Sims

Not many arguments. I can’t see why Gnu Guix is discarded for Nix (similar OS management)…

Hm, I tend to like weird small distros, but I don’t think “This command-line-only OS can help you learn the CLI.” is enough to recommend it. ALL distros have a terminal, and many many of them are similar to this distro.

“Peropesis is a special kind of Linux distribution.”

No, its not, and the reviewer that have ‘reviewed lots of distros’ didn’t even explain what boot system it uses. Is it a Debian variant or ? What does it do that any other distro with a terminal doesn’t do ? (Except not being persistent/installable) Why not just spin up a normal Debian server in lxc ?

There are many other suitable terminal distros out there…

Seem to me that peeps outside of the AI development sphere/interest are not aware of how quickly ‘flaws’ gets fixed. There are still people that don’t think AI will ever be useful - or intelligent - based on some ‘archaic’ performance from many months ago. Reality will hit hard I think.

Personally, I have never seen any development move faster than artificial intelligence, and whatever it can’t do ‘properly’ today, it can do tomorrow or the day after.

Current AI/Agentic status is the clawd family of frameworks + a sota model. However, they are really stupid architectures (Every 30 minutes, the llm is yanked back and presented with the original tasks in an md file - that’s it) and are WAY behind what we can do according to papers/newest development. Papers quickly trickles down to architectures tho, and the next family of agentic frameworks will strike as fast as the clawd phenomenon.

We are not far from general AI - not particularly from llms/transformers, but from the external cognitive ‘harness’ that are build all over. While the harness adds cognitive states to the architecture, many of the typical agentic features are being build into the model itself, so the the cognitive functionality of the harness, are being injected into the models, and the new harness fixes other ‘flaws’. We will see one clawd moment after another, faster and faster, getting better and better…

I hope peeps live in a society that takes care of each other, and don’t treat each other as lazy bums that “just wouldn’t work hard enough”. It’s going to be horrible to peeps in US and similar Capitalist ‘might is right’ societies. There is NO safety net for ‘failure’ there.

Back to article: It was bound to happen within a year or so.

Capitalism makes it worth it to setup an Agent and 50 new channels on YT that does nearly the same. It doesn’t matter how many people gets spammed as long as some views are amassed. It already happens in normal biz on all levels. All AI ‘slop’ is just Capitalist slop. Really, it’s only there to make money.

I find them annoying too, but unfortunately, AI accelerates the Capitalist enshittification we see everywhere…

yes, too much ‘Elon’ vibes…

Read the article and still have zero idea why this project was established, why ‘react’ is suddenly important or what the project really do for Corps. We apparently don’t need AI to produce information slop…

Damn, I didn’t know Fedora was a geo-political tool :-(

Wow. At some point you really should use the same amount of energy actually examining what Russia is saying, and thus what is actually happening. Oh well, I guess its just easier to get a fully formed opinion served by propaganda…

hostnamectl set-hostname moon # 1. Set hostname

docker run hello-world # 2. Host on the moon

You are welcome !

You have to think in ‘information dimensionality’. A yes/no toggle is 0D, a list is 1D, a list of lists (std hierarchy) are 2D, a list of list of list are 3D etc. All information storage types are one of these dimensions. Think of a graph-base file system with nodes and edges between everything. Now, imagine a filesystem where you flick a switch and the whole structure shows another pov ? Maybe you want the whole thing to be shown as file-type hiearchy, or only parts of it. Maybe you need to show movement in the structure, so everything are in a temporal/spatial hierarchy, maybe you are only interested in dependencies ? Relations ? Other ‘weird’ metrics ? …and so on. The main problem is to manage, find and show the needed information in a higher-dimentional fs.

Technically a normal file is also a list, or another ordered structure, but in this sense, they are just a node with further dimensionality.

There’s a TON of information layers locked away in our normal filesystem hierarchy, so OP are perfectly right, and people here have no imagination or even a world model of information structures…

Not sure why people defend an archaic organization form here - reflex ?

You are perfectly right that files and folders are simplistic, and should naturally adapt to the pov that are more information rich/valuable. Hoomans tend to collapse a high-dimensional structures to 0D to 3D, so we can manage the information. In that sense, a std hierarchy is only ONE pov over a ton of pov over the same content. A standard hierarchy is only a low 2D dimension structure that are our first attempt at organizing information. It’s not wrong - just imprecise af.

Anyway, hardlinks are a small step up, can build wild static structures (like a oneshot filesystem in Guix), but is cumbersome to control in multi-dimensional information structures. Likely not what you want, but look into fuse file systems if you want to move on to a dynamic file system hierarchy. An interesting one is a tag file system. It turns a standard limited hierarchy into a much more dynamic file-structure where a file can - and does - belong to a bunch of tags - file type, size, group, comments, whatnot. There are many many fuse fs that can convert anything into a better structured file system. Tagging is a step up from a dumb 2D hierarchy, but maybe a graph file system is the ultimate freeform dynamic filesystem that can present all the pov’s we could possible need ?

Go for it.

Yeah, I lost my girlfriend, so … no judging please !

Great that you chose weights for the final evaluation. Black, and whitelists are just too …0D/binary for any advanced use, and are unfortunately used everywhere.

The title doesn’t sound bad at all - unless one irrationally hates AI. The fact is that the Distro is bad, and it have nothing to do with ‘AI’ integration, or the attempt to do so. Hate the Human behind the release, or the Capitalist religion that enforces ‘Slop’ on all scales - but hating everything that have ‘AI’ in it - is just f’ing moronic…

I have a few super old phones/tablets I didn’t threw out for some reason. IF termux can be installed, does the Android version mean anything for the running services ? My latest Android device is version 7.

Agree. “Are you perhaps too angry?”, “The HELL I’m not! F*ck off !”, “alrighty then”…

“\C-x(”: “s/\([[:space:]]\)\(\(\([[({]\(\([[({]\(\([[({]\([^{][(){}]\{0,1\}\)*[]})]\)\{0,1\}[}][(){}]\{0,1\}\)[]})]\)\{0,1\}[^{][(){}]\{0,1\}\)*[]})]\)\{0,1\}[}][(){}]\{0,1\}\)*\)/\1(\2)/n”

YES! Finally something with a manly learning curve ! Enough with all this baby high level programming languages and UI’s. ALL computer interaction should preferably be in a random mix of brainfuck and ed(2) ! Maybe with a little regex expressions randomly inserted here and there to keep us sharp ;-)

Compute prices are not going down. If you want a budget friendly compute, then you may need to look for NPU’s. They are slower, and is unfortunately also getting more expensive pr TOPS. Electricity prices in the West are going up as a low output gets bought up by datacenters instead. The newest GPU’s are more efficient in re. to power pr Top, but they are not budget friendly.

I have not seen a recent comparison of GPU vs Npu; TOPS vs price vs power consumption, but @ aliexpress, I saw a small 40 TOPS Npu as a nvme stick with 16gb ram that draws 10 watts, or so (search for ‘ai accelerator nvme’). This little thing can scan files 24/7 and your i5 can help out in peaks. Afaik you can run a distributed model that also runs in, and uses your i5 memory/compute, so if you max out i5 memory, perhaps the combined compute is enough for a larger model ? Maybe a few Npu sticks can work together on the same model ?

Alternatively, you could look for the next version of a Huawei gpu card (first had baby kinks afaik), or one of the other upcoming producers from China. They’ll come faster and faster, but are earmarked for local consumption first.

Another suggestion is to buy one of the old P40/80 (from the ‘pascal’ chipset, I think) (or was it k40/80 ??). They should still support a descent range of modern quantization needs and often have 24gb ram. A refurbished miner card cost around 50-70$+ - cant remember exactly tho.

Lastly, you could try something different. If the i5, with enough memory, can run a large model slowly, you could add a dedicated KV cache, so most of the tokens won’t need to be recalculated. Memory and bandwidth are the most important here, but any old server could be upgraded to be a dedicated KV cache server (might need a network upgrade tho!).

Anyway, Ideas are easy and cheap. If I were you, I would ask an AI for a little python app where you can add a product and it returns a graph where it compares with other products and show optimality over time given prices/power and Tops - Gpu vs Npu. Good hunting…

Neat. I had to think about the purpose of running a kernel in a browser, and couldn’t find any at first, but if we just see wasm as a browserless container environment, then it makes more sense than running a linux kernel in a browser. I could see a minimal linux spinning up specialized kernels in different wasm containers, much like vm’s. Not sure what hardware/software roadblocks exist in a modern wasm container environment tho.

MSP’s ??