Tepix

21 ms·

451.
by Tepix 8mo ago
I heard that Windows on phones is about to make a return later this year, thanks to NexPhone.
452.
by Tepix 8mo ago
If you don't use the paypal app, you should be fine, right?
453.
by Tepix 8mo ago
If something truly unacceptable happens, you still have a while to switch to something else, in the meantime you will still have a working system.
454.
by Tepix 8mo ago
The Ryzen 370 lacks the quad channel RAM. Stay away.
455.
by Tepix 8mo ago
Have you thought about getting a second 128GB device? Open weights models are rapidly increasing in size, unfortunately.
456.
by Tepix 8mo ago
Sure, just like Linux doesn't run on a single phone on earth, unless you make some adjustments. I don't see a problem with that.
457.
by Tepix 8mo ago
What's wrong with AOSP though? It's also a Linux kernel. And someone has done the hard work of creating a decent UI and solved the other real issues.
458.
by Tepix 8mo ago
> " If Mastodon's not your jam, maybe star one of my GitHub repos. It's really the least you can do ." I like his sense of humour.
459.
by Tepix 8mo ago
Oh, Linux on mobile. My cynic take: - only works with very few phone models - battery doesn’t last long - bad UI with tiny elements. - not managing a smooth refresh rate - no apps That‘s the pattern we‘ve seen over and over again. The only
460.
by Tepix 8mo ago
Not that inexpensive at 70€ for 5 100GB discs. Why don‘t you store those discs over at some friends or relatives?
461.
by Tepix 8mo ago
What‘s the TUI equivalent for mobile apps? Big UI elements and text, iOS 1-6 style, rendered without bitmap graphics? I.e. something that is lightweight, lightning fast, great to use with just a thumb or so and looks a bit boring and dated,
462.
by Tepix 8mo ago
It‘s not just that - but I complete agree on not using a Personal AI assistant with some cloud service LLM provider. Anyway, by interacting with the world, the LLM can be manipulated or even hacked by the data it encounters.
463.
by Tepix 8mo ago
> Android phones can't support imessage because they are poor low quality phones that cannot handle imessage. Android phones can't use iMessage because Apple never opened it up, contrary to what Steve Jobs was hinting at back w
464.
by Tepix 8mo ago
Board games, card games.
465.
by Tepix 9mo ago
Yes. Radiating heat away is a big problem - I totally agree! But stopping heat conduction is a solved problem! Consider the temperature of your pot handles. That‘s all I was saying.
466.
by Tepix 9mo ago
Give parents extra votes for their children who are not yet eligible to vote. Perhaps half a vote per child for starters.
467.
by Tepix 9mo ago
James Webb, being an infrared telescope, has to be extremely cold: It's Mid-Infrared Instrument is actively cooled to 7°K (-266°C). Most other instruments (NIRCam, NIRSpec, FGS/NIRISS) operate at about 40 Kelvin (-233°C) using onl
468.
by Tepix 9mo ago
What are your solutions for if your AI bot wants to leak your credentials?
469.
by Tepix 9mo ago
I think stopping the heat conductivity is a solved problem - space telescopes manage to do this. But all in all, cooling in space remains a very big challenge.
470.
by Tepix 9mo ago
Check out https://github.com/kyuz0/amd-strix-halo-vllm-toolboxes/blob/...
471.
by Tepix 9mo ago
Prompt preprocessing is slow, the rest is pretty great.
472.
by Tepix 9mo ago
HTTP Header Injection or HTTP Response Splitting is a thing.
473.
by Tepix 9mo ago
Using lmstudio-community/Qwen3-Coder-Next-GGUF:Q8_0 I'm getting up to 32 tokens/s on Strix Halo, with room for 128k of context (out of 256k that the model can manage). From very limited testing, it seems to be slightly worse
474.
by Tepix 9mo ago
llama.cpp with rpc-server doesn't require a lot of bandwidth during inference. There is a loss of performance. For example using two Strix Halo you can get 17 or so tokens/s with MiniMax M2.1 Q6. That's a 229B parameter model
475.
by Tepix 9mo ago
It must be performing a man-in-the-middle for HTTPS requests. That makes it more difficult to do things like certificate pinning.
476.
by Tepix 9mo ago
If you can create a deno sandbox from a deno sandbox, you could create an almost unkillable service that jumps from one sandbox to the next. Very handy for malicious purposes. ;-) Just an idea…
477.
by Tepix 9mo ago
It sounds like a good idea, but there are no guarantees. To verify that nothing is written to persistent storage you'll have to look at target devices in question with the browsers of choice and then do forensics on them. Repeat after
478.
by Tepix 9mo ago
A personal assistant that runs in the standard cloud (anthropic in this case) is madness. That‘s the hill I‘m willing to die on. Run it locally or use a cloud provider you can deeply trust.
479.
by Tepix 9mo ago
Llama.cpp with its rpc-server
480.
by Tepix 9mo ago
You could buy five Strix Halo systems at $2000 each, network them and run it. Rough estimage: 12.5:2.2 so you should get around 5.5 tokens/s.

More ›