prima-facie

I've backed up lots of CS courses, you know those old html websites written by the professors by hand. `wget -mkpnp` works very well for those. Even years after finishing uni, I've used those websites for studying for interviews. Too bad most of them are now gone.

> all of us in math and theoretical computer science and mathematical physics, ... are now in the same boat.

The most important takeaway from this is that mathematicians finally consider us CS grunts their equals.

Google is basically the incarnation of the Immortal Snail meme. OpenAI and Anthropic get millions of USD and are made immortal. But Google is the snail, constantly chasing them, slowly, day and night, relentlessly.

> Immortal Snail, also known as the Snail Assassin, refers to a hypothetical scenario in which a person is given millions of dollars and made immortal in exchange for being hunted down by a snail with a fatal touch for the rest of their existence.

Amazing news! I mainly use the F-droid Basic version which does automatic updating of apps. The regular one needs user confirmation for updates if I remember correctly.

I use grm when I need worktrees, which makes admin a bit easier; https://github.com/hakoerber/git-repo-manager

For one-shotted apps you still need to:

1. know how to do it well 2. maintain it 3. host it (if online)

> What if the world had one GPU per person?

It already does, in the form of mobile phones. It's just that these GPUs are not very powerful _yet_. But the building blocks to enable offline agents on mobile phones are already in place. i.e. Android AI Core framework.

From the article:

To block the scripts open the uBlock dashboard, select My filters, and add:

    ! AliExpress AWSC fingerprinting scripts
    ||assets.aliexpress-media.com/g/AWSC/uab/*/collina.js$script,domain=aliexpress.com
    ||assets.aliexpress-media.com/g/AWSC/fireyejs/*/fireyejs.js$script,domain=aliexpress.com

This is not limited to Bluetooth in any way. In pavucontrol I can see Firefox outputting audio when on AliExpress even though nothing is playing. The uBlock filter fixed it.

This is very cool, thanks

You can manage these easily with the huggingface python app.

hf download hf://unsloth/Qwen3.8-27B-GGUF \ Qwen3.8-27B-UD-Q4_K_XL.gguf

and then see them with `hf cache ls`.

Prune old versions with `hf cache prune`.

For SaaS forges there's also upd.dev launched recently.

For self-hosted forges: forgejo is great but also cgit is pretty good for exporting repos.

With the recent GitHub fiasco I've split my repos across multiple forges. Git was always supposed to be decentralised so why store everything in one place?

Silly question maybe: why use this at all when you already have a well known, decentralised, vendor-independent, almost real-time, and persistent notification system at your disposal? I'm talking about email of course.

> "Didn't you build this like... last week?"

This captures the essence of AI generated work Vs your own work. AI generated work just doesn't seem to persist in my memory.

If I design the application model myself then I can play around with it in my head even when I'm not at my desk. I know it inside-out.

If the LLM generated it then I very likely forget it the minute I stepped out of the office.

I'm interested in this as well. I would assume all of them from the announcement. I.e. they are removing that functionality altogether from regular Gmail (not Workspace).

That's not how you're supposed to use a LLM. This is nonsense.

My point is that a generalist-humanoid robot can be used for nefarious purposes more easily than a specialised robot like a Roomba. A humanoid robot could, in an extreme case, perform a civil arrest or prevent you from leaving your house until the police arrives. Ultimately it's our decision to make.

1. Do we want to be governed by machines?

2. Do we want to relinquish our freedoms to the whims of corporations?

This is exactly my point, thank you.

Doing what's ethically right rarely means taking the most convenient path. Fridges and pantries take a considerable amount of space in our homes. These could become part of an automated storage system.

A lot of these household tasks could be automated without the need for humanoid robots. Roombas can clean your floor, automated rubbish disposal can be integrated into houses and flats, grocery delivery and storage could be done via an automated intake, cataloguing, and storage system. Humanoids are just creepy and you can't really trust them.

Could you sleep easily knowing that you have one in your house? What if you oppose the political views of its creators?

I think this letters tells us everything we need to know about the man. I've rarely seen so much flattery towards the current administration, contradiction, deflection, half-truths and hypocrisy on a single page. This man is scared of everyone: the current admin, the public, and the other companies promoting open-weights.

> Open-weights models that don’t have dangerous capabilities are a public good

Knives should only cut during the day, knives which cut at night are bad.

I found the following config to bring a great QoL improvement:

    FILE: .gitconfig
    ----------------
    [alias]
        ci = commit
        fix = commit --fixup
    [rebase]
        autosquash = true

Now you can do quick commits as save-points, and squash all of them at the end.

    git tag last-good
    git ci -am 'WIP'
    git fix HEAD -a
    git fix HEAD -a
    …
    git rebase -i last-good

I understand this is important mainly for Nvidia GPUs. Is there any benefit at all for this (vs the existing VA-API) on Intel and AMD graphics? VA-API seems to work very well on both of these platforms, as far as I've tested.

Actually not really, Lunar Lake chips with Xe2 graphics have VVC decoding in hardware which can do 8K.

https://en.wikipedia.org/wiki/Versatile_Video_Coding

Phones are already running models locally which can be used in the field for specific use cases. Maybe not for frontier coding just yet.

Also you don't need to be connected to the network to use a local AI in many instances. If all mobile apps were done with a local-first approach, then you could use a local AI to query your emails, lookup already visited pages, summarise recently received documents, and lots more. Lots of apps could use an inbox/outbox approach for receiving and sending updates instead of relying on the network at all times. And this pattern could be greatly leveraged by local agents.

This piece is a bit all over the place. I immediately toggled off the `AI enhancements` and read the draft instead. The internet is already full of AI slop, I find human text a lot more valuable, even if unpolished.

LLMs will not be centralised or restrained to any 'clergy', the rabbit is already out of the hat, and open-weights models exist and are widely used. Probably not as good as the latest Sol and Fable but 95% there.

Codex and Claude Code without a doubt have very good models behind them. But they also have really good harnesses built around them. An LLM is only a brain stuck in a cranium in the dark. It can generate endless code/prose, but it can't walk or see on its own, it needs additional tools. If you read any of the local LLM subreddits you will notice people mentioning again and again that the harness/tool-use/template-tweaking makes all the difference on how a model behaves/on how smart it is perceived.

Some folks are already using Qwen models for their daily work. Maybe it can't work in a hands-off/one-shot fashion like the frontier models, but they can help tremendously if you already have some domain knowledge.

People are excited about local LLMs and it's not going away any time soon.

There's a glimmer of hope with ROCmFP4 which seems to double the current throughput: https://github.com/charlie12345/rocmfp4-llama

The Strix Halo is a great dev machine and a mediocre AI machine. You can run Qwen 3.6 27B at a decent speed, or larger MoE models, and that's about it. For some that's more than enough though, myself included.

Sure but where do you stop spending? :)

https://github.com/jamesob/local-llm

MediaTek Quad-Core chip, 10G Wan/Lan, SFP+, 2G RAM.