Posts
Comments
I struggle a bit to remember what ASI is but I'm gonna assume it's Artificial Super Intelligence.
Let's say that that's markedly cleverer than 1 person. So it's capable of running very successful trading strategies or programming extremely well. It's not clear to me that such a being:
- Has been driven towards being agentic, when its creators will prefer something more docile
- Can cooperate well enough with itself to manage some massive secret takeover
- Is competent enough to recursively self improve (and solve the alignment problems that creates)
- Can beat everyone else combined
Feels like what such a being/system might do is just run some terrifically successful trading strategies and gather a lot of resources while frantically avoiding notice/trying to claim it won't take over anything else. Huge public outcry, continuing regulation but maybe after a year it settles to some kind of equilibrium.
Chance of increasing capabilities and then some later jump, but seems plausible to me that that wouldn't happen in one go.
I guess, why is it a problem.
Isn't the case usually that housing is the single greatest factor between a US and UK standard of life? Or do you not agree?
To quote @Ege Erdil attempting to steelman:
there could be an interest rate effect - as interest rates fall, claims on future rents become more expensive so housing prices go up.
I want to try this as a way of argument mapping alongside a community that might use it.
It seems likely that a proper accounting of the argumetns may involve some false statements.
If it goes well I think it could be useful to me and readers, but I guess it will take several iterations.
The UK has a productivity problem, not a housing one. This from bernoulli_defect:
While housing would increase quality of life and luxury, it’s questionable whether it would fix low British productivity in non-housing constrained industries.
Consider how the Bay Area has had huge GDP growth despite housing shortages as people just cram into bedsits
I don't think lw dialogues match how I think, which is in nested bullet points. I sense from how often I see thinking displayed in this nested bullet point way (AI impacts, Kialo, Rootclaim) that many feel similarly.
There isn't a housing shortage. There are more houses than there are households. This from Ian Mulheirn:
No. Back in 1991 there were just over 3.0% more houses than there were households in the UK according to government data. Today, using the ONS’s latest household estimates, there appear to be 5.2% more places to live than there are households that want to live in them. In fact growth in the stock of dwellings appears to have outstripped that of households over the past 50 years or so. This is a strange sort of ‘endemic shortage’.
He shows this graph, showing that households (blue line) have repeatedly undershot expectations (red and green line)
COMMENT THREAD
If you comment anywhere other than here, Nathan will delete your comment.
May I submit more songs anywhere?
This is one of the best I've heard: https://www.udio.com/songs/aALrHWVtRAhExxKTT7HjdE
To me this reads as a person caught in the act of bullying who is trying to wriggle out of it. Fair play for challenging him, yuck at the responses.
I don't like looking at it. Also it's basically the whole article so it feels unnecessary. Something about it narrowing the text.
I'm on a laptop. But I think people should be able to use their phones to visit almost any website.
I would not put the whole thing in quotes. I find it harder to read.
Yeah and a couple of relevant things:
- The time EA sexual abuse article includes 1 person who isn't an EA and a sort of vibe that iirc includes most tech houses in the bay in the heading of "EA". This is inaccurate.
- EA takes a pretty strong stance on sexual harassment. Look at what people are banned from the forum for and scale it up. I've heard about people being banned from events for periods of time for causing serial discomfort. Compare this to Church communities I've been a part of and political communities and this is much stricter.
When he put it all together, he ended up with a different conclusion from what you would get if you just read the abstracts. It was a completely novel piece of work that reviewed this whole evidence base at a level of thoroughness that had never been done before, came out with a conclusion that was different from what you naively would have thought, which concluded his best estimate is that, at current margins, we could cut incarceration and there would be no expected impact on crime. He did all that. Then, he started submitting it to journals. It’s gotten rejected from a large number of journals by now [laughter]. I mean starting with the most prestigious ones and then going to the less.…
Why doesn't OpenPhil found a journal? Feels like they could say it's the journal of last resort initially but it probably would pict up status, especially if it contained only true, useful and relevant things.
I did a quick community poll - Community norms poll (2 mins)
I think it went pretty well. What do you think next steps could/should be?
Here are some points with a lot of agreement.
This is so cool! Are we gonna get k-means clustering on lesswrong at any point? Find different groups of respondents and how they are similar/different.
My current model is that China is a bit further away from invasion than people think. And a recession in the next few years could cripple that ability. People think that all you need for an invasion is ships, but you also need an army navy and air force capable of carrying it out.
I wish there were a subscript button.
I don't think it's appropriate to put personal forecasts in Wiki pages. But yeah manifold metaculus polymarket. Or maybe just link them so they are more resilient.
Yeah that sounds about right. A junior dev who needs to be told to do individual features.
You're hit thi singularity doesn't sound wrong but I'll need to think
Yeah I was trying to use richard's terms.
I also guess that the less training data there is, the less good the AIs will be. So while the maybe be good at setting up a dropshipping website for shoes (a 1 - 10 hour task) they may not be good at alignment research.
To me the singularity is when things are undeniably zooming, or perhaps even have zoomed. New AI tech is coming out daily or perhaps the is even godlike AGI. What do folks think is a reasonable definition?
Do you have any suggestions for actions I can take to make my piece more like daniel's?
How should I represent forecasts in text?
I added a note at the top
Oh interesting, yeah I was more linking them as spaces to disagree get opinions. Originally I didn't put my own values at all but that felt worse. What would you recommend?
edit for clarity
I am not very good at sizes. But I guess that it's gonna keep increasing in price, so yeah, maybe >$5bn (30%).
Ideally I would like to rewrite most LessWrong wiki pages about real world stuff in this style, with forecasts as links.
Funny! I shared screenshots with several people.
I think an issue is that dialogue is often about two very different models of the world clashing. It takes a lot of work for those two to develop a common language and even then it may just be the two of them. Add to that that they may be ill informed, dialogue is just very expensive. I really like dialogue and yet it takes a lot for me to be in an actually truth seeking state about it or to read others.
I think about this framing quite a lot. Is what I say going to lead to people assuming roughly the thing I think even if I'm not precise. So the concept is pretty valuable to me.
I don't know if it was the post that did it, but maybe!
I've used this a bit, but not loads. I prefer fatebook, metaculus and manifold and betting. I don't quite know why I don't use it more, here are some guesses.
- I found the tool kind of hard to use
- It was hard to search for the kind of information that I use to forecast
- Often I would generate priors based on my current state, but those were wrong in strange ways (I knew something happened but after the deadline)
- It wasn't clear that it was helping me to get better versus doing lots of forecasting on other platforms.
This post. It gives some useful context around the bad outcomes of panels.
As a counter example I think the EA community health team does better than I would expect. A huge global community and they are relatively trusted and the community takes responsibility for itself. Like I don't know that I'd say it does well but I think I'd be hard pressed to find an analogous community that has a better such function.
Feels like they have a better batting average than rationalists in dealing with issues like this (I guess at least partly cos events are more central to EA than rationalism). Many other communities seem to have no responsibility at all - many political groups seem rife with issues but noone taking responsibility over them.
I agree overall that this is a really hard problem, though I suppose I am more optimistic than OP.
Yeah I agree a bit here. This document itself is a nice case study in how things can go wrong. There could be a similar one for "don't ask women out"
This well-encapsulates my concern that there really is a lot of power going to a small group of people/value sets and that I don't make my best decisions when I only listen to myself. If I were to run a country I would want a blend of elite values and some blend of everyone's values. I am concerned that a guardian wouldn't do that and that we are encouraging that outcome.
I am not sure I would characterise 2 and 4 like that.
I think I'd say ozzy's criticisms were more like:
- So-Called Top EAs: Nonlinear was focused on status as a metric rather than helping their interns become better at their jobs
- Sunshine Is The Best Disinfectant: Nonlinear says that any org would have this many errors. Ozzy says let's get it all out in the open then
I am trying to learn some information theory.
It feels like the bits of information between 50% and 25% and 50% and 75% should be the same.
But for probability p, the information is -log2(p).
But then the information of .5 -> .25 is 1 bit and but from .5 to .75 is .41 bits. What am I getting wrong?
I would appreciate blogs and youtube videos.
I think that the main benefit of manifold.love is that it allows friends to do some matchmaking, but I'd do that for free and am not really doing it for profit anyway. I guess currently the other main benefit is as a schelling point for the rat community.
How private are the LessWrong votes?
Would you want to do it overall or blog by blog. Seems pretty doable.
What are legal threats like in the real world?
I've been threated with legal action once (by Jay Z's record company for a parody I made) and it felt like a bet. I probably could win if I spent a lot of money, but I didn't have that money and so I took the song down.
continue to believe that most of what Ben relayed from Alice was true
I can believe she is being precise without conveying an accurate picture. I am not sure that I ever thought that alice's account was the most accurate version of events.
Scanned it, seems pretty fair. In our discussion you convinced me pace should probably have given the team a week to respond, unless other information comes out.
Well I guess I can only talk about my takeaways from Ben's article. Like who gets to say what Ben's article really means? I think probably you should see my reading as pretty different to the median reading. I think I can justify that but if I had realised how differently you all read the article I would have said sooner.
You can read my comments at the time, I don't think I considered Nonlinear as cruel or abusive. I guess that I might describe the worst of their behaviour like that, maybe, but people behave within broad ranges.
This matches my experience too. When I initially made pretty milquetoast criticisms here all of my comments went down by ~10.
Have people thought about doing gene editing stuff at proposera? Seems legal there?
No I'm not saying that.
I am saying about halfway between that and "Ben's account holds up".
What specifically is the most grievous error here.
Things I would do dialogues about:
(Note I may change my mind during these discussions but if I do so I will say I have)
- Prediction is the right frame for most things
- Focus on world states not individual predictions
- Betting on wars is underrated
- The UK House of Lords is okay actually
- Immigration should be higher but in a way that doesn't annoy everyone and cause backlash
Yeah I mean the answer is, just make prediction markets and bet on them. I think we are getting a lot better at that.
(Also I'm a lesswrong user who makes a lot of prediction markets about AI)
In particular:
- A real money version of Yud and Paul's bet https://polymarket.com/event/will-an-ai-win-the-5-million-ai-math-olympiad-prize-before-august?tid=1702634083181
- An attempt at clustering the best AI progress markets into a dashboard https://manifold.markets/dashboard/ai-progress