How do we (or can we) know adversarial attacks don't work on us?

That's not the point I'm making. So in context of vision we pattern match too, and you can see it in the hundreds of optical illusions online. Our visual system relies on built in shortcuts, and they break in different places than a neural network's do. Image models get fooled by things we would never notice, and we get fooled by things they handle fine. IMO that's good evidence that we and these models use different algorithms to solve similar problems. I think the same is probably true of LLM "reasoning", there's some overlap with how we think, but a lot of it likely works very differently under the hood. Good example of this is anthropic's interpretability research where they found claude doing addition in a way that's nothing like the carry the one method it describes when you ask how it did it.

We're pretty confident that adversarial attacks do work on us, actually—we have entire industries built around it. Some of them even start with "adver-"...

There are classes of adversarial attacks targeting them that don't work as well on us, and vice versa.

Hypnosis and optical illusions are adversarial attacks on shortcuts our brain takes.

That the graph exists at all is a non-neutral statement about that graph versus all other possible graphs that could have been drawn and were not.

You can only claim that if you think nothing can every be neutral, which makes the discussion rather moot.

I don't find ads tedious at all. In fact, since I grew up with ads on tv, I miss them. The commercial break was a scheduled time to get a snack, go to the bathroom, get a snack, etc. Today's platforms put the ad breaks in such awkward spots that doing those things is a lot less natural. IMO, it's not the ads that after the problem; it's the way they are used.

So you don't miss ads, you miss a scheduled break. You literally said you walk away when they happened, so you didn't even see them.

That's fine and there are tools to get that scheduled break back - even the official YouTube app has options to remind you to take a break every X min of watchtime and there are some very advanced general screentime limiting apps out there.

As you say YouTube ads aren't scheduled well and they're too short and unpredictable to even go to the bathroom during. Integrated ad reads are even worse in that regard.

You bring your phone or tablet to the bathroom, or are you playing YouTube on an unmovable TV set?

I did without a TV set since I moved in a larger home. Previously I could listen to the TV anywhere in my home, now I can't so instead of pinning me in a place I moved my TV into my phone and tablets. Everything is in an app anyway and for free to air TV I got a couple of RPi 3 with a TV hat and the TVHeadend app.

And when I have guests everybody is watching something different on their phones most of the times.

But these days you just press pause when you feel like it, so you don't have to let someone else pick the moment for you.

This is a problem that can be solved by just pressing pause, you don't need ads for this.

I use the LLM to point me at relevant books and papers. But there is still a trust problem: I am trusting the LLM to point me at reliable, trustworthy sources.

I'm not sure I'm better off with humans though -- I'm not qualified to judge whether a source is a proper authority, not an I qualified to judge whether someone knows enough to point me to a reliable source.

It seems this is a fundamental epistemological problem to which there may never be an answer.

Can you explain the France paradox?

France imports 0 oil directly through Hormuz, but when the Strait closes, all other countries start raising their safety stock which increases oil prices. This makes France's stockouts very expensive mostly due to price (not flows). Which is contagion without direct connection.

Think of the 2008 crisis where companies that had nothing to do with Credit Default Swaps were exposed to the crash due to network effects, potentially crashing the whole economy.

Then, the government intervened by bailing the banks out. Here, it's not really clear if there is even an effective intervention.