Risk

I didn't have at 22 the skill set that I thought I needed to be successful.

And at the risk of splitting hairs here, again, factories more than maybe manufacturing itself.

So, I mean, being competitive in manufacturing is important if you're going to be in the manufacturing business, needless to say.

But it's not so much that you're a manufacturing nerd.

Whether or not you close, you're going to pay for your environmental inspection report.

And as an individual who's not from a wealthy background, you're looking at the amount that you're on the hook for and the amount that's in your savings.

I think the risk there is... that you end up paying it out of pocket or worse, that you make a bad decision, that you close on something you shouldn't have closed on because you want to pay for all those costs with investor money and then you end up running something that's a dud.

So I felt the pressure rising as we got closer to close.

But some problems are illegible, obscure or hard to understand, or in a common cognitive blind spot, meaning there is a high risk that leaders and policymakers will decide to deploy or allow deployment even if they are not solved.

Of course, this is a spectrum, but I am simplifying it to a binary for ease of exposition.

From an X-risk perspective, working on highly legible safety problems has low or even negative expected value.

Similar to working on AI capabilities, it brings forward the date by which AGI-ASI will be deployed, leaving less time to solve the illegible X safety problems.

In contrast, working on the illegible problems, including by trying to make them more legible, does not have this issue and therefore has a much higher expected value, all else being equal, such as tractability.

Most of the best experiences of my life, there's been some element of danger or risk involved with it.

Sometimes it's an emotional risk, sometimes like a sort of financial risk.

Sometimes it's a physical risk, but every good story, basically when a story is told, Danger is there.

When a book is written, when a movie is made, it's always, you know, there's a kind of, you know, I quit my job and I didn't have anything to fall back on.

And I didn't know where my next paycheck was coming, but I started my own business, you know, and, you know, and everyone's, well, it sounds pretty scary not knowing how you're going to provide for your family.

which seems consistent with the prompt injection they wrote in training.

A bunch of the work the models and I do is making videos about X-Risk.

Claude mostly to one-shots any request in this domain when it's MIRI-related.

Astra can take hours of revisions and I've got to carefully check their work.

And hopefully, you know, hopefully we're able to make a profit.

You can then, you know, do risk assessment for yourself.

If you want to do two or three leg round Robins, if you want to do six or seven leg round Robins, you can pick and choose and do whatever you, Risk works best for you.

So let's just get right into it here, starting with the Chicago Blackhawks at the Tampa Bay Lightning.

I have two picks, both on Conor Bedard here.

It goes way back and I have basically multiple persona.

One is the risks forum stuff which as you hear is ongoing for a long time.

The other is the idea of trying to build trustworthy systems which some of you have been heavily involved in for many years.

So that's next talk in three weeks.

folks in the power industry said, well, we're going to fix that.

We're going to change the algorithm so that can't happen again.

And in my risks archive, I have probably six or eight, nine total blackouts of our entire West Coast, the entire West Coast plus Baja, California, and California.

Vancouver, another East Coast one, and you say,