tmzt

While it's true that early Linux ARM devices where embedded and generally only supported a single configuration, they didn't actually use devicetree.

Originally, embedded Linux ARM devices used a board file with a platform bus and hard-coded device metadata. The bootloader had to pass a machine id which told the kernel which hardware you were running on and which board file to use.

You can see remenants of this in the kernel still, though it's quickly being removed. I'm actually working on a hybrid kernel with the goal of bringing modern Linux support (on an lts branch) to old MSM7x300 devices, like the Evo 4G Shift I intent to use a tmux console/cyberdeck.

On another note, ACPI/UEFI doesn't always give you a clean abstract surface to work with either. ACPI is notorious for building in OS checks into it's compiled bytecode to the point that Linux often lies to it about what OS is running.

Has anybody thought about encoding AST tokens as LLM tokens, similar to how different words can have different meanings and that's reflected in their embedding?

Also working on something similar and using a dual LLM architecture (small router, larger deep thinker) with offline models, as well as determinisitic skills encoded as TSX.

It's evolved into a mesh-based operating system, gained it's own GPU-based AI library/runtime, and even molted and extended itself to ESP nodes.

Getting closer to a full release sometime in May. For now, pieces are released on my github.

If that type of scanning works and discovers these attacks, why not make it part of the publishing process?

Using LLM to target an FPGA would be interesting, generate the Verilog/VHDL.

is there a link to the project, the one in the comments won't open on mobile

If it helps I've been doing a lot of projects with TSX (via OCX) as a file format.

It works nicely for these things because it already supports typed parameters and can be treated as a a raw AST without JS engine.

Its also easy to manipulate with the same library.

Because one is for lawyers? It doesn't seem like a hard choice.

Same here. Then you see SOTA in a browser from Ex0byt, online 10x training (JIT-Lora), TurboQuant (Google), etc. Just saw KV prediction mentioned in this thread, so looking into that too.

I'm adapting all of this to Rust+WGPU with compute shaders if you want to follow along.

See this repo: https://github.com/tmzt/shady-thinker

Goal is Qwen3.5 27b on a Pixel 10 Pro running GrapheneOS.

Gemini Nano is supposedly doing it on device. It looks like something similar should work with Apple GPU and ANE.

Or a grinder and a camera. See CCC of years past.

What is the API key used for?

s/PROGMAN/EXPLORER/g

I built a MIT-licensed vite plugin (deadsimpleseo.github.io) to help with SEO for developers of React apps.

It bundles and executes a static HTML pipeline in a node vm executing within the vite process, then outputs that to a folder with an index.html for each of the .md and .tsx in your seo source folder. The result does not rely on javascript or fetching content to render, it's static HTML and there's no rehydration step.

This allows it to reuse your main app layout, headers, footers, etc.

I built it for a one page tool where I wanted crawlable links in the footer served from a standard nginx setup.

I've been looking for a modern equivalent of the white CRT VT420 as a distraction-free coding environment.

I could see building one with an RP2040 and OLED or Eink display.

If it helps, you can disable some of those limitations on Android:

https://www.reddit.com/r/AndroidQuestions/comments/16r1cfq/p...

I've used a similar drive on my Chromebooks for the last decade.

I went looking for a USB-C version and was surprised not to find one.

I would love to see one with SSD speeds if such a thing is possible.

Does anybody have something like this but usable in a remote environment like a self-hosted k8s?

Specifically, I'm looking for something I can leave running while sporadically connecting from a remote VSCode or electron app.

Right. But of course, intel was busy spinning off their Xscale business to Marvell. If they had seriously invested in it, they could have owned the coming mobile revolution.

They did push hard on their UMPC x86 SoCs (Paulsbo and derivatives) to Sony, Nokia, etc. These were never competitive on heat or battery life.

Oddly it works as a neologism as well. "moment" conveying that it was short-lived or fleeting.

A memory or artifact of a short period of time.

Perhaps the next question we are asking is "what happens if you give a statistical model symbolic input" and the answer appears to be, you get symbolic output.

Even more strangely, the act of giving a statistical model symbolic input allows it to build a context which then shapes the symbolic output in a way that depends on some level of "understanding" instructions.

We "train" this model on raw symbolic data and it extracts the inherent semantic structure without any human ever embedding in the code anything resembling letters, words, or the like. It's as if Chomsky's elusive universal language is semantic structure itself.

I'm not sure if this would work for you, but there are inexpensive devices that plug into an HDMI port. They appear to the computer as a monitor. I use them for screen sharing to a remote display, but they should enable to think there is a monitor attached. It negotiates the display information as if it was an actual monitor.

Here's the pack of three I purchased on Amazon.

Woieyeks 3 Pack HDMI Dummy Plug https://www.amazon.com/dp/B0CKKLTWMN

This could be done using something like llamafile, in the sense that it would be a small universal executable containing a package.

It could even support running as a self contained application on Windows, with all needed DLLs provided.

There is also support for doing this without being root:

https://developer.android.com/develop/connectivity/usb

I watched the video and am curious about the protocol loading apps from add-ons. Do the resistors j7st say what k8nd of connection is used (i2c, serial, etc)?

Also, I've been looking for cyberdeck and clamshell designs using CM4 and am impressed with how you sourced the keyboard. Is there a reason the replace the matrix chip with an i2c extender and not use the original chip?

Hello there,

Any idea why Microsoft didn't use this in WSL1 or 2?

Is it more efficient than hyper-v with hardware acceleration?

Can you see it being useful again? Or does it make sense to have a hybrid where the code runs using hardware acceleration but the timers are cooperative?

At one point I believe DSL used Kdrive which has a slightly different DDX and driver structure but the same DIX.

I searched archive a while ago and couldn't find it. It's probably referencing the announcement but I clearly remember a pull quote on the ELKS site, probably the same one quoted above.

I happened upon the site when first playing with a (z80) Gameboy compiler (probably SDCC) and wanting to port Linux to it.

I fully admit i could be imagining it.

Yes, I should have said 386.

I would say Gemini Live is getting there. It's lacking integration with NotebookLM and Keep. It would be amazing if I started a project conceptually and wanted to move to code it could fire up VS Code and let me get to work.

Gemini's home automation works nicely and it can understand comments like it's too dark in here or it's cold inside and act appropriately. This is using the Android app as an assistant, not live mode.

OpenAI's implementation is apparently similar but I haven't tried the voice mode as a free user.

I haven't tried Apple Intelligence yet on my M1 and don't have an iPhone, so I can't compare.

I've been looking at offline capabilities with open weight models but they aren't there either. A full speech-to-speech model [1] working on an M1 Mac would be incredible.

[1] https://arxiv.org/abs/2410.00037

The first versions of Android that the public saw were very similar to the OS on a BlackBerry or Danger HipTop. The G1 even used the same mechanism to deploy the keyboard.

As far as the rounded corners, I remember seeing a reduced Google Reader view of Engadget later that year that had every device looking the same from the top third up. I really wish I had a screenshot.

There is now a lot of cross inspriation and features that are copied in both directions, as well as both implementing the same thing at around the same time (Intelligence and Gemini).

On the Android side, Pixel gets most new features first while Samsung offers their own take. Samsung is generally ahead of their direct competitors in terms of hardware.