Occasionally think about topics discussed here. Will post if I have any thoughts worth sharing.
Tomás B.
Interesting MATH update: https://ai.googleblog.com/2022/06/minerva-solving-quantitative-reasoning.html
What was it that eventually did work for you?
I’m usually the dumbest person in the room and I still post occasionally.
I’m not sure it’s productive to engage with this stuff. Taking a GRAND STAND may feel good, but in many cases people end up becoming useful foils. Block liberally, don’t engage, focus on what actually matters.
And then there were times when even the soft-spoken
Tathagatha listened to the words of his disciple, who had
digested all of the things he had preached, had meditated long
and fully upon them and now, as though he had found entrance
to a secret sea, dipped with his steel-hard hand into places of
hidden waters, and then sprinkled a thing of truth and beauty
upon the heads of the hearers.
Summer passed. There was no doubt now that there were
two who had received enlightenment:
IIRC you were looking into these ideas more seriously, any progress?
I always assumed that “Why don’t we give Terence Tao a million dollars to work on AGI alignment?” was using Tao to refer to a class of people. Your comment implies that it would be especially valuable for Tao specifically to work on it.
When I’ve talked about this, I’ve always meant both literally hire Tao and try to find young people of the same ability.
I would recommend just playing around with a language model. goose.ai/NovelAI are cheap and quite good if you can’t get GPT-3 access.
Anyone who thinks boxing can happen, this thing isn’t AGI, or even an agent really, and it’s already got someone trying to hire a lawyer to represent it. It seems humans do most the work of hacking themselves.
Beauty and the Beast
NeoVim, too. I think PyCharm has it, too.
Regarding the weird mediocrity of modern AI, isn’t part of this that GPT-3-style language models are almost aiming for mediocrity?
Would a hypothetical “AlphaZero of code” which built its own abstractions from the ground up—and presumably would not reinvent Python (AlphaCode is cool and all, but it does strike me as a little absurd to see an AI write Python) - have this property?
Regarding the Einstein stuff, do you think Einstein’s brain had significantly more compute than a 100 IQ person? I would be very surprised if Einstein had more than, I don’t know, twice the computing power of a normal brain. And so the claim that the distance between an idiot and Einstein is far smaller than that between the idiot and a chimpanzee still seems true to me.
I never got around to the scenario-based planning post, but things sure have changed in 10 months!
I still think this is an important question.
100 mg first thing in the morning and no more after is what I do. If I need a stimulant later I use nicotine which has a much shorter half-life.
What is the solution to the Alignment problem?
With the rise of effective appetite suppressants, has this lowered your enthusiasm for the bracelet?
:goose10:
I’m pretty sure I will “win” my bet against him; even two months is a lot of time in AI these days.