They’re basically just auditing skills using another LLM. It’s better than nothing, but hardly effective.

I don’t think there’s anything other than reducing permissions that can make these types of tools secure.

Things are only to get more interesting with the MacBook with an A chip instead of an M chip, and the MacBook Pro with touchscreen later this year.

I fear at some point we’re going to lose macOS.

Hopefully linux support improves by then to take the defectors if MacOS goes away.

macOS is the main reason I’m using mac, though. the whole integration with the Apple ecosystem is the important thing, the hardware itself is overpriced

You’re forgetting just how convenient a tablet is and how little average people care about the latest and greatest triple-A games.

I've found a tablet convenient in 3 situations: Watching video, reading ebooks or displaying sheet music. (And a single tablet is rarely very good for more than 2 of these at a time.) Otherwise it's either too cumbersome or the I/O is too useless.

Just browsing the web on the couch. It's so much better to have a 10" screen than whatever your phone has.

And even on an iPad you can put a video running in the corner while you browse HN or Lobste.rs.

Yeah, it appears they have a lot of things backwards, for example:

> “We were very disappointed,” Franklin said. “That means perpetrators of crime, people who are maybe engaged in domestic abuse or stalkers, they can request footage and that could cause a lot of harm.”

The whole point is that they should have been collecting data on perpetrators of crimes only in the first place, not a massive dragnet.

It feels so in bad faith as there’s a claim ICE could use the footage if it’s public domain. They already can and that’s one of the big arguments against it you clowns.

beyond that it also implies "but the bad guys could then trawl all this data; only us, the good guys should be able to do that!"

Yeah, that's even more important, isn't it. It begs the question, framing the conversation as "Who should have access to this invasive, privacy-destroying footage" rather than "WTF? Stop this."

Stateless vs stateful concurrency management is very different, though; I can roll back / replay a mail box, while this isn’t possible with shared locks. It’s a much cleaner architecture in general if you want to scale out, but it has more overhead.

Swagger / OpenAI is to trigger things in the backend, this is to trigger things in the frontend (which may, in turn, trigger things in the backend).

Ahhh, okay I misunderstood it then thanks.

This implies that OpenAI must build and release and maintain a model without any safeguards, which is probably the big win and maybe something Anthropic never wants to do.

I don't think that is the correct conclusion.

But they won't be releasing it, they will be leasing it to DOJ and all their other customers will get the safeguarded model.

Sorry, meant to write DoW / department of defense, not DOJ.

People forget that the code is the spec. Usually it’s more effective to maintain comments.

Heck, given that LLMs are language driven, why not bring literate programming back? Knuth would be proud.

I must be a fossil living under a rock, but: were they ever gone? As the amount of new CLI based applications I install on a monthly basis is always far more than the amount of new GUI based applications.

I think a lot of new developers used GUIs very exclusively for a very long time. Agentic workflows and Claude code have brought CLI tools to the forefront again.

A good example of this that I've noticed is a lot of newer devs were using GitHub Desktop or VS Code to manage git operations, but Claude gave them peek under the hood and now they're using it directly a bit more. Claude Code is a great gateway drug to CLI and TUI addiction

Isn’t Elon the Elon of AI?