LessWrong 2.0 Reader

Perplexity is still my daily driver, due to the superior UI. I go to elicit or you.com for harder problems.

hmm actually, I think I was the one who was wrong on that one. https://en.wikipedia.org/wiki/Synaptic_weight seems to indicate the process I remembered existing doesn't primarily work how I thought it did.

a particular kind of cis women who feel entitled to be extremely rude and intrusive because they assume "women = inherently benign".

Seems to me that some women believe that when they do something, it is fundamentally different from when a man does exactly the same thing. (Something like the fundamental attribution error [? · GW], or xkcd#385 but with reversed genders.) For example, if the woman gets angry and yells at someone, it is because that person was really annoying, or she was tired, etc. Simply, she acted that way because of external reasons. But if she sees a man get angry and yell at someone, it's obvious: men are inherently aggressive. (Or maybe, if she is a good feminist, it's because men are privileged.) This way, she can condemn a certain type of behavior and be really emotional about it... and then go and do exactly the same thing -- because in her mind, it is not the same thing at all.

Or to use the example from the article, men are inherently rude and intrusive; she faced an interesting situation and was naturally curious about it. That's not the same thing at all!

If you SEE a coin flip come up heads (and examine the coin and perform whatever tests you like), what's your posterior probability that the coin actually exists and it wasn't a false memory or trick in some way?

Not enough to make any practical difference to any decision I am going to make. Only when I see the extraordinary evidence required to support an extraordinary hypothesis will it be raised to my attention.

Do you measure the goodness of outcome using current utility function or future utility function?

The current one.

Second example - it is intelligent to care about future instrumental goal.

Only if it is a future instrumental goal that will be used to achieve a current terminal goal.

I think its reasonable for the conversion to be at the original author's discretion rather than an automatic process.

Maybe create a Quotes Thread post with the rule that quotes have to be upvoted and if you like them you can add a react.

When you are considering finite tape length, how do you deal with $x^{(k-1)}$ when $k=0$ or $x^{(k+1)}$ when $k=N$ ?  

I still don't quite get it. We already have an Ilya dusk ever who can make type 1 and type two improvements, and don't see the sort of jump's in days your talking about (I mean, maybe we do, and they just look discontinuous because of the release cycles?)

Thanks for writing this post!

I don't know what the correct definition of AGI is, but to me it seems that AGI is ASI. Imagine an AI that is on super expert level in most (>95%) subjects and that have access to pretty much all human knowledge and is capable of digesting millions of tokens at a time and and can draw inferences and conclusions from that in seconds. "We" normally have a handful of real geniuses per generation. So now a simulated person that is like Stephen Hawkings in Physics, Terrence Tao in Math, Rembrandt in painting etc etc, all at the same time. Now imagine that you have "just" 40.000-100.000 of these simulated persons able to communicate at the speed of light and that can use all the knowledge in the world within millisecond. I think there there will be a very transformative experience for our society from the get go.