mendigou

I am former AWS and this is pretty accurate.

The other factor to add here is that, with some exceptions, the whole company feels like a Rube Goldberg machine and very few people care about what happens outside their cog (because they’re not incentivized to do so).

The slower car should signal the moment they decide they need to merge into the lane, then wait for an appropriate gap.

I’m sorry but that is an insane interpretation of turn signals. If you only turn them on once you’re about to execute the maneuver, they’re useless. The whole point of them is to warn other drivers you intend to do something so they can anticipate your movement and allow you to do it safely.

How are turn signals explained in the driver handbook (or equivalent rulebook) in northeastern US? As far as I know they clearly state that turn signals are about intent, so they should be turned on the moment you decide you need to turn or switch lanes.

I would be interested to hear what advantages you find Antigravity CLI has over Claude Code.

I keep hearing this thing about apps and I'm confused. I write code on my MBA M1, run orbital simulations, run a media server... All like I would on Linux. What am I missing on userland?

I understand there is low level system stuff I can't control, but I've made my peace with that. When using Linux I hadn't touched its internals for years.

People in Germany and the Netherlands must extra hardened I guess, since everyone bikes around there, even with their kids in tow.

How is it worse for men?

Where do you get these?

After so many years, so many people still believe in this user control paradigm.

Giving users control works for the slim percentage of power users. Most users will end up obliterated by scammers and other unsavory characters.

Perhaps there is a way to give control to today's users (that includes my non-technical mother) and still secure them against the myriad of online threats. If anyone knows of a paper or publication that addresses this, I'd love to read it.

Indeed there was a split second of primal flight response, and then awe. Not like you have a long time to think about it!

I was lucky enough to see this one live.

It was astounding. It became day for a few seconds and then it broke off into a half dozen pieces or so. It left an incandescent trail behind.

It looked incredibly close. I had the feeling we were going to hear it coming down nearby in a few seconds, but it ended up in the Atlantic Ocean (actually not that far from us since we were by the ocean, but far enough).

So glad we were having drinks outside and I wasn't grabbing anything in the house!

Not disputing all of that, but telling the GP flat out "no" is incorrect, especially when distributed training and inference are the only way to run modern massive models.

I haven't touched this in a while, but you can train NNs in a distributed fashion and what GP described is roughly the most basic version of model parallelism, where there is a copy of the model on each node, each node receives a batch of data, and the gradients get synchronized after each batch (so they again start from the same point like you mention).

Most modern large models cannot be trained on one instance of anything (GPU, accelerators, whatever), so there's no alternative to distributed training. They also wouldn't even fit in the memory of one GPU/accelerator, so there are even more complex ways to split the model across instances.

How is Germany similar to the US? Perhaps in the enshittification of society, but certainty not in the forces that drive innovation.

Germany could be that innovation hub in Europe, and it seems incredibly resistant to do so. And all of the EU is trailing behind, unfortunately.