There’s something uniquely painful about debugging your own old code. You remember the intent, but none of the decisions that led to the implementation. It really does feel like archaeology.

You design the sandbox so the agent starts in that layer. The next thing you can do is to limit the network access, this is what I'm working on right now.

Or do you mean something else?

Every major power dislikes election interference until it's strategically useful. The surprise isn't that countries try to influence each other. The surprise is that anyone still treats it as exceptional.

Hot take but academia and industry don't have fundamentally different goals anymore. The real challenge is making sure students don't end up satisfying neither. A PhD that isn't published enough for academia and isn't impactful enough for industry is the worst outcome.

I wonder how much of that is scale rather than danger. Really large objects can feel unsettling even when you know they're harmless. VR seems especially good at triggering that response.

I wouldn't be surprised if "the best model" becomes a niche concept over the next few years. For most production systems the winning architecture may end up being an orchestrator that knows when to call a cheap model when to escalate to a stronger one and when to combine multiple outputs.

there are so many branches of evolution here, I also see them converging to a "best model becomes a niche concept" too

the supercycle is on device models, and one of those evolutions is models baked into chip die, and you just upgrade chipsets every few years instead

so it's the hyperscalers that will take the L in that environment

Isnt that the idea from gemini cli?

I think people send too much time optimizing the notebook instead of optimizing what goes into it. I've seen beautifully organized notebooks that were never opened again and messy scraps of paper that someone referred back to after years.

There was a post about something similar here a few days ago, it called it a "dirt notebook" https://news.ycombinator.com/item?id=48954149

This is one of those cases where security mechanisms ended up shaping the workflow itself.

Constraints like these are such a good way to learn programming. When you have limited memory and a slow CPU, every variable and every line of code has a purpose. It's hard not to come away with a better appreciation for how software actually work.

One thing i have always linked about Jane street projects is that they tend to package ideas that have existed in research or niche system into something developers can actually use. Even if you never adopt the library, the design docs are usually worth reading.