no_carrier

Pre this whole LLM era, Googled would've been roasted if they created a chatbot that hallucinated to users. I feel like this was the main reason they never turned it into a product and then ended up on the back foot.

Their Flash model is going head to head with the SOTA models. It's completely the opposite of embarrassing.

Thais will quite commonly bomb fish their neighbour's dam as often as they feel like it. I'm not convinced that they're also not doing it in the ocean, too. It may be happening away from the main diving sites.

Your source is a site called "Unherd" and the story doesn't exist on any other platform.

That's someone getting deported and has nothing to do with the claims that people were being forcibly removed into quarantine zones. This just didn't happen.

None of your three examples provide any evidence of hazmat crews forcibly taking away infected people. In fact, your links have absolutely nothing to do with anything at all. It's reasonable to assume that police would arrest people breaching curfew orders.

Under no circumstances was anyone in Australia forcibly removed by people in hazmat suits.

I'm more interested to know about the endpoint that was responding to all IPs.

Also, minor nitpick:

> Without that we would have been reduced to tracing through switch ARP tables (for switches smart enough to report that)

You'd only need your switch to have a viewable mac address table rather than an ARP table.

Android now runs a Debian VM, too. It's early days, but I'd be surprised if they don't get Linux GUI apps integrated into Android in desktop mode soon.

Every single LLM out there suffers from this.

There was a time in the 90s when quite a number of governments were aware of this Microsoft problem and insisted on open formats, so that important govt records and data could be open down the track. It seems like all of that was forgotten and traded for 'convenience'.

That's not exactly for slowness reasons. The creators of uv have stated that if pip followed the same algorithms they'd see similar performance benefits. People greatly overstate the Python performance penalty.

Tabs are the recommended indentation method for Go code. It is interesting that they switched after the first example, though.

You absolutely can run docker on the CLI in WSL2. The only requirement is you have systemd running in your WSL distro, which is fully supported.

Niri looks promising to me. After I discovered PaperWM I moved on to Material Shell, which was still the best workflow I've come across to date. It was basically gnome with a 2d grid: you could scroll workspaces left and right and then again on the vertical axis. The author sort of abandoned it (it was a Gnome extension) in order to go all in on their own desktop environment with the same concept. I'm guessing that'll take a while to get off the ground.

> Then came the SQLModel problems. The author pushes it very hard in the FastAPI docs

No it doesn't? The front page for FastAPI contains a pretty lengthy tutorial with no mention of SQLModel. The only time SQLModel gets a significant mention is on a page explaining connecting a relational DB, but it makes it clear that any DB at all can be used. Something has to be chosen for the tutorial, so it makes sense the author would choose their own.

If SQLModel isn't right for you then you're the only person to blame. I've been through that tutorial before and settled on plain old SQLAlchemy.