First, I'd paste it into notepad just to make sure.

The Meta Rayban glasses have a light when it's active. Now granted, many people do pay a mobile repair person to permanently disable that light. But personally, I think the people that do not disable that light shouldn't be penalized. And for that reason, I think the law is badly written.

They're not banning AI generated writing, they're banning unverified content.

Go ahead. Do it then. The second queue already exists anyway. It's called a blog. Anyone can create one. And no one can prevent you from publishing to it, or from archiving those blogs permanently.

A single user can use AI to generate terabytes of unverified claims every single day. At some point, something has to give. Even with blogs, search engines will have to be judicious about what's worth indexing and what is just unverified noise.

The claim that "false premises invalidate everything built on them" only holds when the error is unrecoverable and propagates non-linearly.

But in many cases:

The error is consistent and quantifiable, so results can be recalibrated

The relational findings survive even if absolute values shift

The direction of effect holds even if the magnitude is off

A good counterexample of this is carbon dating. Early carbon dating used an assumed atmospheric C-14 ratio that turned out to be slightly wrong. The premise was false, but scientists didn't throw out decades of dating results. Instead, they developed calibration curves (using tree rings, coral, etc.) to correct the systematic offset.

No, Saleforce degraded their service as soon as they took over. It's been a slow death ever since.

If Salesforce is not extracting a pound of flesh every year from each customer they have, they're not happy.

Yes, but when when you tell a scientist to uproot their entire life from DC and move to Kansas or to the middle of nowhere. And that even if they move, that the future of their agency is pretty much over, or at least put on life support.

That could be considered a form of constructive dismissal (even if they call it just "quitting").

The project was doomed to fail from the very beginning.

Apple is too privacy focused. To make AI competitive, you have to be willing to share private data with it.

Right now, Gemini Live on Android sucks terribly, but I am more hopeful for it..

I'm not the original poster.

Nexperia, I don't know. But the US did have issues with China having access to the technology from its Dutch sister company ASML.

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.bloomberg.com%2Fnews...

So it's not far-fetched to think that the US has something to do with this issue with Nexperia as well.

That couldn't possibly happen!

Foreign companies (Tesla excepted) are not allowed to own more than 49% of Chinese-based companies/joint ventures.

To be fair, (with the exception of Tesla), foreign companies are not allowed to own more than 49% of Chinese companies.

So if international law was truly applicable, then there would be parity between nations about foreign majority ownership.

When you buy your ticket, you can just specify that you're disabled and that you need a paper ticket as a special accomodation. After all, they already have these questions for people who need other accomodations (for wheelchairs or food). It shouldn't be too hard to add one more to the list.

And for the passengers that don't have the foresight to check that box when they buy the ticket, I'll bet Ryan Air will be more than happy to supply a paper ticket for an extra $75 fee per boarding pass (or per leg of the journey).

And yet, I know French workers that were able to bootstrap their own businesses in part thanks to the generous severance packages they got from their former employers.

Zoox doesn't have a steering wheel or pedals. You do not have control of it (except for the emergency red stop button).

Right now, Zoox can not even deviate from a hardcoded preprogrammed route, so it's a long ways to be true self-driving either.

Notice how the routes it does in San Francisco or in Vegas are always the same circuits.

Getting a DUI on it wouldn't make any sense.

You're thinking too small.

If we can do this with a plane, we can do this with a continuous giant ski lift (or a sky lift).

You're not wrong, but you should know that Copilot is behind the foundational models.

And using Copilot as a benchmark for anything isn't an accurate representation of what's possible right now.