve the Navier-Stokes equations. Something most people thought wouldn't have been possible a few years ago. I think your article would be more convincing if you said something about it. The cutting-edge from the latter half of 2026 are on a different level and should be addressed. And AI solving millenial problems comes with itself concerns (like what is the point if too many solutions exist and not enough humans to verify them).
Under "LLMs are concious", it requires extraordinary evidence to prove or disprove this. I think you should word it differently. Focus more on how the question itself isn't clearly defined and less on your opinion that it's no where near human level intelligence. This section seems pretty biased towards your side when you yourself admit "we don't even really know what consciousness is".
Under "Harms -> Environmental", it would be fun to tie this back to "AI will solve". Rather than solving climate change, it is making it worse :).
I want to convince people to stop using LLMs, but I don't think I would point them to your article right now. It seems too emotional, whereas in my experience you need to present facts neutrally and let people decide. Otherwise the people on the other side of the argument will never change their opinions.
About local LLMs: these are niche products and not what the vast majority of LLM users are actually using. And even then. To remain useful, new versions of these models need to be trained constantly and that training still relies heavily on corporate infrastructure, labor, and data.
@alanxoc3 About Navier-Stokes: if you followed the controversy there, it would be clear that the claims made by openAI are somewhat doubtful. And I'm not going to address every bit of marketing these companies put out.
'Under "LLMs are concious", it requires extraordinary evidence to prove or disprove this.'
No, it doesn't. Making a ridiculous claim and them demanding others to disprove it is the practice of quacks and scammers. That's not how actual research works.
@aesthr Yeah, I know about the controversy, but it is a controversy. Which means we don't know if it's true or false. And even if some work was stolen, some parts would have been original, otherwise the proof would have been completed by a human first. But I am very aware that I didn't read it nor would understand the proof, so take my point with a grain of salt if you want.
Local LLMs, it's your article, so completely your judgement. I just thought maybe you hadn't considered it.
And for the last point about "LLMs being conscious". I'm not trying to fight for either side. The fact that still no one can define what consciousness even means, means that you shouldn't tell them they are wrong and they shouldn't tell you that you are wrong. That's the point I'm trying to get across.
Your immediate reaction was that the other side is "quacks and scammers". I promise I know people who have the same opinion about me when I bring up concerns about gen-ai. This kind of language just divides people further rather than changing opinion. But what both sides can agree on is that consciousness is not even defined, so neither side can be proven or dis-proven and it's basically pointless to talk about it altogether.
I'm trying to be constructive, and as with any criticism, you don't have to listen to it. Your article was inspiring nonetheless and I appreciate the amount of effort and sources you cite. It definitely helped me strengthen my opinion.
@alanxoc3 were your replies generated with an LLM?
@aesthr No. I haven't used an LLM in many months.
@aesthr I really want to be on your side. I'm the only one out of my entire social circle that is anti-LLMs. I care mostly about the psychological impacts it is having. And have concerns about how it will affect my family. But you are being very rude to someone who shares your views.
@aesthr you came into her mentions asking her to change her writing. You were being rude and you're now complaining about the responses you got to your rude, un-asked for opinions. Am I doing the same to you? Yes.
Not everyone has to agree with you and whatever your hangups are about what's been written, and if you go around giving unasked for feedback you should expect some pushback.
Let us build together the largest resource on what is, and what is not #solarpunk!
https://qum.cz/is_it_solarpunk/
Leave a comment, if you definitely know about something that either is, or is not solarpunk!
Thanks!
@prahou Fcking off bigtech is solar punk
@DeltaLima is swearing solarpunk?
ヂオ » 🌐
@dio@kanoa.de
@prahou taking a dump in the woods ~feels~ solarpunk (ecofriendly like, of course, no wiping with paper I guess, burying it like a cat).
@prahou Weaving textiles is solarpunk. Natural fibres definitely, but I think we can weave a little bit of synthetic polymer as well, as a treat.
@prahou an overpriced townhouse with a perfectly manicured lawn and a Ford F150 in the driveway in the suburbs of northern Virginia is not solarpunk
their hardware kinda is though, no? so a hacked and repurposed ALPR would be quite solarpunk.
@mangofeet @hi @mohs @twinkles what is alpr?
@gumnos @hi @prahou yes, but I think that this argument will not be able to stand alone, as a tech could have a very wasteful impact during production or after it's usage period. It also does not say how long each period is. So it would be one constraint to flag a tech as solarpunk, but certainly not the only one.
I feel like this debate is a little simplistic in its ambition, but when it comes to batteries, renewable lithium batteries may be a necessary transitory stage away from fossil fuels, but sodium batteries, based on abundant resources which do not require lots of environmentally destructive mining, is a better choice for solarpunk
@hi @prahou I would say that Organic Flow Batteries could be far more Solarpunk
https://www.frontiertechhub.org/insights/organic-flow-batteries
@prahou
tesla nc 470 hifi turntable is solarpunk
however, the continued vinyl record production requires oil and is, puzzlingly, not solarpunk
@prahou a centralized directory of what is and isn't solarpunk isn't solarpunk
@liaizon Great!
@prahou @maddiefuzz @lambda @liaizon Dreaming of cats may or may not be solarpunk!
@catsalad @maddiefuzz @lambda @liaizon added!
@prahou @catsalad @maddiefuzz @lambda @liaizon
If you’re named Jeff, that’s not Solarpunk, please leave immediately.
I will find you
LLMs are not solarpunk
sabotaging datacenters is solarpunk
setting Teslas on fire is solarpunk
a lawn is not solarpunk
a wild garden is solarpunk
feeding birds is solarpunk
drinking soda is not solarpunk
staying hydrated is solarpunk
@nest Thanks! Not sure about the violent ones, but I added it all for now!
@prahou I think that, instead of allowing cow flatulence to accelerate the greenhouse effect, we should instead capture all of the cow farts into the balloon of an airship, and use that as transportation. That would give a lot of inspiration to watercolour landscape artists.
@prahou wait, if smoking cigarettes isn't solarpunk but batteries are, would vaping be solarpunk?
@bugbear need more input from other experts, but i guess so!
@prahou @bugbear no. if you want to inhale something, you need to build a widely unpractical appliance like the solar powered coffee maker from lowtechmagazine https://solar.lowtechmagazine.com/2025/11/how-to-brew-solar-powered-coffee/ which does not use batteries or PV
@bugbear as someone who's not smoking but vaping , and doing this for four years now only with the energy from my off-grid solar system (no joke. i also charge my power banks for the to-go situations, so my vaping device is 100% solar powered since then), i would say this is definitely solar punk from my perspective :) ☀️ @jollysea @prahou
@prahou
Also I must add to the pooing in the forest and field, that peeing in these locations is also solarpunk
@prahou
And photographing naked hairy girls with an old flexaretík, whether or not they are performing any of the above activities or are even in nature? That, of course, is solarpunk.
They don't even have to be hairy. But I know not whether disposable razors are solarpunk. I'd say they are not.
(Recommended developer is Caffenol. Caffenol is solarpunk.)
@prahou i think if you re use existing batteries it sounds really solarpunk to me. The production of new ones probably isn’t most of the times.
@prahou gonna need the experts to chime in on this one
Is livecoding/demoscene solarpunk?
4kb limit, sounds solarpunk, but often it involves significant wattage from an oversized discrete GPU...
And most of the popular livecode platforms run on the web... Is JavaScript solarpunk?
Many questions!
@the_dot_matrix @prahou When I think of JavaScript, I think of Brendan Eich, who hates gay people, Laurie Voss, who hates unions, and Vercel, who hates the entire planet. I think a sweaty Bulgarian kid filled with yerba mate livecoding with Chicken Scheme on a Raspberry Pi is probably solarpunk, however.
@xan @the_dot_matrix JavaScript is not solarpunk. I know this.
@prahou @xan @the_dot_matrix is HTML solarpunk? Feels like it to me, very open, kind of robust/long-living, and a bit high-tech even
@prahou Mending is solarpunk! TEMU / SHEIN is not solarpunk! Thrift stores are solarpunk! Department stores are not solarpunk! Handmade holidays are solarpunk! Landfillcore is not solarpunk!
@prahou@merveilles.town Smoking is vegan and therefore still solarpunk. And the moon reflects whose light?? Still solar punk! If the moon isn't in this movement then neither am I
@prahou Sharing excess produce from your garden, allotment or greenhouse, with neighbours and friends is Solarpunk.
@prahou Is stealing (I mean borrowing) dishes from your vegan landlord solarpunk?
@timthelion yes!
@prahou What about smoking %100 organic weed in your room?
@timthelion @prahou it's better to use solar bong (with solarpunk certified lenses), if your room has windows
@prahou What about attending a gathering of vegan cultists that other cultists went to by plane but you went to by train?
@prahou
Programming with an LLM is not solarpunk.
Programming with meat _can_ be solarpunk. Except JavaScript. Frick JavaScript.
@prahou Is trans šicí kroužek that is %60 trans people and %40 liberal white cis women solarpunk so long as only šlapací šicí stroje are used?
@prahou could you add links to the comments that suggested each? I just want to talk to the germknödel haters. They are/can be vegan and delicious!
Also, the SR-71 Blackbird. Unfortunately, not solarpunk 😔✊
@prahou
burning fossil fuels is solar punk (you are using billion year old solar energy)
dyson spheres are not solarpunk (you block out the sun and that's very Mr. Burns (Mr. Burns is not solarpunk))
@envagyokabela hmmmmmmmmmmmmm
@prahou
don't believe me? I'm an expert, I'll have you know!
Let me demonstrate:
mutts are solarpunk
purebred dogs are not solarpunk
@prahou
gonna provide a smol list of boolean inquiries
is lisp(any tbh) solarpunk?
is werc solarpunk?
is plan9/9front solarpunk?
is musl solarpunk?
is solarpunk solarpunk?
is jupp solarpunk?
is mksh solarpunk?
@quarkstuff @prahou is it coincidental that there are 7 binary inquiries, and thus all of them can be answered at once with a single ASCII character?
@prahou hello i think the clearest vision of solarpunk i have seen is the start of the novel "a half-built garden". a key element there is that the solarpunk experience for those characters is not good, at all, but it is working
@prahou i strongly object that germknödel are not solarpunk!
but solarpunk is definitely solarpunk, kinda wild that you missed that one tbh
@prahou is solarpunk even solarpunk?
@prahou god scrolling this thread makes it look like just a bunch of syllables and sounds now
Semantic Satiation is fascinating!
https://en.wikipedia.org/wiki/Semantic_satiation
(stupid meat CPUs)
The assembler on my new disk operating system is working, so once I finish the basic editor I'll be able to write the remaining programs from inside the OS.
@crc Watching you do this work is lighting up our childhood whimsy for pocket computing again.
(we are very distracted by that custom keyboard up top tho 👀)
@and that keyboard is my main one at work. It's a keyboardio atreus with linear a key caps & kailh pro purple switches. It uses a modified Dvorak layout as the base layer, and my preferred split of symbols on the left & a number pad on the right on the secondary layer.
Pocket computing is pretty nice. It brings back a lot of early PDA vibes for me, though I haven't actually written any PDA-type programs for it yet. Soon though...
what are your thoughts on using ai to discover vulnerabilities or assist with research in general?
@hi such use is still using the fashtech thingy while burning down the one planet we have and accepting thievery, slavery and abuse
train their own models ethically and run them locally or simply hire experts to do the actual work?
@hi we can refuse to interact with the slop extruder directly, and we can “not encourage people to use the fashtech”; beyond that… iff you even get the people to respond themselves… it’s up to the individual
(practice shows that the noise to signal ratio makes it not worth it anyway)
@hi @mirabilos based on recent vulns hyped up by ai companies that ended up being just marginally more than a nothingburger... I think we can safely ignore most of them just fine.
I wonder if the current AI craze is just a more graspable version of capitalism: everything (deviant) is made into a less edgy, more reproduce-able form; flooded onto the market; relies on invisible chains of exploitation; wealth accumulation is obviously unevenly distributed. Everyone is supposed to like it but most are actually just exhausted and tired.
🐡 Hello server30! A brand new machine just joined the fleet.
Dell PowerEdge R630 with 2× Xeon E5-2690 v4, a comfy 256G of RAM, and a PERC H730 Mini herding 2x146GB 15K SAS + 2x900GB 10K SAS disks.
That is 30 servers, actually 32. :)
Thanx for all the support!
Grab your VM here https://openbsd.amsterdam 🎉
@OpenBSDAms brand new a decade ago maybe..
@Nux maybe... and it still works!
@prahou love it!
@sizeofvoid thanks! (it's cropped from this whole other thing! https://merveilles.town/@prahou/117474180887027517 )
I have a tiny home datacenter that’s free to run because of solar power (net metering). When I move to a new house, I want to increase the amount of solar and the amount of computering I’m able to do for free. The part that’s most in need of an update is power. I have whole-house batteries that take over during a power outage, but my server doesn’t survive the switch. It’s like 100ms. I have an expensive UPS to handle this, but I think I need a capacitor-driven PDU. That seems to not exist.
Financial summary of our #Git hosting service based on #OpenBSD and #GameOfTrees in Q3/2026:
Our financial situation is profitable. Our expenses with 3 production servers (web, mail, repositories) are now fully covered by our subscriber base. Our one and only repository server is almost booked out.
There is not enough money left over to pay anyone for their time working on the project so we remain in full volunteer mode. We will keep a buffer to prepare for renting additional servers once this becomes necessary.
Expenses:
Total hosting costs in Q3 2026: 781€
Payment processor fees: 52.66€
EU VAT: 147.11€
Amount total: 980.77€
Funds received from subscribers in Q3 2026:
EU amount total: 701.00€
non-EU amount: 454.00€
Amount total: 1155.00€
Funds received on #OpenCollective in Q3 2026:
EU amount total: 538.65€
non-EU amount: 171.50€
Amount total: 710.15€
Procession of Fools \\ HAPPY BIRTHDAY, #OPENBSD!
#unix_surrealism #technomage #comic #analognowhere #phone_bad_book_good #mobilek #ai
@prahou
#welcometoelderfield vibes all over the place.
Good.
@holedigger17665 i don't know what that is!
@prahou
A recent videogame. One of VERY few videogames of the decade which may and can wear "LOVECRAFTIAN" on its sleeve rightfully. Its art style have quite something in common with this strip you've posted.
# sysupgrade
# syspatch
# sysmerge
# pkg_add -ua
# reboot
#
my favorite features of 2.96:
Because Ventoy often sucks when it comes to non-Linux ISO files, I decided to go the good ol’ way with colored USB sticks.
It took a bit of time, but I found the proper color range 
and more.
@joel Ventoy lost me after the xz incident, and years later, the binary blobs and lack of reproducible builds are still an issue 🤷♂️
@ricardo I’m still using it. But I really started to be annoyed to not be able to properly boot (Free|Net)BSD with it.
I've been thinking more about #openbsd the last couple of days, I've not used it since 3.9 and back then it was the core of a security product. It was perfect for that. Does anyone use it these days?
@Bredroll yes. So long as you can live with a limited selection of packages/ports #OpenBSD is actually one of the easiest of the BSD's to get setup on a laptop. I can, and do, operate happily with it. I also have some stuff that I need available on the public Internet running on @OpenBSDAms - a wonderful hosting service that supports the OpenBSD project too!
@hi props goes to LLVM22
At least it's not "This is where the OpenBSD LLM model data is…" 😆
@hi I seem to recall that there was also some new dæmon added to the base install too, so that might bump up the install-size. But browsing through https://www.openbsd.org/plus80.html and https://www.openbsd.org/80.html I don't see it mentioned, nor does hunting through @brynet's post history of #80HYPE turn up whatever I'm (mis?)remembering.
Ah, found it: job@ added rtrd(8)
https://bsd.network/@brynet/117281727508782334
(I just hadn't gone far enough back in @brynet's feed 😆)
@hi what is WAF for this mouse/keyboard multimedia installation? In the times I did this (~2003~2006), that implied only me launched the multimedia sessions :-/
that helps my 4 year old child learn how to run simple commands:
find grep ffplay
keeping filenames short so she can read it
@hi you weirdo prepare them for the Walking Dead world! 🙃
@hi true! So sad society doesn’t make use of this talent for a sane future :-/
but sadly, that's not really possible. we depend too much on the outside world for things like education, jobs, social life, groceries, and so on
Been trying to organise my thoughts on genAI for a blog post for a couple of days now and just end up getting angry at myself each time 🫠
how many times in your life have you dd'd the wrong hard drive?
| never: | 291 |
| just once: | 123 |
| 2-3 times: | 71 |
| 5 times or more: | 21 |
@prahou did you just experience a disk destroyer incident?
@j-g00da no! but I thought I did for a second, so it made me reflect.
it's only ever happened to me once many years ago, when I overwrote my computo with a lunic installation
@prahou After the first time, I learned to type echo dd ...
and think about it before I removed the echo.
Not a lot more often after that.
@prahou If i recall 3 times, one was really painfull bc I ddestroyed about 2 weeks worth of data processing and analysis for my masters thesis :)
@prahou I was a professional sysadmin with many years' experience, super great at my jobs, so of course I don't understand why the top choice only starts at five 
Are we talking per week or something?
@prahou Never, at the cost of doubting and second-guessing myself and triple-checking the commands. An acceptable price, all things considered.
Still managed to do a `rm -r` instead of `mv -r` this one time, though.
@prahou literally one time, two days ago, accidentally wrote a raspberry pi image to a desktop partition and discovered my mortality
As @brynet has been pointing out... it's time for some #80HYPE 🧙🏻♂️ 🐡
Some of the highlights we are looking forward to! This is just a teeny tiny subset of all the amazing things coming in 8.0! https://www.openbsd.org/80.html
OpenBSD vmm/vmd(8)
- Added MP support for up to 64 VCPUs per guest VM
This is what a dmesg looks like:
https://dmesgd.nycbug.org/dmesgd?do=view&id=9166
OpenBSD httpd(8)
- Setting custom HTTP response headers.
- New blocking/filtering rules based on received HTTP headers (e.g: User-Agent: )
OpenBSD relayd(8)
- Added Lua patterns(7) support for relayd(8) filter rules; improved glob(7) documentation.
- Added ECDSA support to the relayd(8) CA privsep engine.
- Switched the default TLS cipher in relayd.conf(5) set from 'HIGH:!aNULL' to 'secure.'
Thank you @mlarkin, dv@ and @sizeofvoid for these specifically, and of course all developers who are working tirelessly on #OpenBSD
How do you pronounce Pluggable Authentication Modules?
| As an initialism: "P.A.M.": | 3 |
| As an acronym: "Pam": | 135 |
Closed
I can't say that it ever even *crossed my mind* to pronounce it "P.A.M." ("pee eh em" just sounds weird) It's always been "Pam".
@gumnos Thanks, that's what I thought too! Also fairly sure that's what the people at Sun said when I watched videos back in the day.
But I've also been on several calls now where both were used, so not sure if we're now in a GIF/GIF situation.
there are those who pronounce it "Pam" and those who are wrong 😆
(I'm now very curious about the demographics of those on your calls who pronounced it "P.A.M.")
@gumnos P.A.M. seems to be pronounced more among younger engineers, and those from East Asia, at least in my experience.
But also not universal, an old salt (his words) said it as such. But then, was he saying that because he was talking with younger people? Not sure.
@rubenerd No, no, no!
According to Gartner, PAM stands for "Privileged Account Management"
(the confusion is hilarious when you're implementing a pam module for PAM)
Dustings: Let the sun melt it off.
Snow storm: Big pole with a brush on the end.
Ice storm: Well shit...
@hi I'll have to know the answer to this question the coming winter, too...
(OTOH, I did sleep in an igloo a couple of times, and it's remarkably well-lit by the sun inside, so you'd still get some % of power even when covered)
for better or worse winters getting warmer here and snow is not staying for long
@hi
I try to trigger a roof avalanche by throwing snowballs at it. Hey, it did work (twice, partially)! It was fun, but winter production is just so "meh", even with clean panels. :-/
@hi If they are on a steep enough angle, the snow slides off as soon as any light filters through the snow and starts generating power.
The other alternative, which we use on our non-roof mounted, is to change the panels to vertical so the snow never sits on them. The reflection of light off the snow makes up for the miniscule loss of power from not being more directly aimed at the sun, plus it eliminates any and all issues with clearing them off.
september was the first month when i got an electric bill since i installed battery in june.
money wise the balance is still positive, but winter is coming...
but at least it's possible to go #offgrid if i really cut down consumption (survival mode)
Huge Font/Small Screen Challenge:
Open a terminal with a ridiculous font-size that results in a tiny number of columns/rows, ideally 40x25 or less, e.g.
$ xterm -fa 'xft:JetBrains Mono' -fs 42
(maximize the window)
$ echo $COLUMNS $LINES
40 10
and attempt to use your preferred $EDITOR/$VISUAL.
How readily can you edit text?
I was given one of these little Xte Ink readers a few months back. I put some books on it and carried it around, and I tried to like it, the way I tried to like my Pinebook. I have something against electronics rotting in drawers, so I kept trying, despite the feeling that the people behind it did not care for it one bit.
Unlike the Pinebook, it's well built; it only needed a not-terrible operating system. I've jumped through these hoops before, and it's no different from working on any other embedded platform, except that this time I only had to write the thinnest Varvara emulator I could get away with, and only ever flash the device for critical changes.
Since the firmware is just a ROM, I wrote and debugged it all in Uxn11. The device emulator only had to pass the tests, and once it did, whatever I ran on my laptop ran on the device. The result is a document viewer that works equally well on everything with an emulator!
I'm not sure exactly where this device fits in my life right now, but it will help with one thing: bringing tickets and QR codes to events, which is always a headache for me without a smartphone or a printer. I might as well have something to read on the way there.
@neauoire there's also already an alternative firmware for it. Crosspoint?
@neauoire I just got one and been experimenting with available custom firmware, seems there's quite a community around it.
What format or conversion do you use for QR codes and tickets? I see that PDF is not well-supported given the hardware. BMP images perhaps?
@rembrandx All my images are icn files, so that's what I use.
https://wiki.xxiivv.com/site/icn_format
Severance voice: "Your outie knows how many bytes, sectors, tracks, and cylinders his hard drive has".
some days it feels like this
@gumnos The lack of bike helmet is also telling
@gumnos I had a need to use Microsoft Word (fat application) for the first time in a long time today.
Oh My ${DEITY}!!!
The AI had gotten BAD!!!
If you highlight anything there is a balloon / hover card for “Edit with AI”.
It is super slow to respond.
It reopens the sidebar that you closed to respond to you after it eventually gets around to it.
I feel like Word from Office 97 may have been peak.
Wordpad (RTF) is about all I want.
I LONG for Lotus Notes.
—
I was documenting a 53 step process for something at work with text copies (not screen shoots / pictures) in it. You can search for and copy text out of it.
—
Yes, I did put a zero width space in the AI(s) above.
So apparently there's this new activitypub server called LittleFedi which is in many ways like GoToSocial (which is what I'm running), except it has first class OpenBSD support (GTS requires using an unsupported compile flag). I don't think I can just migrate in place, which is a bit annoying, but maybe in the future I'll find a way to switch over. https://littlefedi.org/
make. It is nice and tiny.@rooneymcnibnug I've tried snac before gotosocial but I could not deal with the UI and the poor compatibility with my fediverse client of choice (phanpy). I really like the idea though!
@gosha I would not choose something that is 99% chance slop ("little computer people", really?) over gotosocial, but ymmv.
@sigrid Hmm, what makes you think it's slop? The FAQ says the author doesn't use AI (https://littlefedi.org/faq.html), and the vulnerability reporting timeline (https://littlefedi.org/security.html) is long enough to suggest (to my naive view) that no LLMs are used there either.
@gosha because I do not believe it not for one second. Took month(s) just to release the source code, yet no VCS, just a tarball? I only see a single person related do the project, so I am only left to assume "little computer people" are agents, and the rest are excuses to avoid everyone finding out the king is naked.
@sigrid @gosha I put the whole "Why this exists" blurb into pangram.com and it reports 100% AI-generated text. I did this with a few of the docs sections as well and it came up the same.
Reading through the rest of the website (ignoring that the style is very generic and looks like it could have been AI-generated), all this other text also feels AI-generated. Like: "littlefedi init asks the questions that are painful to answer later:" and "The config file is the seed and the console is the override"... way too dramatic and punchy compared to normal docs written by a human.
@alderwick @sigrid @jbauer @gosha yeah i had the same though (i even helped test it). bummer.
@d6 @alderwick @sigrid @jbauer Oh well that's annoying, but I guess it saves me worrying about it :D Thanks for the sleuthing work, team!
How to use dithered brushes (screentones) in GIMP (and have it behave like it would in traditional monochrome drawing programs)
@lanodan i don't know the proper terminology! i wanted to be able to do this and couldn't find any how-to's, so i figured it out myself. maybe there's a better way.
Wallpaper support because of course.
@neauoire Serif font plus that specific wallpaper makes it look like a fancy magazine
@lains I just need to make a folder called VOGUE
@magnetic_tape @lains everything I do is voguecoding
https://wiki.xxiivv.com/site/concatenative.html
@neauoire i already ported uxn to my kobo, ill try tk add your firmware as one of the launchers ^__^
One of my favourite thing about uxn for embeded development is that I have a universal emulator. While there are no xte-ink emulators out there, I still get to prototype entirely without having to constantly upload new builds on the device for testing.
@neauoire
Do you keep (or intend to publish) a devlog about this port? I'd be curious to know how you manage to do your port, the steps, and which, if any, hurdles you had to overcome.
@magnetic_tape I haven't but I'll try to write about it afterward, it's extremely aphazardly, I often go backward 10 steps X)
https://git.sr.ht/~rabbits/uxn4
Wrote a little image viewer in the firmware so that I can bring QR codes at venues(I don't have a smartphone).
https://git.sr.ht/~rabbits/uxn4/tree/main/item/src/firmware.tal
@neauoire this is so cool!
are the books going to be simply text files or do you plan to add some kind of epub viewer?
@neauoire Is there any way to read .epubs on uxn? Your port looks interesting, though my XTeink is serving as my primary e-reader.
Starting on the actual book page rendering today, at last. Now that I've done the easier tasks for the firmware.
@neauoire is the processor fast enough to render Hershey fonts on that thing ? it could be a perfect showcase for the Hershey fonts I've been working on, since I'm drawing them at ~40px cap height, so similar to uf5 fonts
@neauoire Going to make an ePub reader out of it?
@liebach not epub but yeah, just something I can read text files and draw images that I can carry with me.
We've got a text reader!
https://git.sr.ht/~rabbits/uxn4/tree/main/item/src/firmware.tal
#uxn
@neauoire seeing this makes me realise my Light Phone 1 might make a candidate for UXN. I could make a handy little manual library or something out of it.
@gahlord it would yeah! I've never seen one of those in person before but if anything it's like the xte, only with more features.
@neauoire @gahlord It’s unfortunately a quite bad Android phone underneath. Maybe something like PostmarketOS could run on it though..
There is something about the screen driver that prevents touch inputs during the slow screen refreshes and that makes it unbearable to text on. So many flaws that it becomes e-waste :/
@neauoire are text files limited to 64kb? Or is there streamed "chunking". Epub would be neat but the spec makes head spin
📵 I am interested in the "no smartphone" part.
I would love to have a life without a smartphone, but I think it's not possible, if I want to make money with the stuff I like to do and big-tech social media is the only way to have this life and for this you always need a camera or the smartphone.
May I ask how you earn money and how you stay in contact with people?
I boycott Whatsapp and it's already hard to get people to Signal. Can't imagine how to stay in contact with people without a smartphone, because they refuse to use XMPP.
Thanks a lot 🙂
@utopify_org I live almost entirely from donations, but I don't think there's anything I do that would be applicable or replicable by others. Even if I explained in detail how I do it, it would be of no use.
📢 syspatch time!
Log message:
Release nsd unbound libressl procargs sysvipc iked wgbind sysvsem uidrange errata.
https://www.openbsd.org/errata78.html
It's typical of #conlang projects to have a translation of the Lord's Prayer as a showcase template of the language. But I didn't want that, for some reason, I had in my head to do this one instead.
I ran photoshop on a £0.60 Raspberry pi pico processor emulating a 1980s Apple Mac. It inspired me to think about simple computers.
I see folks building bedrock-type abstractions more and more these days and adding on @tbsp's list, I tried to gather every single virtual machine designed explicitly to fend of bitrot or digital obsolesce by targeting a frozen specification I could think of.
I was wondering, I'm sure I must have forgotten some, if you can think of one that is not on the list, could you please let me know!
https://wiki.xxiivv.com/site/virtual_machines.html
@neauoire @tbsp
Dunno if this counts but I thought you might find it interesting. https://gitlab.com/mschwartz/inspiration
Making games and throwing them back through the time machine.
Uxn68 emulator by @d6.
https://phial.org/d6/uxn68k/
One of the first game we ever published for #Uxn was Donsol, five years ago, to celebrate, we've made a whole new version of the game with new monster sprites, better resolution and animations.
We had originally made it with the Playdate in mind, but they didn't want it so we're making available to play it directly in the itchio page, or via your favourite Uxn emulator!
https://hundredrabbits.itch.io/donsol
We hope that you enjoy it.
Have a good autumn. :)
Rek & Dev
roi on solar panels is not that impressive if you have grid connection, but it still saves us a lot especially in cold sunny springs and autumns
for being able to go fully off-grid from march till november solar is my favorite
@hi in my opinion some aspects of fp are useful when writing C (consts are great, pure functions, functional pointers), but for a full fp experience you need tco, garbage collection (and it means no hardened OpenBSD malloc for you). Libraries-wise Glib will give you some useful structures using native malloc, but bdwgc for example is not going to use it, and because it's conservative, performance will be abysmal.
Why use C? OCaml is right there, works great on OpenBSD
i'd like to keep things as simple as possible, and c works out of the box in #opensbd. i'll try sticking with native malloc and see how far i can get
@hi well, fp simplicity for the developer comes at a cost for underlying platform, but I get the idea. sys/queue.h and sys/tree.h have some structures you might find useful.
It might depend on how you go about it and what you expect of it. Basic anonymous-functions might be possible (depending on the version of C), but closures (and managing their storage lifetime) might be more challenging.
map: doable with macros (the sys/queue.h header has good examples of things like the *_FOREACH macros). They tend to be macros because the exact method of iteration would differ depending on the data-structure over which you're iterating (iterating over a single-linked-list is different from iterating breadth-first over a tree is different from iterating depth-first over a tree, all of which are different from iterating over a trie or btree)
reduce: I'd have to think a bit harder on that, but I suspect it would involve similar FOREACH-style macros
immutability/pure functions: using "const" on appropriate functions/parameters and ensuring that you don't modify parameters. There might be some complexity if you're trying to support older C standards (IIRC, the ability to return non-simple types from functions was added later), so they might end up being semi-pure functions where a parameter is used to designate the function-output location for more complex data-types like structs/unions
recursion: has always been there in C 🙂
Feeling sad about AI
https://video.infosec.exchange/w/rYgQFwrqs9mpWydyoUCakY
I talk about why I've been feeling sad and encourage myself and you to feel some hope.
Transcript: https://artificialworlds.net/blog/2026/09/11/feeling-sad-about-ai/
Fixed a bug in the notification page that made snac hang forever while trying to read abnormally big files.
Improved support for text-only web browsers: it's now possible to configure a set of web browser user-agent strings that will receive simpler HTML in the private timeline web UI. Basically, it consists in avoiding details / summary HTML tags as much as possible.
Added some fixes to media proxy code.
Fixed EmojiReact code to allow any emoticon defined in emojis.json, not only those with colon-wrapped identifiers.
Fixed a bug in notification filtering (paging was sometimes incorrect).
Added a notification filter for Webmentions.
Mastodon API: Don't return reactions count as string (contributed by mkljczk), implemented GET /v1/media (contributed by clairemont).
Fixed bug in documentation examples (contributed by Sprite_tm).
Updated Ukrainian and Russian translations (contributed by wincentbalin and koru).
If you find #snac useful, please consider buying grunfink a coffee or contributing via LiberaPay.
Remember when 📱 were invented and didn't have much RAM so they ran one program at a time, and the programs were designed so when you switched away they'd remember what they were doing and resume it when you came back? And then the 📱 got a LOT of RAM so now multiple programs can run until RAM ran out and one got ejected? And then the programs started taking up so much RAM only one can run at a time? And then it turned out the programs no longer remember what they were doing when they get ejected?
OK, our little computer/server/NAS is online at Crowd Supply: https://www.crowdsupply.com/mnt-research/mnt-station
The main idea was to have a high-end enclosure for the classic MNT Reform mainboard to give it a second life, but you can as well build an open hardware computer from the parts.
We're also publishing 3D printable versions of the cases, and there will be more photos and use case details in upcoming campaign posts.
The Chat // Have a nice chat unencumbered by the meaning of real words. Made for #picojam26 https://itch.io/jam/picojam26/entries (thanks for organising, @illuminesce !)
Play online: https://www.lexaloffle.com/bbs/?pid=194820#p
Or from #picotron terminal: > bbs://thechat.p64
When your compiler knows what to compile without giving it arguments, and when the operating system knows what artifact to run without having to tell it. 
Don’t move buttons from under the mouse.
Thanks @The2lb3oz4dr10grOfHedgehogs for the video
Reading @j-g00da's comics via @jn's zine viewer on Uxn. Fits the medium like a glove.
https://codeberg.org/neuschaefer/dreams/releases
https://kamiokan.de/
Very cool!
An NES homebrew ROM showing live ADSB aircraft traffic, fed over serial through controller port 2 on real NES hardware. A companion app does the networking the NES can't, then streams it to the NES.
I am seriously considering switching back from Neovim to nvi. I constantly vacillate between the richness and bloat of Neovim and the scarcity and minimalism of nvi. Sure, Neovim does have some nice features. But is it worth this huge increase in complexity? The size of Neovim's codebase is probably like 20 times that of nvi's, while its benefits seem rather marginal to me (based on my way of using it - your mileage may vary).
We always add, add, add instead of subtract, subtract, subtract...
@thorstenzoeller While coworkers and I were debating and comparing IDEs including Vim and Emacs, one of the smartest people I ever worked with was just quietly coding using, I think, TextPad. He said it has the only two features a programmer needs: Syntax highlighting, and the ability to run an arbitrary command line, and optionally parse the output to create a pane of results as though they were compile errors, with messages and (filename; line numbers), which can be clicked on to view the line
On Friday, my son asked me why C++ is better if the Python version of the code is simpler. He specifically had asked about a "hello world" example, so we wrote both examples together. They both run instantly, so it's not a great comparison. I showed him how to tell which is faster with the `time` command. We also talked about the difference between the work that Python and C++ are doing. I'm pretty sure he understands what we talked about, but who knows?
█▓▒░ 𝗣𝗜𝗖𝗢⬩𝟭𝗞 𝗝𝗔𝗠 𝟮𝟬𝟮𝟲 ░▒▓█
🌐 https://itch.io/jam/pico-1k-2026
WHAT?
👨💻 Make #Pico8 (or #Picotron) Games/Demos/Tools in ≤ 1KB
WHEN?
📅 SEP 1ˢᵗ-30ᵗʰ 2026
Hope to see you at #Pico1k jam! 🤓
Our new Amsterdam servers are live. The EU can now be the primary home for Fastmail data, and like our US locations, it's our own hardware, racked and run by our own people. Nothing rented from a big cloud provider.
Already a customer? Switch regions in your settings. Thinking about it? You choose your region at signup.
Our CEO Bron's post walks through exactly what lives where, because we figure you'd read the fine print anyway.
https://www.fastmail.com/blog/fastmail-offers-eu-data-region
basically, my meter shows the net value across all phases, while the electric company counts import and export separately and charges for both. so instead of having almost zero import, the actual import was much higher...
trying to configure my inverter to balance the phases better
exported over one mwh at peak hours. can't wait for the next electricity bill... :)
More surreal rendering tests for #picotron 0.3.1, this time with tline3d textures x colour tables. Notes and gif:
https://www.lexaloffle.com/bbs/?pid=193908#p
like = you run it on server
boost = you run it on desktop/laptop :)
❤️
reply = not subject to artificial pleas to boost or like 😆
(but yes, several OpenBSD server instances and two laptop instances)
I'm using OpenBSD on little servers only for a private case not on a public way, and less on a professional way ( I don't work in IT I'm just curious about it ).
One machine is a little NAS and another is a homelab hosting actually a snac instance and a postgresql service.
Also my 20 years old computer, I'm actually use, works well on desktop with OpenBSD thanks to dwm :)
"coding is solved" they said
"use AI" they said
me working on OpenBSD writing C: dead software walking
https://rsadowski.de/posts/2026/dead-software-walking-relayd-and-httpd/
Checking that quad boundaries join up with no holes or overlaps. This one kind of reminds me of a colour plate from an 80s siggraph paper. #picotron
Apple keeps releasing new and more expensive hardware that folks want us to support but we can't even afford to buy the new stuff ourselves to test it on.
I mean.. my iPad is an M1. My iPhone is still a 14. And my Mac is an M1 Max. So... I'm a few years behind over here across the board.
Now there's talk of a foldable iPhone? I'll probably need a second mortgage for it which wouldn't be a good idea because I don't know if selling apps will even survive another year.
The math isn't mathing.
I know I shouldn't be surprised that it just works as is, but still cool, I can build compile uxn binaries directly inside dos.
wcl -bt=dos -ml -fe=bin/uxnmol.exe src/uxnmol.c
📢 EuroBSDCon2026: The Schedule is Here! 😈⛳️🐡
The schedule is now published!!
You can find it at: https://events.eurobsdcon.org/2026/schedule/
Grab your tickets 🎟️ at https://tickets.eurobsdcon.org/eurobsdcon/brussels/
For everything else, peek at https://2026.eurobsdcon.org/
More information is added all the time.
EuroBSDCon 2026 in Brussels, Belgium 🇧🇪
September 09-13, 2026
#RUNBSD #FreeBSD #NetBSD #OpenBSD #EuroBSDCon #EuroBSDCon2026 #BSD #Conference #Register
cons
having all my tools within arm's reach has been an amazing boost to my productivity. more importantly, it brings me peace of mind because i actually know where my things are now... ❤️
i think i own too many things, but i guess that's the price you pay for living in the middle of nowhere
i'm finally getting all my boxes sorted and everything organized for the first time since i moved in last year
For photos, a combination of nnn and nsxiv's thumbnail feature.
For video, mpv and the shell like some sort of animal.
@hi feh and mpv for simple stupid viewing. thunar, bash+awk+sed, geeqie when mass renaming is required. digikam to store, tag and forget until next couple of years. gotosocial to expose some to the world.
maybe someday, I’ll add immich for private/family memories.
@hi There is a docker packaged iCloud backup app that does that the other way around. I use it to sync all my iPhone photos on my NAS. Useful? https://github.com/steilerDev/icloud-photos-sync worked great to grab 13k pics from my Apple account.
@hi I use PhotoSync. You choose from WebDAV, SFTP etc…
@hi PhotoSync? It supports WebDav and SFTP amongst others. https://apps.apple.com/gb/app/photosync-transfer-photos/id415850124
@hi syncthing? I don't have an iphone, but I find this https://apps.apple.com/us/app/synctrain/id6553985316
@hi radon is something that just comes from the ground, not related to the type of the car, and it needs to be evacuated through the regular ventilation. I've learned (not too late thankfully) that builders even in EU countries have less strict requirements for garages, so for the last 10 years every time I need to deal with garages or any other ground level places long term - I bring a radon/voc sensor with me.
Everything I Know About Making Art On The Computer I Learned From @prahou.
https://triapul.cz/automa/i_did_not_kill_stanley_lieber
https://rabbits.srht.site/decadv/
#occ2026
Five years ago(as of June 8th), I decided to give up on Acme and make my own text-editor. The plan was to make something I'd use for the rest of my life, improve it as I went along. It was to be written in this new language, and my list of ideas for it included a ton of extremely bizarre UX choices I hadn't really tested for my own. But if we ignore the more quirky esoteric things about it, there was two pretty simple choices made for it:
Undo depth: zero.
Largest file: 64kb minus the editor's own size.
Undo is probably the easiest one to understand. As you'd expect, some other mechanism emerged to take its place, I wouldn't just let it eat up my work each time I made a mistake. What ended up replacing it was being able to append and cut off chunks of snarf buffers, when I have something I might want to keep, I push it the top of the snarf file. When I'm satisfied with the change, I pop it off, or add something else on top. This has served me well enough, I figured that if I ever needed undo I might add it, but the occasion never came!
As of today, Left sits at 17677 bytes, which means I can edit a file, at most, a 45kb long file. This turned out to be less of an issue than I expected, mostly it made it so I split projects into chunks, but more importantly, it means that every single byte of the editor has to be accounted for. If I add something, I know its cost and the cost is directly making for a worse editor. This has been a constant force of optimization and code hygiene over the years.
A new MNT Pocket Reform mini laptop variant with Quasar 6490 (Qualcomm QCS6490) processor module (8GB RAM / 128GB UFS) preinstalled is now available in our shop: https://shop.mntre.com/products/mnt-pocket-reform
Enjoy!
just took some pictures of the new MNT Pocket Keyboard (small standalone USB-C keyboard with integrated optical trackball and RGB backlight) that's coming to our shop at some point next week. the case is a mix of aluminum (top plate) and Multi Jet Fusion printed Nylon (HP PA 12 S). in my opinion it feels really nice in the hands. open source hardware, of course.
need to find something or write a little wrapper for ffplay
@hi for OpenBSD on protectli you should be fine with a connection to the serial port. If you want a simple to use KVM-over-IP solution, I would highly recommend PiKVM. You might even be able to attach an USB-2-serial adapter to your PiKVM system.
@hi I used to use this https://www.tumfatig.net/2022/pikvm-using-a-raspberry-pi-zero-2w/
It was quite nice. I also used it to record the "OpenBSD workstation" :)
current setup:
openbsd -- usb dac -- amp -- speakersnext up: make everything more user-friendly or teach my family how to use the command line ;)
openbsd -- hdmi tv
then i discovered that openbsd doesn't support audio over hdmi and my openbsd server doesn't have analog audio output, obviously.
my tv had been connected to a stereo amplifier anyway, so my setup is actually even better now: a dedicated usb dac connected directly to the amplifier.
also, lots of cables :)
at least the dac box can draw power from my server over the same usb cable.
very excited to switch to fully local streaming ❤️
@hi My daily driver is similarly comfortable to me, but pretty unusable by others…lots of keyboard shortcuts, superfluous window-chrome hidden, lots of dark-mode comfort for my eyes, lots of CLI/terminal interactions 😆
I Did Not Kill Stanley Lieber: How to draw (with 9front)
The penultimate paper on paint(1) and the 9front art machine.
https://triapul.cz/automa/i_did_not_kill_stanley_lieber
#unix_surrealism #plan9 #9front #art #guide #computers #oldcomputerchallenge
We really appreciate the work done on the TLS settings in relayd(8) and httpd(8).
It makes it even easier to get a tried and tested secure connection to your favorite OS and service!
"Both relayd(8) and httpd(8) now have the "secure" list of allowed crypto methods for HTTPS, which include TLSv1.3 and the TLSv1.2 AEAD cipher suites.
The previous list was "HIGH:!aNULL" which contain non-perfect-forward-security methods and this change may cause old clients to not be able to connect."
Source: https://undeadly.org/cgi?action=article;sid=20260629165750
Hat Tip to @sizeofvoid and others!
it's easy to script everything. wrote a script to download the latest version of routeros and upgrade all my mikrotik devices: routers, modems, access points (via capsman)
and when you need gui then winbox is amazing too. nice solid ui and it's only getting better over time
@hi I'm sure the new bZ is nice, but not so much nicer that I'd consider upgrading (esp. that we got ours, an AWD with a heavy discount when the dealerships were getting rid of their old test-drive cars when the new model was coming in, essentially we got it 40% cheaper...)
now considering good old woodworking instead...
any good cnc machines that can produce small parts?
any similar machines? (also made in eu)