c16

No, because think of the children. /s

> My main problem with Mullvad right now is that its servers are blocked almost everywhere or generate an excessive number of Captchas

Why not buy a Low End Box and run one of the WG Setup scripts to get you going? You'll lose the anonymity, but its your box, a clean IP, significantly cheaper than a commercial vpn.

Sometimes the winning move is to not play at all. He's finding out what happens when people stop playing his game.

Why is this not criminal?

As other have mentioned, the code isn't the valuable part. But also there are alternatives now to these LLM providers.

They should have voted on if pineapple deserves to be on pizza while they were at it. The guy had to go.

It's been fun to see people re-writing and flashing their own smartwatch software with LLMs. Looking forward to the first TV OS rewrite video.

WiFi / 5G faraday cages exist, but they have been downrated for affecting the signal quality somewhat...

True. This then boils down to a quality vs speed decision. the 3.8 27b is far better than 3.6 A3B from my experience. I'm happy taking the speed hit, given local models aren't as intelligent as frontier models. Anything that can get me closer to my CC experience both in reasonable speed and intelligence is worth it. With that said CC can also be slow at times, so it's locally the difference in experience is not always noticeable.

> Running a large model locally comes down to one thing: how much RAM it actually needs in memory.

Not completely true. It's memory AND memory bandwidth. You can have 1tb of memory but if you have awful memory-bandwidth you'll also have slow tok/s. A3B helps with this, but so does MTP.

From my experience, you'd be better off running the dense 27b-mlx with MTP than the 3.6 version with A3B. You say your model is ~20GB of ram, but the 3.8:27b-mlx is 18GB and gets me very reasonable tok/s, and greater speed if you disable thinking when not required.

Qwen3.8:27b-mlx on 64GB MBP M4, I can get up to 42tok/s, more often than not in the ~30 range.

I like to think Stripe put $50B in a paypal account, and they got locked out with no way to get it back.

There are many topics, personalities and politicians we hear about daily who have no merit.

Qwen's advances do (currently) have merit.

Finding a vps from https://LowEndBox.com is the way. DigitalOcean is a rip off these days.

+1 to the long list of people hoping for Qwen3.8-27b A3B.

I really want a mini for running local models, but there's no way I'm buying one for the next few years. I'm planning on waiting for models to get better and more powerful, as well as for hardware prices to come down.

> I'd say a healthy 75% cut to token prices and they are back in competition.

Surely you don't want them to be the reason the bubble bursts?

I've written my own model harness and use Qwen3.8-27b-mlx with it. I don't want to say it's as good as Claude (I use Sonnet primarily), but it's not far off. What a time to be alive.

Big thank you to the Qwen team. 3.6 A3B was shocking good, and now I'm hoping they release an 3.8 A3B model too.

Edit: Having used qwen3.8:27b-mlx on MBP M4 64GB, I get around ~45 tok/s. A3B would be great for smaller devices, but it's definitely usable. As I understand it it's a mixture of MLX and MTP.

Some advice I got from another HN Mac user was to run local models in energy saver mode. You'll get slightly reduced tokens, but the laptop won't overheat and the fans won't go wild.

Gmail has categories. So they already are.

I've been using Goland/IntelliJ since I graduated. Been paying for it for years. But the resource hog and I guess focus on writing a mediocre AI side panel really killed it for me.

The main improvements lately were:

- line/function prediction

- Claude Code plugin

And I can use Claude Code in any terminal. So I've switched to zed.dev and use Claude there. Wish it had the line prediction though.

pretty sure the guy from TPB had a micro-payments/donation startup around the time when everyone was using Facebook like buttons on their websites.

As an ad-blocker user for years, I agree. On the flip side, how do blogs, news, and every day ad-serving websites survive without them? If it dies, it dies? Maybe this is not a discussion for here right now. But it's definitely something I've thought of. Not everyone want to pay for news.

That I don't know how to return odd or even in javascript?

Update: This works super well. Thanks.

I miss the days of playing Halo as a kid, and jumping on internet forums. MSN being the primary chat app that everyone used. Facebook was in its infancy, but everyone who had hobbies or a community was on a purpose made forum. People who knew how to write html/css/php build basic websites and blogs. Gaming clans came together, and xfire/steam was a great way to talk to and play with the same people.

Now days, I don't have Facebook, I don't play games, and the only forum I call home is this one. Times have changed, but so have I. At least I can reminisce on the good times.

Will give this a try later. Enjoy working with A3B Coder, but the heat coming out my 32gb M5 is a lot. This might be the trick - Thanks!

My experience so far has been Qwen3.5:32b-a3b-coder via Claude Code on a MBP 64gb M4, and a MBP 32gb M5. Just found about qwen3.6 so downloading that currently.

on 64GB M4 I find it's able to do things fairly well. The few times I run out of tokens, I hop over to that and I'm mostly unimpeded. I compare it to the Haiku models, where you have to go in and be surgical about your changes, or like others have said, guide a junior.

on 32GB M5, I find that it works, but around the 30% ctx threshold it slows down quite substantially, so more need to be surgical in your requests. I'll often just have my IDE open and Claude. But maybe I've been too comfortable talking to Sonnet/Opus and so forget I need to be more deliberate in my requests.

My finding here is that the harness is a big part of the problem. CC seems to be very good with Qwen in my experience. Better than OpenCode.

I also run DeepSeek for some other non-structured data tasks and to generate a to-do out of that. That's not coding, so won't go into that, other than to say it's very competent as a small model left to run in the background and automate small parts of my life and process.

tl;dr it's totally doable on a 32gb mbp using ollama, but be precise in your requests and guidance.

With DigiID, as with this, I never understood why countries give critical infrastructure contracts away from the country it directly impacts, provided they have a mature tech ecosystem. I thought the whole point was that it was critical?