bee_rider

I’m actually not sure where the gap is, in the US, on foundries. It might be more like people with a bachelors and some experience.

We could also maybe use more semiconductor physics PhDs?

In the future, will a 2 year old model be useless? In 2 years? 10 years?

This is kind of an interesting thought (underneath the grumpiness, which I guess is where the downvotes came from). Programmers seem to hold aviation processes in very high regard, but their industry is also under cost pressures. The “UX” of riding in a plane does kind of suck, I wonder if some of that comes from the costs involved.

Not too surprised to see pressure to reduce the number of pilots.

I do wonder with polls like this, that include middle-of-the road options like “fairly important,” if there could also be some change in how we communicate our preferences. Maybe we’ve got a general creeping uncertainty about everything.

I think the “it’ll come crashing down” argument comes from cheaper inference, not more expensive. The frontier labs have to pay for these massive training runs. If inference gets cheap enough that you can run it yourself, why run it on their side? Open weights are (much) less than a year behind.

Means-tested disaster recovery assistance could be worth considering I guess.

Maybe that was a reasonable thing to do, no need to re-litigate the past. But, we collectively made a prioritization decision. We don’t get to just ignore the downsides of that decision.

Lots of areas have natural hazards. Should we help out people hit by tornadoes or floods?

With climate change induced problems, at least there’s some justification. We collectively (as a country… you and I personally might not have been alive or voting when the mistake was made) failed to elect governments that would prevent it, so we we collectively get to pay for it.

I think they describe the CPU using the weights? I mean in some sense a CPU can be viewed as just a function, right?

Graphs theory, like in general? Just as one example, an awful lot of numerical linear algebra research is devoted to the handling of sparse matrices (ones where most of the entries are zeroes), which can also be looked at as graphs. Likely any CAD program with a finite element analysis tool uses some graph theory under the hood.

It might not be tested. But, it would seem very weird if anyone other than the person using the LLM was accountable for the output.

None of these analogies work very well because providing services across the border without crossing it doesn’t scale very well in-person. I should have just said that directly, but the image of cross-border bagel service was too funny to pass up.

WRT your second paragraph, I think I addressed it already in my second paragraph.

If someone put a bagel shop on the Swiss/French border and sold bagels across it (using, I don’t know, a long pole or something) I guess there would be some questions.

Realistically, Switzerland and France have a relationship; treaties, etc. Quad9 has to comply with Swiss law, and they probably hope that making reasonable good-faith moves to block French users will cause Swiss courts (if it comes down to that) to see them as the more sympathetic party.

I don’t see how you can claim there’s a dichotomy when the “second thing” is something you’ve conjured up from your own assumptions.

The lighthouse thing seems much more odd. That humans take social shortcuts and justify them as rational is… I mean, the only thing that makes it not the most obvious “contrarian take” about humans, is that it isn’t really all that contrarian at all.