yesterday at 5:25 PM
For the last five months or so, opus and fable have both been able to do this task as a soneshot prompt. An inference engine is almost pure compute and to a first approximation the only operation your hardware needs to support to do llm inference is multiplication [the rest of an inference engine is "some trig"].
So to get any model working on $OLD_HARDWARE, the hardest part of the task is now sourcing the hardware itself and giving the agent ssh access -- a modern agent will be able to source a compiler, write the implementation, optimize it, test in qemu (or any other simulator) and push it onto the device.
Example prompt: "write an inference engine for qwen3 0.6b in pure assembly that runs on this device. Write a script to attach the weights to the resulting binary. Prefix is supplied as a stream of u32 token IDs on stdin. The engine streams u32 token IDs on stdout."
yesterday at 5:52 PM
I agree with you, and I'm using claude and other models heavily, the idea is not in the soneshot - because actually as you see in the logs I posted on github, it took around 107 + 50-ish experiments to figure out which models to run which and push the speed.
I'm treating it as a search algorithm project, where I develop the methdology and search for pushing the limit of devices and searching in a big space of possible kernels and so for what works for a specific hardware.
so I'm looking at it from a different prespective, it takes a lot of brute forcing, and claude can do this well, the whole game (to me) is which directions you push the model into to reduce the search space. and how that learning from very small and tight devices you can scale these techniques.
But in essence, I don't disagree with your sentiment - the goal is not 0.6B models, these are extremely trivial as you mentioned and as I showed in the experiments, the goal is 16B+ models, and this as far as I've seen, requires smart methods to search for optimizations and customizations both to the model and inference engines.
yesterday at 6:08 PM
Why did you both say "soneshot"? Same typo of oneshot?
yesterday at 11:25 PM
Soneshot is a much better technique for such and such type of projects
yesterday at 6:22 PM
I just asked myself the same thing and wondered if this was some new technique that completely slipped past me. - But it certainly must be a double typo
yesterday at 6:27 PM
orrr we can listen to paranoia that these 2 comments are just an elaborate advertisement
yesterday at 6:18 PM
What is “soneshot”?
yesterday at 8:56 PM
I think it's a typo of "oneshot", referring to a single prompt being able to handle the entire task.
yesterday at 10:21 PM
Yes that was my typo. Sorry about that.
yesterday at 5:56 PM
So true, people think they can prompt AI, tell it to create something amazing and "Novel" so they can feel cool and bragg about it. Shame on such people.
yesterday at 9:27 PM
It's okay to share fun thoughts even if you haven't put blood sweat and tears into them.
yesterday at 5:57 PM
no one said it's amazing or novel, it's just a fun experiement I'm not selling this :) just out of curiousity to explore.
yesterday at 6:04 PM
yesterday at 8:44 PM
lol sure
"claudia slop me up some 600m, commit, push, make 500 pages of slopped README.md, github stars+++"
great, you should definitely do this, there will be absolutely no problems whatsoever and it'll also run at peak speeds while of course being completely unusable because who wants to run a slopped 600m
t. at 1200+ commits on my "oneshot"
yesterday at 9:43 PM
No one will run a slopped 600M or even 16B for anything serious today, the whole goal is to experiment, to push both limits of what can be done with the model, and the hardware itself.
It's fun times actually. And I think such experiments and others posted are both showcasing the limits people are pushing.
yesterday at 3:23 PM
Recently I started this open source project at my AI company, where I'm trying to make LLMs working
Now Reached 16B models running on iPad Air 1st gen and xiaomi tv box, both have less than 1GB of usable memory, and reached some decent speeds given the hardware.
I'm open sourcing the effort done, kernels developed and results for everyone to re-produce and push more, also the modified models are on hugging face : https://huggingface.co/hyperionaidev
I want a more challenging device, what do you think?
yesterday at 4:46 PM
> a more challenging device
A Playdate
yesterday at 5:41 PM
Aw man I would love to be able to use the crank to back up the agent dialogue, maybe make it branch forward in a different direction. Make the dialogue into a game
yesterday at 8:38 PM
I made myself Remote Control Deluxe(tm). Which is essentially a remote control connected to my Claude and VLC.
I hold A to speak. B to cancel. Turn the crank really fast and really slow → send the prompt!
I still do use it sometimes.
The app talks via serial and WiFi.
yesterday at 5:55 PM
pretty interesting, I don't own one, but it's possible probably with an emulator to figure this out, thanks!
yesterday at 6:24 PM
Yes this has been my exact thought for awhile now.
Perhaps the real value with AI is going to come from...surprise here...a whole bunch of different low resource AIs networked together living and working and yes playing thru a networked "society"
Sounds kinda familiar doesn't it?
yesterday at 9:24 PM
vigorously turn the crank to speed up the output. ha.
yesterday at 6:09 PM
A Speak and Spell that replies using the voice processor
yesterday at 7:39 PM
Is it even still around? :D would be interesting if actually accessible idk
yesterday at 4:46 PM
perhaps the pi 1st/2nd gen?
I'd thought of the ESP32 devices but their RAM is too small
yesterday at 5:07 PM
Pi 1 is too easy. My PicoLM [fork] already runs on it.
It also runs on MS-DOS 80386 2MB RAM, so...
And iPhone 1. And Digital UNIX.
yesterday at 11:19 PM
I was looking for the lowest claim in this thread, as I managed to get a nano LLM running in the ≈3mb of SRAM of an FPGA with zero DDR in the loop at 20k tok/s, but 2mb on a 80386, great job.
Have you got a write up for it?
yesterday at 10:01 PM
I have a basement full of Pentium 4-Core 2 Quad machines lying around. Has anyone strung together machines of this era to make something work? I always imagined that even if I were able to provide something like 4-8GB of ram per box the token speed would be measured in tok/hr or something.
yesterday at 5:55 PM
Raspberry Pis in general are too easy for this task they don't even need modifications, older iphones seem more interesting due to some work on comptability.
yesterday at 10:03 PM
Machines with unified ram work better? whats the bare minimum Ram you need? Maybe those old SGI or Sun machines could have a second life. SGI boxes made the Terminator come to life on the big screen so now their second act could be to help bring the real terminator to life :P
yesterday at 6:07 PM
A PlayStation 2 dev kit (with 128 MiB of RAM instead of the standard 32 MiB).
Between the VUs, the R5900's SIMD (MMI), the GS and the I/O buses (FireWire, Ethernet, IDE, Memory Cards...) to stream model weights, the smallest Bonsai models (or aggressively quantized Qwen3.5-0.8B) and some really cursed programming might just eke out something the PlayStation 2 has no business running.
yesterday at 8:55 PM
This would be fascinating. I've sometimes wondered how computing would have been different if we'd known how to build modern LLMs in the 2000s, despite the much poorer hardware (leaving aside how you could possibly train such a thing back then).
yesterday at 10:04 PM
Wonder if its possible to solder on more RAM to the PS2 like you can with the Xbox.
yesterday at 4:22 PM
> I want a more challenging device, what do you think?
One of those vapes with a screen on it.
yesterday at 4:35 PM
challenge accepted.
yesterday at 6:33 PM
literal vaporware?
yesterday at 6:10 PM
This is very cool. Love the trend of optimizing small models
I’ve been working on small local models and it’s amazing how fast the tech has progressed
A year ago, I was creating conversational characters by finetuning a qwen 2.5 0.5b param model, and it worked pretty well. But no reasoning or math capabilities, and still required some decent specs to run; could maybe do well enough in a beefy raspberry pi
But now, I have a locally trained 6.5M param model that does reasoning, solves multi-step math problems, does simple physics problem simulation, tool calling, RAG and conversational chat. Total size about 30MB packaged for ONNX, and runs with <10ms latency in the browser
yesterday at 6:57 PM
A year ago 1B parameter models couldn't do much, today with the correct harness and/or fine-tuning they can do more or less serious work (not frontier tho).
I believe in a 6-8 months it will get even better and my TV will be happy.
I didn't try fine tuning so small models honestly but looks interesitng to try.
yesterday at 5:16 PM
It's technically sweet regardless, so don't take this as even a little bit of negative pushback, but I am curious what the utility of these small models is? Are they smart enough to be useful on practical tasks? What are people doing with them?
yesterday at 5:27 PM
Not op but my opinion is that It doesn't matter if it's not currently useful. I think it's important that the community focuses on this line of research because big companies will not. It's in their best interest to make LLMs resource hungry
yesterday at 6:15 PM
Yeh; this. They see how GTA6 looks and know a hypothetical scenario like a future Nvidia 7000 series with 96GB VRAM and 512GB system RAM in everyone's home means the entire tech, media, knowledge work labor system; the rich peoples social moat; evaporates.
It's all quite clearly collusion among a few dozen, hundred. No way their education/experience or data analysis engines have not made them aware.
They’re just used to Boomer and Silent Generation politicians blithely enabling them cause line go up for them when they do!
Curious to see how well it works on Millennials and GenZ; the elders were educated before STEM education had matured in a lot of ways. Younger generations have not had debunked fantasy rammed down their throat as scientific fact; Nostradmus types of the mid-late 1900s pointing at some ignorant math assuming it's true because it's written in a scientific notation! Duh!
Science education of the past gave rise to thinkers like the Oceangate CEO. Nihilists who figure there's no such thing as forever and we all die anyway; so YOLO no matter who I take down with me.
yesterday at 6:49 PM
If you work in a highly regulated space subject to e.g. HIPAA, and you work with ancient on-prem systems within a slow moving and risk intolerant organization -- for instance, virtually every state run healthcare system in the US -- then these are all you can hope for.
In these settings, data is insanely messy beyond your wildest nightmares and hidden away in free-text fields that small models like this are great at cleaning up, normalizing, parsing, categorizing, etc
yesterday at 5:48 PM
This is an interesting question, so the idea is, based on experience, fine tuned small models can do narrow tasks very well (my experience running models fine-tuned on customer data as low as 1-2B is very promising) also the goal is to find a breakthrough were instead of the line of work (which I respect) that tries to squeeze much bigger models into smaller hardware like strata and maya, I'm working backward to discover/search for custom kernels and modifications till we exhaust the limit of what a smaller machine is capable of, so instead of people doing 64gb machines and old gpus to run a deepseek, I want to see what is the biggest usable qwen small model can run with reasonable speed on smaller devices, and by scaling as we discover more tricks, normal laptops can unlock more compute - for example a qwen 3.8 27B can do pretty serious work when coupled with a specialized harness - but we are yet to figure out what is the smallest machine that can run it with usable output and usable speed, and this can unlock interesting business.
today at 12:10 AM
I find this compelling “merely” as an engineering challenge, but I’m curious if you’ve observed an actual practical benefit?
By that I mean, have you found a case where a fine-tuned, small model is outperformed by stuffing a heavily tuned larger model in its place?
yesterday at 7:56 PM
I'm told that qwen 3.8 27B is good enough for practical programming work, so if that's doable then I could easily see the utility. Or... the biggest qwen I see here is Qwen3.5-2B; is that actually capable of programming or agentic work with code?
yesterday at 8:34 PM
It's not, but tiny models have so far been on the same track of increasing capabilities as large ones. The parameters needed to do useful work keeps going down. It hasn't reached 2B yet, but it's very possible it'll get there. In the last month we've gotten 4B models that can do small tasks.
yesterday at 5:58 PM
Speech recognition, speech generation.
This alone is quite useful in small/limited devices/contexts.
Realtime log parsing and summarization is another application I've seen for little LLM's..
Basically, anywhere you'd use an LLM, but at much, much smaller scales of ontology.
yesterday at 7:28 PM
speech models are what I'm trying next because imagine a nice portable meeting transcription device that can hold around without internet and for NDA-ed meetings and so (with permission of course)
yesterday at 5:24 PM
Also curious on this. They are too dumb to use with tools and if you ask it how to bake a pancake you only get rubbish.
yesterday at 5:53 PM
I mostly agree today, but I'm trying to push for the future where smaller models are more capabable, and I believe it will happen + the fine tuned cases I mentioned in other commments.
On other notes : there is no free lunch, you have to sacrifice something for the other
yesterday at 5:30 PM
[dead]
yesterday at 9:31 PM
The AI botnet is going to fill every crevice of civilization, do you want AI ants? This is how we get AI ants.
yesterday at 9:36 PM
Or maybe AI bees :(
yesterday at 4:47 PM
in a similar vein, somebody got tinystories LLM running on the nintendo wii, including gpu acceleration [0].
0: https://github.com/brysontang/iiama2.c
yesterday at 5:43 PM
this looks very interesting, it can't run a real (read : working on some real tasks) LLM, yet, but it's actually very interesting device.
yesterday at 6:03 PM
Here’s another repo that might be interesting
https://github.com/slvDev/esp32-ai
It shows using Per Layer Embeddings to run models that normally take 100s of MB of ram run efficiently in just a few KB of ram and keeps most of the model on disk. It’s somewhat similar to MoE so it doesn’t stream in the whole model from flash for every token, but takes it even farther. If some of these devices have 1GB of ram and a few GB of disk there might be some interesting options here.
I’ve been investigating getting something like that running on the Wii too, though I haven’t made much progress in a while. Some basic calculations says a 1B model might be possible, at least a few hundred M
yesterday at 7:27 PM
I would actually like to see the Wii one, I don't have one unfortunately so I can't directly try on it but this esp is interesting I would like to replicate and/or improve the results on it, this technique is similar to the streaming i did from flash but probably more efficient, I need to look it up and understand it more. Thanks for sharing it
yesterday at 10:09 PM
There is a lot of random e-waste being tossed. Think TV's, Wifi routers, IOT devices. I mean many of these should run right? WRT54G was an extremely common router so something like that? Or the old Laserjet printers that were everywhere.
yesterday at 6:54 PM
Thanks for sharing! I'm exploring a similar direction, but optimising for extreme model size. I have gotten an LLM down to 970 KB, with some interesting results compared to GPT-2 (small).
yesterday at 7:00 PM
Super interesting, not a CS/AI guy here, but I'm curious as to how you can train such a small-sized model give you as accurate an answer or good hypotheses to get you going
yesterday at 6:55 PM
This is extremely interesting, do you have a public repo or write up to share or not yet? if not yet please sign me up when you release it, Cheers!
yesterday at 7:01 PM
I'm planning to have a blog post up in the next two weeks or so (simedw.com), still have a lot of experiments to run, and training even tiny models takes a bunch of time.
yesterday at 7:39 PM
following!
yesterday at 4:08 PM
This is really cool man. I like this on going trend of resurrecting old tech with AI and this is yet another avenue to explore.
I'd be interested in seeing some older flag ship mobile devices
yesterday at 4:21 PM
I have a few phones and tablets laying around, will try to play with..
yesterday at 8:25 PM
I'm doing https://github.com/rcarmo/go-pherence for the same reason. We should swap notes.
yesterday at 9:55 PM
Is this the new version of "Can it run Doom"?. Maybe, "can it run Qwen?"
yesterday at 5:31 PM
What can these small models do? Are they just for novelty?
yesterday at 6:01 PM
as of now, it's just for novelty and fun, without serious fine tuning on domain data, you can't really do much with the smaller models.
yesterday at 6:07 PM
So does it make sense to try and run models on older 4gb GPUs? I assume the energy efficiency isn't worth it?
yesterday at 6:11 PM
well, the whole idea honestly is curiousity about limits of hardware more than actual application today plus the idea that in the future people can run more capable models (in terms of capability @ small parameters) on such devices.
It makes more practical sense today to just go into 4gb GPUs and many people are working on this already.. on my side of things, it's more on the exploration side for now, until we see what good models get released that can use this capability.
There are many people on github and x running decent models especally MoE models on older GPUs and it's doing nice, I'm personally tinkering with this at home.
yesterday at 8:17 PM
Unless you have electric resistance heat, and it's cold.
yesterday at 6:02 PM
> Device #1 is a Xiaomi TV Box S (3rd Gen): Amlogic S7D, 4× Cortex-A55 @ 2.5 GHz, 1.9 GB of RAM (about 1 GB usable)
Whoah!! A tv has almost 2GB of ram?! Does it need it to run Chrome or Electron? /s Ouch
yesterday at 7:26 PM
no electrons were hurt in the process of making this :(
yesterday at 4:38 PM
This is great. I've recently been playing with repurposing some old thin clients as homelab Linux machines
Not my site, but same idea: https://www.parkytowers.me.uk/thin/
All this old/obsolete gear is plenty capable when used with the right software.
yesterday at 6:34 PM
Funny, I use that exact website to find passively cooled hardware to use as a server.
I did get Qwen3.5 0.8B working at 5 tokens per second on the iGPU of a Dell Wyse 7020. I mostly use it to complain to; I have not yet found an agentic task for it. But also the CPU does plenty of small services and terraria servers and such.
https://nortyspock.github.io/blog/running-a-language-model-o...
yesterday at 4:47 PM
Out of interest, what have you been doing with them?
yesterday at 5:21 PM
Running a pihole, a WeeWx weather station, a NAS backup server, a wireguard VPN, Zot container registry.
yesterday at 10:03 PM
but why that matters in the age of scaling llm ??
yesterday at 6:09 PM
machine spirits are coming for my calculator
yesterday at 6:05 PM
This is the lord’s work right here. Models at the the intersection of small and useful — forthcoming models, admittedly— will be in the toolset that the next generation of poor kids will use to build a more egalitarian world.
yesterday at 7:25 PM
Thanks for your kind words :) hope to do something impactful one day with it
yesterday at 5:57 PM
It appears that it's not you doing that but claude doing that.
yesterday at 6:58 PM
That's programming in 2026. Right now, Claude it doing my job for me, but if it goes well enough I'm the one who's going to be promoted.
yesterday at 6:00 PM
most of the problem is actually brute forcing, so I'm using a brute forcing machine and just giving it pointers to reduce the size of search space for tricks, still fun, nothing frontier I know, but it's nice to see what we can do with our little brute forcing machines.
yesterday at 6:03 PM
The fact that you've commented two things under the same comment kinda feels like it's not your doing but the agents doing
mrashad74 3 minutes ago | parent | next [–]
most of the problem is actually brute forcing, so I'm using a brute forcing machine and just giving it pointers to reduce the size of search space for tricks, still fun, nothing frontier I know, but it's nice to see what we can do with our little brute forcing machines.
reply
mrashad74 4 minutes ago | parent | prev | next [–]
well, yes claude is doing that, and I'm directing it and building the experiments and direction only, I didn't hide it at all :)
Everyone is using LLMs now, so it's acceptable to use it for a fun experiment too I believe.
reply
yesterday at 6:20 PM
not really it's just a mistake of my side replying to differnt comments - I don't not use any agents for interacting online :D
yesterday at 5:58 PM
well, yes claude is doing that, and I'm directing it and building the experiments and direction only, I didn't hide it at all :)
Everyone is using LLMs now, so it's acceptable to use it for a fun experiment too I believe.
yesterday at 5:11 PM
ESP P4 or ESP S31?
yesterday at 5:42 PM
challenge accepted, I have one lying around, will see.
yesterday at 9:32 PM
AI skynet is winning. The spread of the slop is unstoppable now.
yesterday at 7:33 PM
[dead]
yesterday at 5:42 PM
[flagged]
yesterday at 4:43 PM
[dead]