jbstack

There's definitely appeal in key-driven window managers in general. Projects like i3 and Niri are popular. But you can get that with any Linux distro (albeit not many have it set up that way by default). You don't generally choose a distro just for whatever DE/WM it happens to start with.

As someone who uses Signal and WhatsApp, I would be extremely happy if those companies blocked access in my jurisdiction if such laws are implemented. In fact, if they agreed to insert backdoor access I would never trust them again, even if they subsequently reversed it.

We have to be willing to suffer personal inconveniences to stand up to these types of policies.

With all the current concerns about rogue and mis-aligned AI, and about how we've made essentially zero progress on making sure that AI is safe and trustworthy, NOW you want to put backdoors in encryption algorithms? We should be doing everything we possibly can to make our systems more secure, not less.

This type of policy was terrible when they first thought of it. Now it's outright negligent and dangerous.

I do something similar. I have a base NixOS image in Incus, with whatever tools apply to every project (e.g. Git, OpenCode) already installed. When I work on a project, I spin up a VM instance, use nix shell to add any project-specific tools, then share only the project folder from the host to the guest. This way, the worst the agent can do is destroy my project folder, and I can always restore that from another clone of the repo.

I know a lot of people are using containers for sandboxing, but given how capable the latest models have shown themselves to be for breaking out of sandboxes, I prefer the extra isolation of VMs for this.

I do all this locally - it's an interesting point to able to turn the laptop off but keep the agents running. I might consider running some of these on my homelab server just for that.

> Permits

Obtained by your AI agent, granted by another AI agent.

> Materials, trucks, equipment etc.

Ordered by your AI agent, delivered by a self-driving truck or drone, loaded/unloaded by robots to/from the warehouse/site.

Most of this stuff can already be done today in some form, so its just about improving the autonomy and integrations along the supply chain.

> Xephyr

Given that X11 is becoming more and more obsolete over time, what's the Wayland option?

There's a lot of middle ground between doom-scrolling type content and a 10 hour block of non-stop immersion. It doesn't have to be one or the other.

Many factories in China now already have 90% of their work being done by robots. Is it really so hard to imagine that we reach a point where they can be prompted?

You're assuming a handover happens in a single definable moment. More realistically it will be one small decision making power at a time, so that at any given moment it never really feels like control is being given up. Boiled frog.

> in Germany people are reluctant to give their IBAN to others

Are they worried people might send them money without their permission?

Doesn't matter if you aren't asking the type of questions where hallucinations are relevant e.g. you're seeking pure reasoning rather than factual information.

> I thought the supreme court said the tariffs are illegal without congressional approval

The ruling wasn't as generalised as that. What it said was that the president doesn't have the authority to impose tariffs under the provisions of the International Emergency Economic Powers Act, which Trump had cited as the basis for some tariffs. That made those tariffs effectively illegal, but the ruling said nothing about tariffs which had other legislative provisions as their legal basis. The tariffs that you see today cite some other rule as their basis rather than the IEEPA.

You want both. The harness and the sandbox do different things.

The harness says "You have access to tool X, Y, and Z, but not A, B, C".

The sandbox says "If you try to use X to access a forbidden resource, I'll prevent you from reaching it".

> Moreover I am not sure it is even good advice?

I think it is. He isn't saying to learn how to train a LLM so that you can go on to train LLMs. He's saying to learn it so that you gain a deep understanding of how LLMs work. Ordinary startups can still benefit from things like training or fine tuning highly specialised smaller models, knowing how to select and configure an appropriate model for the task at hand, knowing what software to use and why, understanding what's going on behind the scenes instead of treating everything like a black box, having a higher level of intuition about LLMs generally, etc.

Most computer science courses do in fact teach things which are lower level than coding, such as how transistors work.

I think this is the only sensible way to work with agents, if you care about code quality and reliability but still want the benefits of AI. There seem to be three camps that people more or less fall into: (a) AI is terrible/bad/evil and should never be used, (b) you should one-shot everything and be happy if it seems to "work" when you try it, (c) the middle ground, where the AI writes code which you carefully review.

I definitely prefer (c). But I get why (b) can feel necessary. If your competition is using (b) there can be pressure to do the same just to keep up.