I'm writing a book about epistemology. It's about The Problem of the Criterion, why it's important, and what it has to tell us about how we approach knowing the truth.
I've also written a lot about AI safety. Some of the more interesting stuff can be found at the site of my currently-dormant AI safety org, PAISRI.
Posts
Wiki Contributions
As AI continues to accelerate, the central advice presented in this post to be at peace with doom will become incresingly important to help people stay sane in a world where it may seem like there is no hope. But really there is hope so long as we keep working to avert doom, even if it's not clear how we do that, because we've only truly lost when we stop fighting.
I'd really like to see more follow up on the ideas made in this post. Our drive to care is arguably why we're willing to cooperate, and making AI that cares the same way we do is a potentially viable path to AI aligned with human values, but I've not seen anyone take it up. Regardless, I think this is an important idea and think folks should look at it more closely.
This post makes an easy to digest and compelling case for getting serious about giving up flaws. Many people build their identity around various flaws, and having a post that crisply makes the case that doing so is net bad is helpful to be able to point people at when you see them suffering in this way.
I think this post is important because it brings old insights from cybernetics into a modern frame that relates to how folks are thinking about AI safety today. I strongly suspect that the big idea in this post, that ontology is shaped by usefulness, matters greatly to addressing fundamental problems in AI alignment.
I'm less confident than you are about your opening claim, but I do think it's quite likely that we can figure out how to communicate with orcas. Kudos for just doing things.
I'm not sure how it would fit with their mission, but maybe there's a way you could get funding from EA Funds. It doesn't sound like you need a lot of money.
Completed
The Typical Mind Fallacy is the most important bias in human reasoning.
How do I know? Because it's the one I struggle with the most!
9Back when I tried playing some calibration games, I found I was not able to get successfully calibrated above 95%. At that point I start making errors from things like "misinterpreting the question" or "randomly hit the wrong button" and things like that.
The math is not quite right on this, but from this I've adopted a personal 5% error margin policy, this seems to practically be about the limit of my ability to make accurate predictions, and it's served me well.
What does this mean?
I can't think of a time where such false negatives were a real problem. False positives, in this case, are much more costly, even if the only cost is reputation.
If you never promise anything that could be a problem. Same if you make promises but no one believes them. Being able to make commitments is sometimes really useful, so you need to at least keep live the ability to make and hit commitments so you can use them when needed.