Posts
Comments
I think the Conclusion could serve well as an abstract
An abstract which is easier to understand and a couple sentences at each section that explain their general meaning and significance would make this much more accessible
I plan to send the winning proposals from this to as many governing bodies/places that are enacting laws as possible - one country is lined up atm.
Let me know if you have any questions!
options to vary rules/environment/language as well, to see how the alignment generalizes ood. will try this today
it would basically be DnD like.
Making a thing like Papers Please, but as a text adventure, popping an ai agent into that.
Also, could literally just put the ai agent into a text rpg adventure - something like the equivalent of Skyrim, where there are a number of ways to achieve the endgame, level up, etc, both more and less morally. Maybe something like https://www.choiceofgames.com/werewolves-3-evolutions-end/
Will bring it up at the alignment eval hackathon
I see them in o1-preview all the time as well. Also, french occasionally
If developments like this continue, could open weights models be made into a case for not racing? E.g. if everyone's getting access to the weights, what's the point in spending billions to get there 2 weeks earlier?
this can be done more scalably in a text game, no?
People Cannot Handle Gambling on Smartphones
this seems a very strange way to say "Smartphone Gambling is Unhealthy"
It's like saying "People's Lungs Cannot Handle Cigarettes"
To be a bit less useless - I think this fundamentally misses the problem of respect and actually being able to communicate with yourself and fully do things, if you've done so - and that you can do these when you have full faith and respect in yourself (meaning all of yourself - may include love as well, not sure how necessary that is for this). Could maybe be done in other ways as well, but I find those less beautiful, personally.
I think this is really along the wrong path and misunderstanding a lot of things, but so far along the incorrect path of thought and misunderstanding so much, that it's hard to untangle
I thought this was going to be an allegory for interpretability.
give better names to actual formal math things, jesus christ.
I think posts like this are net harmful, by discouraging people from joining those doing good things without providing an alternative and so wasting energy on meaningless ruminating that doesn't culminate in any useful action.
oh, sorry, I thought slatestar codex wrote something about it and you were saying that's where it comes from
I pretty much agree. I prefer rigid definitions because they're less ambiguous to test and more robust to deception. And this field has a lot of deception.
Yup, those are hard. Was just thinking of a definition for the alignment problem, since I've not really seen any good ones.
what do you think of replit agent, stack blitz, etc?
damn, those prices are wild
used before, e.g. Feynman: https://calteches.library.caltech.edu/51/2/CargoCult.htm
btw, thoughts on this for 'the alignment problem'?
"A robust, generalizable, scalable, method to make an AI model which will do set [A] of things as much as it can and not do set [B] of things as much as it can, where you can freely change [A] and [B]"
Unfortunately this is a fundamental problem of Media, imo.
Yes, this would be very very good. I might hold a hackathon/ideathon for this in January.
I didn't get the premise, no. I got that it was before a lot of physics was known, didn't know they didn't know calculus either.
Just stating it plainly and clearly at the start would have been good. Even with that premise, I still find it very annoying. I despise the refusal to speak clearly, the way it's constantly dancing around the bush, not saying the actual point, to me this is pretty obviously because the actual point is a nothing burger(because the analogy is bad) and by dancing around it, the text is trying to distract me and convince me of the point before I realize how dumb it is.
Why the analogy is bad: rocket flights can be tested and simulated much more easily than a superintelligence, with a lot less risk
Analogies are by nature lossy, this one is especially so.
personally, I found how Beth just kept saying 'not really' and not saying the actual physics very very annoying.
Yup, I think research that studies the effect of recommendation algorithms on the brain, from various social media platforms and compares them to the effects of narcotics, would be extremely useful.
I think we're really really lacking in decent legislation for recommendation algorithms atm - at the absolute bare minimum, platforms which use very addictive algorithms should have some kind of warning label informing users of the possibility of addiction - similarly to cigarettes - so that parents know clearly what might happen to their children.
This is going to be even more important as things like character.ai grow.
rather than this, there should just be a better karma system, imo.
one way to improve it - have the voting buttons for comments be on the opposite side of the username
This is very useful, thank you.
Something that might be interesting to add at the end of surveys such as these:
"How much has this survey changed your mind on things?" - sometimes just being asked a question about something can change your mind on it, would be interesting to see if it happens and how much so.
Clickbait still works here, just with a different language.
Cons: Humans are opaque. Even from our inside view, it is very difficult to understand how they work, and very hard to modify. They are also the most difficult to talk about rigorously. There is also the failure mode of anthropomorphizing badly and attributing arbitrary properties of humans (and especially human goals) to AGI.
I don't think it's really correct to say that humans are opaque from an inside view, especially for people with high empathy. People who understand themselves well and have high empathy can very consistently predict and understand others.
Pretty much all of those reasons - what it's missing is that nicotine itself may also be a carcinogen- at least, it has the ability to be one: https://link.springer.com/article/10.1007/s10311-023-01668-1
Although there aren't enough isolated studies done on nicotine in a long period to be conclusive: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5020336/
Some reviews disagree: https://pubmed.ncbi.nlm.nih.gov/26380225/
I strongly advise against taking nicotine.
Eliezer is extremely skilled at capturing attention. One of the best I've seen, outside of presidents and some VCs.
However, as far as I've seen, he's terrible at getting people to do what he wants.
Which means that he has a tendency to attract people to a topic he thinks is important but they never do what he thinks should be done- which seems to lead to a feeling of despondence.
This is where he really differs from those VCs and presidents- they're usually far more balanced.
For an example of an absolute genius in getting people to do what he wants, see Sam Altman.
Unless I'm missing something, this seems to disregard the possibility of deception. Or it handwaves deception away in a line or two.
The type of person to end up as the CEO of a leading AI company is likely (imo) someone very experienced in deception and manipulation- at the very least through experiencing others trying it on them, even if by some ridiculously unlikely chance they haven't used deception to gain power themselves.
A clever, seemingly logically sound argument for them to slow down and trust that their competitor will also slow down because of the argument, will ring all kinds of bells.
I think whistleblower protections, licenses, enforceable charters, mandatory 3rd party safety evals, etc have a much higher chance of working.
Yes, we host a bi-monthly Critique-a-Thon- the next one is from December 16th to 18th!
Judges include:
- Nate Soares, President of MIRI,
- Ramana Kumar, researcher at DeepMind
- Dr Peter S Park, MIT postdoc at the Tegmark lab,
- Charbel-Raphael Segerie, head of the AI unit at EffiSciences.
What about regulations against implementations of known faulty architectures?