Hank Green

[Hank Green @hankgreen]

Original tweet deleted...you can still see it here along with a post on how I'm changing things going forward.

https://t.co/Ixfcxv0XTF

[Hank Green @hankgreen]

@MrGartin @callmemeghna The annoying “_____ WORD” categories are consistently the third or fourth categories guessed! People just like getting the puzzle!! Even I do that when I’m tired.

[Hank Green @hankgreen]

Everyone in the world who has been confused by the difficulty rankings from 4x3 for the last week or so... I created this very clever system where it updates the colors based on how difficult people actually find the puzzles, except I screwed up and assigned the colors wrong.

Blue has been the easiest category for *days*.

I knew something seemed off but was too busy to investigate. Today though it was like "WELL THAT JUST CANT BE RIGHT!??!"

Tomorrow it will be fixed!

[Hank Green @hankgreen]

@transgendererer Do not lick the subway salt.

[Hank Green @hankgreen]

I’ve been approaching the last 60 days or so with a real “can do” attitude that is frankly not appreciated by my current or future self.

[Hank Green @hankgreen]

@callmemeghna No, you are a fairly small minority!

[Hank Green @hankgreen]

@theo More people than play my dumb browser games wtf.

[Hank Green @hankgreen]

@RRanney1 4x6 (assuming you don’t mean 4x3) is on a weekly difficulty ladder, so Monday is easier and Sunday is hardest.

[Hank Green @hankgreen]

@ramez Clears up like 3 distinct confusing phenomena in 700 words. Wild that it has not travelled farther.

[Hank Green @hankgreen]

Oh… OH!!

[Arvind Narayanan @random_walker]

To understand and empathize with how workers in many or most fields outside software experience advances in AI capabilities, I propose a little thought experiment. https://t.co/QZG24aRiBe

[Hank Green @hankgreen]

My uncertainty is so high that it seems very difficult to imagine that I'd be able to predict the precise thing to do to lower existential risk...

Sitting doing nothing might do that...or not. Telling people that theirchildren are going to die might do that...or not! Working the problems that definitely exist right now might do that...or not.

So I think the thing to do is the last thing because at least then I'm working problems.

[Hank Green @hankgreen]

Well, I mean, the framers of the Constitution didn’t /say/ that compressing the sum of human culture into high-dimensional vector space wasn’t fair use, so…

[Hank Green @hankgreen]

Sorry, not at all saying that the probability is zero...I think it's weird that people so rarely give ranges and, when they do, the /ranges/ do not include zero.

All data in science has a number and a CI. I need CIs and, when the unknowns are very unknown, I would expect CIs to include zero. There is a very good chance that one of the things on the causality chain to x-risk is impossible, so if you multiply the chain of percentages...you get zero.

[Hank Green @hankgreen]

@jonathanpassey @BenBackus4 @technollama Interesting. Do you think that the value of a car exists solely in the car? Or is there some value to the gasoline?

[Hank Green @hankgreen]

The ship is made of a material we do not understand floating in liquid we do not understand headed into a landscape of icebergs we do not understand and indeed may not exist.

We have no idea what might keep it afloat...might as well start with cleaning the deck.

(also, I don't think the activation energy is high! I have seen humans obsessed with near-term human extinction everywhere for my entire life. It is an amazing way of feeling like you matter, which is like our number one priority.)

[Hank Green @hankgreen]

Might as well dig this one out from a deeper thread...curious if people have thoughts.

Re: the AI industry's focus existential threats.

[Hank Green @hankgreen]

I place different weights on things that might happen and things that definitely are happening. One thing that bugs me about existential-risk debates is that history is full of people at the forefront of genuinely important ideas who could not tell the difference between “this is a very big deal” and “this is the end of everything.”

AI may pose existential risks. But it has already enabled an enormous transfer of value and power away from the people whose work made these systems possible.

Too easy to let a possible future tragedy eclipse current harms. Too easy to think "WE ARE BECOME DEATH" because that feels horrifying but also kinda awesome.

It's important to focus on reality because otherwise the odds that you end up in a cult get really high!

[Hank Green @hankgreen]

@GameChangerSam I was really hoping this one would make it in.

[Hank Green @hankgreen]

Thanks for pushing me for more concrete language. It’s not just that it’s never happened before. It’s that when there are no physical systems to test and no previous examples for the probabilities to rest on, assigning numbers to vibes doesn't make them not vibes.

Humans have quite a history of finding the end of the world over-plausible. That makes me suspicious of people who hear "there are harms being done right now that we should be focusing on right now" and reply with "It's more important to focus on x-risk"

[Hank Green @hankgreen]

I'm glad to be talking to you about this...it seems to me that you're listing things that have happened before...where there is data. How this transition will go ecnomically, I can see forecasting that. Predicting superintelligence enslaving humans seems very "human brain got stuck on a cool story" to me and when I see people's probability chains they read like "if/then" chains in a way that does not reflect how reality works!!

A big problem I have with people's pDooms is that the range never includes a zero. There are zero-probability events! Things that literally can't happen. Lots of them.

[Hank Green @hankgreen]

@BenBackus4 @NathanpmYoung @technollama I think the burden of proof of “I can give the probability of an outcome that has never happened before” probably lies with that claim.

[Hank Green @hankgreen]

@NathanpmYoung @BenBackus4 @technollama I find it confusing that people know the odds of something that is, to me, very obviously unknowable. This is not how odds work.

[payton mitchell @paytmitch]

“being annoying on twitter will never get you anywhere”

okay then explain this