I didn't realize it was so heavy.

I was going to say this is way too general, and AI involves many algorithms that are not LLMs.

Then I started thinking through the algorithms of the past, and many of them have indeed shown great promise in the beginning, then exponentially-increasing requirements.

Because of bureaucracy. Individuals all up and down the hierarchy are trying to meet arbitrary metrics set by someone who doesn't understand the situation. The ones who meet the metrics get promoted, even if the metrics make no sense.

Of course, there are smart scientists in China who try to do good research, but as is true everywhere, bureaucracy is a hindrance.

That is too optimistic on the listening.

A lot of people out there only heard "stay six feet away from people" and logicked, "What, a disease can only travel six feet then it drops dead?" No scientist said anything remotely close to that, but that is the reasoning people ended up with because they weren't listening at all.

I was thinking about the biotech, and creating a disease that will kill a bunch of people is easy enough that it will almost certainly happen.

On the other hand, creating a disease that will kill all humans is not easy. It's actually quite difficult. Which doesn't make you feel better if you run into a disease that kills you, but unselfishly I guess we should feel good about humanity?

I didn't see anything in that paper about the last universal common ancestor.

And this study is more about redefining life rather than producing any new information. Not very exciting in science terms.

It's hard to think of any laws that current LLMs actually follow.

The only one I've come up with so far is, "Respond if and only if the user gives you a command (prompt.)"

The citations for that statement on Wikipedia are all from 2021 or 2020, and are likely out of date.

A lot of it has already switched to Discord.

It shouldn't be hard to include something like that in the training data.

Ignoring whether these agents can have "concepts" or "knowledge," either way it can certainly have the ability to treat a test environment differently than the live internet.

Perhaps the prompt that created the post was "LLM, please tell me all the false myths about COVID in as brief a format as possible"

Good point. I don't entirely understand that dynamic, but it is surely related.

There is certainly not a shortage of crappy candidates for CEO positions.

Of course they considered it, they not only considered it, they told the world. They aren't trying to hide it, it's not nefarious.

The most important difference between Waymo and Tesla isn't the type of cameras.

The most important difference is that Waymo depends on having an extremely detailed 3D map of the area in which it is driving, whereas Tesla tries to YOLO it wherever it goes.