whitehexagon

There are not many sites I whitelist for javascript, or even bookmark these day. Really glad I tried your game, it's fun and nicely executed. Well done to you both.

Agreed. I can only get dried nettle leaf tea here, although seems to be going out of fashion, same with Ginkgo.

I got spoilt for teas when I was contracting in Germany. Nettle was very common along with some other traditional teas, including one that was good for flu/fever that I can never remember.

Agreed. I had no problem with the conBee II itself, but it did take me a while to figure out different IKEA ZigBee devices required different 'secret handshakes' to tigger pairing. eg number of button presses within a second, I think one was 5x, another 4x, and lights were different again. Still, I prefered that hasle than having a hub dialling home.

But it was pretty stable once it was setup. Just occasional reboot on the rPI but I think that was my flakey SMS gateway code.

The ZigBee range used to include a bulb and remote, so you just leave the mains switch on.

Otherwise, I used floor lights in the past with WiFi switchable sockets before I switched to ZigBee. The WiFi ones wanted to dial home.

Although I have been trying to only buy 'needs' not 'wants' recently, I did stockpile a few IKEA ZigBee gadgets before they retired them. One of the few product lines their MBA's hadn't destroyed.

I was working on some Golang code, talking to them via the very open ConBee II ZigBee gateway. Great fun, and very fast once I got subscribe vs polling working. So now I get an SMS for door access, but kinda hopefully never for a water leak.

No interest in yet another 'standard', especially since Matter seems to mandate PKI device attestation. ZigBee just feels more open to me, and I have enough eWaste devices with expired certificates.

some background:

https://news.ycombinator.com/item?id=44507971

Agreed, GPU is the expensive route, especially when I was looking at external GPU solutions.

Using Qwen3:32b on a 32GB M1 Pro may not be "close to cloud capabilities" but it is more than powerful enough for me, and most importantly, local and private.

As a bonus, running Asahi Linux feels like I own my Personal Computer once again.

I only just got rid of my spy-phone, why would I want something with even more intrusive spying on me. I certainly would not inflict that upon others. A sure way to lose friends fast.

BigTech has burned so much good will at this point, that every new venture just feels like a timer ticking down to a bait and switch for ad revenue, subscriptions, pro features or just selling our data to the highest bidder.

And what happens to 'local' data when the 3 letter agencies want access. No thanks, sounds completely dystopian. If the data is there, someone will find a way to abuse it.

Sorry for this, but it sounded like the start of a joke, and I couldnt resist:

Three AI billionaires went out for chicken and beer, they had no money to pay, but promised to buy a million meals over the next five years, but demanded the owner pay for a server-farm in compensation for the restaurants upcoming success.

Qwen3:32b on MBP M1 Pro 32GB running Asahi linux. Mainly command line for some help with armv8 assembly, and some SoC stuff (this week explaining I2C protocol). I couldnt find any good intro on the web-of-ads. It's not much help with Zig, but then nothing seems to keep up with Zig at the moment.

I get a steady stream of tokens, slightly slower than my reading pace, which I find is more than fast enough. In fact I´d only replace with exact same, or maybe M2 + Asahi with enough RAM to run the bigger Qwen3 model.

I saw qwen3-coder mentioned here. I didnt know about that one. Anyone got any thoughts on how that compares to qwen3? Will it also fit in 32GB?

I'm not interested in agents, or tool integration, and especially wont use anything cloud. I like to own my env. and code top-to-bottom. Having also switched to Kate and Fossil it feels like my perfect dev environment.

Currently using an older Ollama, but will switch to llama.cpp now that ollama has pivoted away from offline only. I got llama.cpp installed, but not sure how to reuse my models from ollama, I thought ollama was just a wrapper, but they seems to be different model formats?

[edit] be sure to use it powered, linux is a bit battery heavy, but Qwen3 will pull 60W+ and flatten a battery real fast.

I was wondering what it would take to fork the internet. Or more seriously, what it would take to create an open source / hardware alternative. Maybe something true to it's routes of academia, ie no commercial activity permitted, especially adverts! Maybe even limited to some of the original simple protocols like HTTP, FTP and maybe RSS?

When I was in Holland many years ago there was a university initiative to connect the whole town with it's own hobby network. I bought some directed antennae to connect, and it worked, but I seem to recall it was line-of-sight WiFi stuff and thus very limited range.

I also worked for an airline that had line-of-sight laser communications, but again it was only short distances.

Are there any modern alternatives that would make such a people sponsored/focused hobby network possible?

The first thought I had was long-wave radio, but I cant imagine there is much bandwidth there, but might be okay for a text only protocols?

Not sure. As a developer I like the concept, but the more I read of the legal, the more concerns I have. It reads more like an online tablet, but without service guarantees. Maybe some of the legal was for the website, it wasn't clear to me. But there was also mention of social media, complete game usage collection, and phone and email direct marketing, defaulted to opt-in.

Sure, but then maybe it should be two models; one with WiFi for adult gaming, and one that is 100% offline safe model for kids to play with. With the added advantage that the SD card based games can be managed by parents to help manage 'game time'.

This looks like a perfect match for the music synth ReacTable, something that never ported well to ipad because of the lack of physical pieces.

But my concern would be that this becomes just another Ad platform, but targeted at kids.

For me a board-game is offline time. So I would picture this with no WiFi and SD card based games. Which could still be profitable via an other-device app store. But would also avoid temptation for developers to add these more addictive online/networked games.

>"The truth is that we have failed to avoid an overshooting above 1.5C in the next few years. And that going above 1.5C has devastating consequences."

So we were originally predicted to hit 1.5C in 2040, and now it's predicted to be 2029, but last year was already globally 1.5C hotter.

I dunno, the whole thing makes me want to seek a forest to hide in, although I think Europe had record forest fires last year too.

It feels like the only game plan in town, is to rush build new fossil fuel power stations, and hope that AI reaches 'God in a Box' stage soon enough to be able to pray to her to save us all? Would such a solution even be shared with the world if it reduced profits?

And it is all about profits, right? So for November I have decided that I will reduce my spending down to essentials only. Needs not Wants. Anyway, nothing I bought this year really improved my quality of life, apart from some extra solar panels and an FM radio.

I worry that this global push for 'Know Your Developer' and the attempt to make them legally liable for what they produce, is going to destroy open source, An 'open' linux included.

After that, certified locked down BigTech 'Personal Computing' will be the only menu choice.

Well done to you both. I'm only a 1/4 way down your useful TODO list after 12 months. I got bogged down in setting up IRQ vector tables on armv8 and took a huge detour to refresh my assembly skills. So I feel some of the journey you have been on. It takes a lot of patience, but can be very rewarding. Congrats!

Agreed, a very refreshing change.

I cant reccomend anything apple or android based, but the PinePhone has some reasonable low-level documentation on itś SoC (Allwinner A64). And there are a great series of tutorials of how Apache Nuttx was ported to the PP. Here's one of the earlier ones: https://lupyuen.org/articles/uboot

I've been working through them myself but still dont understand the boot process, and a long way from radio chip functionality. But I suspect you'll find radio chips quite locked down, since they are licensed for connection to mobile networks. Often there is just a unknown binary blob running the show.

I do a full shutdown, since it boots from cold into Asahi in ~30s (M1 Pro). Plus I´d rather have a laptop that I know is shutdown, than this crazy 'we might need to do an update whilst it's in your backpack' fake sleeps.

I think my next laptop will be M2 Pro/Max, but honestly at the moment I dont see the need for anything faster. And having discovered Asahi, I dont think I could ever go back to a BigTech commercial Ad platform OS. This feels like My computer.

I love the feeling that I own my computer again, thanks Asahi team. But I am so attached to my MBP M1 and remix 42, that I was kinda hoping that this was the final release at the end of the universe, maybe with small patches for the remaining bug-blatterings. It certainly feels like I could call this 'my last computer', and if not, then my next computer will be an M2 running Asahi.

Anyway, keep up the good work. I am currently bashing out some bare-metal Zig and assembly for PinePhone, so I get a small taste for the challenges you face, although I have some documentation. How you manage without that I have no idea, incredible work!

One thing I have been wondering is that with your approach, is it possible to run a Asahi VM inside Asahi? It would be really nice to be able to have sandboxes for all these misbehaving modern apps like browsers etc. I got the impression from earlier posts (is it muvm doing this?), that things like the GPU pass-though dont affect performance much, so it sounds like a sandbox VM approach could be very performant and powerful. Indeed, these days I run run every 3rd party app in a sandbox if possible.

I still have fond memories of traveling down to London to visit a computer trade show, and leaving the Psygnosis stand with a giant cardboard Lemming, window deco Lemmings, and a much prized Lemmings Demo disk, which is still sat in a box here with the full game.

I think that was the same show I came away with DevPac, probably my best tech investment ever, and maybe a 10?MB hard disk. I often wonder what happened to the assembly code for my two player pacman game, level editor included.

68k & Amiga chipset was a dream to code for. Although I'm having some good fun at the moment with some armv8 assembly & Zig on the PinePhone.

27B$ funding for a new data center, and massive AI staff cuts, does that mean they already reached their 'god in a box' breakthrough? Or just decided ad slop bots are progress enough. Feels like mixed signals anyway given the industry investments that are happening around AI.

I'm running a 4k monitor from the HDMI port on my M1 MBP with Asahi. I was thinking the same, so discovering a hdmi port was quite a nice suprise, especially since only 3xUSB-C

It's a nice thought experiment, but for finished software we would need finished, and well documented hardware, like the gameboy example, and that would be in conflict with most of the profit models of BigTech. It's a shame because we probably already have 100x the processing power at our fingertips than we'll ever use, and certainly more than we 'need'.

But let's not forget to include a finished programming language (other than C89), finished tools, and a finished OS. Although these days it seems like modern programming languages are just academic playgrounds for feature experimentation. And outside of open-source, OS 'enhancements' seem designed mostly for delivering more ads or subscription services.

But even if all that could be solved, I have never worked anywhere that would be willing to pay for more than 80% of what an engineer considers 'finished'. Because we kinda know that the last 20% would be 80% of the budget. Those long term support and stability savings of a well engineered solution are impossible to sell upfront.

I dunno, didnt they already crack the 400GB/s memory bandwidth some years ago? This seems like just another small bump to handle latest OS effects sludge.

Now the M1 range, that really was an impressive 'outperform' moment of engineering for them, but otherwise this is just a clock-work MBA driven trickle of slightly better over-hyped future eWaste.

To outperform during this crisis, hardware engineers worth their salt need to designing long lived boxes with internals that can be easily repaired or upgraded. "yeah but the RAM connections are fiddly" Great, now that sounds like a challenge worth solving.

But you are right about the software. Installing Asahi makes me feel like I own my compter again.

I'm currently hacking a toy OS in Zig on the PinePhone, and I have to say the documentation is a bit painful, or sometimes just missing, for parts of these complex SoCs, and that is meant to be a fairly open platform.

But the modem binary blob is a whole other world, and I am not sure how they could tackle that, since my understanding is that this is partly done for carrier licensing reasons? ie. to avoid abuse on the cellular networks. So isn't an open source radio driver also going to have to be licensed in the same way, and then ultimately shared as another binary blob?

The PinePhone compromise seems to be 'isolating' the modem & blob at the end of a USB link. Although I'm not 100% sure how that works yet, since I only just got the graphics & fonts working.

But even that is a bit of a puzzler, since I'm currently framebuffer-to-lcd based, but I know there is a Mali GPU hiding somewhere. I suspect that will also involve another blob. Anyway, the framebuffer approach seems fine for now, it is booting in ~2s, and the less binary blobs involved the better.

It will be interesting to see what FSF can achieve. But, personally I think they would be better focusing on a fully open-hardware dumb phone, and build upon that.

I think my 2001 MBP M1 Pro is ~200GB/s memory bandwidth, but it handles qwen3:32b quite nicely, albeit maxed out at ~70W.

I somehow expected the Spark to be the 'God in a Box' moment for local AI, but it feels like they went for trying to sell multiple units instead.

I'd be more tempted by a 2nd hand 128GB M2 ultra at ~800GB/s but the prices here are still high, and I'm not sure the Spark is going to convince people to part with those, unless we see some M5 glutenous RAM boxes soon. An easy way for Apple to catch up again.

I bought a decent 4k monitor to go with my 4k media box, and the required 4k cable, but wouldnt play because 'for your safety, blocked because not a TV' or some such nonsense, think it was HDCP? Annoying enough for me to cancel 4k paid tv, plus they were busy agressively reducing the broadcast bitrate. Also ended up with a load of bluray 4k discs I cant use, and they wonder why people download stuff.

Nice. I´d like something stylish like that for my solar inverter, the fans drive me mad.

Nice, what did you settle on?

I'm playing with qwen3:32b on an Asahi M1 currently, and it does more than enough for me, but might be tempted if there is a bigger model within reach.