Posts
Comments
You probably should have said 'yes' when asked if it was AI-written.
This started happening in Hawaii, and to a lesser extent in Arizona. The resolution, apart from reducing net metering subsidies, has been to increased the fixed component of the bill (which pays for the grid connection) and reduce the variable component. My impression is this has been a reasonably effective solution, assuming people don't want to cut their connection entirely.
I agree with you that basically anything in the stock market has much less counterparty risk than that. I disagree with basically all non-trading examples you give.
It's not just the stock market, it's true for the bond market, the derivatives market, the commodities market... financial markets, a category which includes prediction markets, cannot function effectively with counterparty risk anything like 5%.
My sense is around 1/20 Ubers don't show up, or if they show up, fail to do their job in some pretty obvious and clear way.
If the Uber doesn't show up I'm not sure that's counterparty risk: you haven't paid anything, so it seems more like them declining the contract. The equivalent for a prediction market would be if you hit 'buy' and the button didn't work, not for when you have paid the money and then don't get the result taken from you. That's much less bad than if the trade went through and then was settled incorrectly.
I think that's false, at least the statistics on wage theft seemed quite substantial to me. I am kind of confused how to interpret these, but various different studies on Wikipedia suggest wage theft on-average to be around 5%-15% (higher among lower-income workers).
I think those studies have significant methodological flaws, though unfortunately I can't remember the specific issues off the top off my head, so this may not be very convincing to you.
I agree this is true for gas and water (and mostly true for electricity, though PG&E is terrible and Berkeley really has a lot of outages).
According to the first google hit, PG&E said the average customer suffered 255.9 minutes of outage in 2013, which is a lot higher than I expected, but is still only 100*255.9/(60*24*365) = 0.05%
In most domains except the most hardened part of the stock market counterparty risk is generally >5%.
This seems quite wrong to me:
- High Yield Corporate Bond OAS spreads are <5% according to bloomberg, and most of that is economic risk, not "you will get screwed by a change of rules" risk.
- Trades on US stock exchanges almost always succeed, many more 9s than just one.
- If I buy a product in a box in a supermarket the contents of the box match the label >>95% of the time.
- Banks make errors with depositor balances <<5% of the time.
- Most employers manage to pay fortnightly wages on time without missing one or more paycheques per year.
- Once you're seated in an Uber or Taxi they take you to your destination almost all the time.
- Your utility company fulfills its obligations to supply your house >>95% of the time under all but the most extreme circumstances.
- Most employees turn up >95% of non-holiday days, and most students maintain >95% attendance.
A bit dated but have you read Robin's 2007 paper on the subject?
Prediction markets are low volume speculative markets whose prices offer informative forecasts on particular policy topics. Observers worry that traders may attempt to mislead decision makers by manipulating prices. We adapt a Kyle-style market microstructure model to this case, adding a manipulator with an additional quadratic preference regarding the price. In this model, when other traders are uncertain about the manipulator’s target price, the mean target price has no effect on prices, and increases in the variance of the target price can increase average price accuracy, by increasing the returns to informed trading and thereby incentives for traders to become informed.
Yes, sorry for being unclear. I meant to suggest that this argument implied 'accelerate agents and decelerate planners' could be the desirable piece of differential progress.
This post seems like it was quite influential. This is basically a trivial review to allow the post to be voted on.
L'Ésswrong, c'est moi.
I agree in general, but think the force of this is weaker in this specific instance because NonLinear seems like a really small org. Most of the issues raised seem to be associated with in-person work and I would be surprised if NonLinear ever went above 10 in-person employees. So at most this seems like one order of magnitude in difference. Clearly the case is different for major corporations or orgs that directly interact with many more people.
I think there will be some degree to which clearly demonstrating that false accusations were made will ripple out into the social graph naturally (even with the anonymization), and will have consequences. I also think there are some ways to privately reach out to some smaller subset of people who might have a particularly good reason to know about this.
If this is an acceptable resolution, why didn't you just let the problems with NonLinear ripply out into the social graph naturally?
If most firms have these clauses, one firm doesn't, and most people don't understand this, it seems possible that most people would end up with a less accurate impression of their relative merits than if all firms had been subject to equivalent evidence filtering effects.
In particular, it seems like this might matter for Wave if most of their hiring is from non-EA/LW people who are comparing them against random other normal companies.
Sorry, not for 2022.
I would typically aim for mid-December, in time for the American charitable giving season.
After having written an annual review of AI safety organisations for six years, I intend to stop this year. I'm sharing this in case someone else wanted to in my stead.
Reasons
- It is very time consuming and I am busy.
- I have a lot of conflicts of interests now.
- The space is much better funded by large donors than when I started. As a small donor, it seems like you either donate to:
- A large org that OP/FTX/etc. support, in which case funging is ~ total and you can probably just support any.
- A large org than OP/FTX/etc. reject in which case there is a high chance you are wrong.
- A small org OP/FTX/etc. haven't heard of, in which case I probably can't help you either.
- Part of my motivation was to ensure I stayed involved in the community but this is not at threat now.
Hopefully it was helpful to people over the years. If you have any questions feel free to reach out.
Alignment research: 30
Could you share some breakdown for what these people work on? Does this include things like the 'anti-bias' prompt engineering?
I would expect that to be the case for staff who truly support faculty. But many of them seem to be there to directly support students, rather than via faculty. The number of student mental health coordinators (and so on) you need doesn't scale with the number of faculty you have. The largest increase in this category is 'student services', which seems to be definitely of this nature.
Thanks very much for writing this very diligent analysis.
I think you do a good job of analyzing the student/faculty ratio, but unless I have misread it seems like this is only about half the answer. 'Support' expenses rose by even more than 'Instruction', and the former category seems less linked to the diversity of courses offered than to things like the proliferation of Deans, student welfare initiatives, fancy buildings, etc.
Thanks, that's very kind of you!
Is your argument about personnel overlap that one could do some sort of mixed effect regression, with location as the primary independent variable and controls for individual productivity? If so I'm so somewhat skeptical about the tractability: the sample size is not that big, the data seems messy, and I'm not sure it would capture necessarily the fundamental thing we care about. I'd be interested in the results if you wanted to give it a go though!
More importantly, I'm not sure this analysis would be that useful. Geography-based-priors only really seem useful for factors we can't directly observe; for an organization like CHAI our direct observations will almost entirely screen off this prior. The prior is only really important for factors where direct measurement is difficult, and hence we can't update away from the prior, but for those we can't do the regression. (Though I guess we could do the regression on known firms/researchers and extrapolate to new unknown orgs/individuals).
The way this plays out here is we've already spent the vast majority of the article examining the research productivity of the organizations; geography based priors only matter insomuchas you think they can proxy for something else that is not captured in this.
As befits this being a somewhat secondary factor, it's worth noting that I think (though I haven't explicitly checked) in the past I have supported bay area organisations more than non-bay-area ones.
Thanks, fixed in both copies.
Thanks, fixed.
Should be fixed, thanks.
Changed in both copies as you request.
- I prioritized posts by named organizations.
- Diffractor does not list any institutional affiliations on his user page.
- No institution I noticed listed the post/sequence on their 'research' page.
- No institution I contacted mentioned the post/sequence.
- No post in the sequence was that high in the list of 2021 Alignment Forum posts, sorted by karma.
- Several other filtering methods also did not identify the post
However upon reflection it does seem to be MIRI-affiliated so perhaps should have been affiliated; if I have time I may review and edit it in later.
13 years later: did anyone end up actually making such a book?
The labels on the life satisfaction chart appear to be wrong; January 2021 comes before December 2020.
Well, with hemispherectomy, those problems are no more. Hemispherectomy is a procedure where half of the brain is removed. It has been performed multiple times without any apparent complications (example).
I was skeptical until I read the example. Now I am convinced!
It's hard to sell 1 million eggs for one price, and 1 million for another price.
Are you sure this is the case? It's common for B2B transactions to feature highly customised and secret pricing and discounts. And in this case they're not selling the same product from the customer's point of view: one buyer gets a million ethical eggs, while another gets a million ordinary (from their point of view) eggs.
Thanks for writing this; ordered.
A teacher in year 9 walked up to a student who was talking, picked them up and threw them out of an (open) first floor window.
Worth clarifying for US readers that 'first floor' in the UK would be 'second floor' in the US, because UK floor indexing starts at zero. So this event is much worse than it sounds.
Thanks, added.
At the moment, the poor person and the rich person are both buying things. If the rich person buys more vaccine, that means they will buy less of the other things, so the poor person will be able to have more of them. So the question is about the ratios of how much the two guys care about the vaccine and how much they care about the other thing... and the answer is the rich guy will pay up for the vaccine when his vaccine:other ratio is higher than the other guys. This is the efficient allocation.
It might be the case that it is separately desirable to redistribute wealth from the rich guy to the poor guy. This would indeed allow the poor guy to buy more things. But, conditional on a certain wealth distribution, it is best to allow market forces to allocate goods within that distribution.
(For simplicity I have ignored macroeconomics in this post, but the same argument broadly goes through if you don't.)
Hey Daniel, thanks very much for the comment. In my database I have you down as class of 2020, hence out of scope for that analysis, which was class of 2018 only. I didn't include 2019 or 2020 classes because I figured it takes time to find your footing, do research, write it up etc., so absence of evidence would not be very strong evidence of absence. So please don't consider this as any reflection on you. Ironically I actually did review one of your papers in the above - this one - which I did indeed think was pretty relevant! (Cntrl-F 'Hendrycks' to find the paragraph in the article). Sorry if this was not clear from the text.
Larks, excellent name choice for your AttackBot.
Thanks! I figured it was in the spirit of a DefectBot to defect linguistically as well, and there was a tiny chance someone might be doing naive string-matching.
You will have to wait for next time's obituary I'm afraid! I think Isusr should have a good grasp on the philosophical and ethical traditions I was attempting to channel with CooperateBot - while the insights are deep, I think the lengthy code is quite clear on the matter.
I actually have no idea - I guess we are just two naturally very cooperative people!
Cool competition! It makes me wish I had had more time to put into CooperateBot. At present I would say it instantiated a relatively naive view of cooperation, and could do much better if I invested more time considering the true nature of generosity. Looking at the obituary I suspect that CooperateBot may not last much longer.
Holding constant the total amount of taxes you pay, it is better not to get a refund. This is the perspective you should take at the beginning of the year.
Holding constant the amount of taxes you have already paid, it is better to get a refund. This is the perspective you should take at the end of the year.
I attempted to produce a rough estimate of this here (excerpted below):
... One (BERI funded!) study suggested that banning large gatherings reduced r0 by around 28%.
Unfortunately, protests seem in many ways ideal for spreading the disease. They involve a large number of people in a relatively small area for an extended period of time. Even protests which were advertised as being socially distanced often do not end up that way. While many people wear masks, photos of protests make clear that many do not, and those that are are often using cloth masks that are significantly less effective than surgical or n95s in the face of repeated exposure. Additionally, protests often involve people shouting or chanting, which cause infectious droplets to be released from people's mouths. Exposure to tear gas can apparently also increase susceptibility, as well as cramped indoor conditions for those arrested.
It's hard to estimate how many new cases will be caused by the protests, because there doesn’t seem to be good statistics on the number of people at protests, so we can't model the physical dynamics easily. A simple method would be to assume we have lost the benefits of the ban on large gatherings over the last week or so. On the one hand, this may be an over-estimate, because fortunately most people continue to socially distance, and protests take place mainly outside. On the other hand, protesters are actively seeking out (encouraging others to seek out) boisterous large gatherings in a way they were not pre-March, which could make things even worse. On net I suspect it may under-estimate the incremental spread, but given the paucity of other statistics we will use it as our central scenario.
If the r0 was around 0.9 before, this suggests the protests might have temporarily increased it to around 1.25, and hopefully it will quickly return to 0.9 after the protests end. Even if we assume no chain infections during the protest - so no-one who has been infected at a protest goes on to infect another protester - this means the next step in disease prevalence would be a 25% increase instead of a 10% decrease. Unfortunately the exponential nature of infection means this will have a large impact. If you assume around 1% of the US was infected previously, had we stayed on the previous r0=0.9 we would end up with around 9% more of the population infected from here on before the disease was fully suppressed. In contrast, with this one-time step-up in r0, we will see around 12.5% of the population infected from here - an additional 3.5% of the population.
Assuming an IFR of around 0.66%, that's a change from around 190,000 deaths to more like 265,000. Protesters skew younger than average, suggesting that this IFR may be an over-estimate, but on the other hand, they are also disproportionately African American, who seem to be more susceptible to the disease, and the people they go on to infect will include older people.
See next year's post here.
I still found this helpful as it allowed me to exit my directional Yang and Buttigieg positions with negative transaction cost.
I would like to add that I think this is bad (and have the codes). We are trying to build social norms around not destroying the world; you are blithely defecting against that.
This case is more complicated than the corporate cases because the powerful person (me) was getting merely the appearance of what she wanted (a genuine relationship with a compatible person), not the real thing. And because the exploited party was either me or Connor, not a third party like bank customers. No one thinks the Wells Fargo CEO was a victim the way I arguably was.
I think you have misunderstood the Wells Fargo case. These fake accounts generally didn't bring in any material revenue; they were just about internal 'number of new accounts targets'. It was directly a case of bank employees being incentivised to defraud management and investors, which they then did. If ordinary Wells employees had not behaved fraudulently, all the targets would have been missed, informing management/investors about their mis-calibration, and more appropriate targets would have been set. In this case power didn't buy distance from the crime, but only in the sense that it meant you couldn't tell you were being cheated.
For more on this I recommend the prolific Matt Levine:
There's a standard story in most bank scandals, in which small groups of highly paid traders gleefully and ungrammatically conspire to rip-off customers and make a lot of money for themselves and their bank. This isn't that. This looks more like a vast uprising of low-paid and ill-treated Wells Fargo employees against their bosses.
...
So that's about 2.1 million fake deposit and credit-card accounts, of which about 100,000 -- fewer than 5 percent -- brought in any fee income to Wells Fargo. The total fee income was $2.4 million, or about $1.14 per fake account. And that overstates the profitability: Wells Fargo also enrolled people for debit cards and online banking, but the CFPB doesn't bother to count those incidents, or suggest that any of them led to any fees. Which makes sense: You'd expect online banking and debit cards to be free, if you never use them or even know about them. Meanwhile, all this dumb stuff seems to have occupied huge amounts of employee time that could have been spent on more productive activities. If you divide the $2.4 million among the 5,300 employees fired for setting up fake accounts, you get about $450 per employee. Presumably it cost Wells Fargo way more than that just to replace them.
In the abstract, you can see why Wells Fargo would emphasize cross-selling of multiple "solutions" to customers. It is a good sales practice; it both indicates and encourages customer loyalty. If your customers have a checking account, and a savings account, and a credit card and online banking, all in one place, then they'll probably use each of those products more than if they had only one. And when they want a new, lucrative product -- a mortgage, say, or investment advice -- they're more likely to turn to the bank where they keep the rest of their financial life.
But obviously no one in senior management wanted this. Signing customers up for online banking without telling them about it doesn't help Wells Fargo at all. No one feels extra loyalty because they have a banking product that they don't use or know about. Even signing them up for a credit card without telling them about it generally doesn't help Wells Fargo, because people don't use credit cards that they don't know about. Cards with an annual fee are a different story -- at least you can charge them the fee! -- but it seems like customers weren't signed up for many of those. This isn't a case of management pushing for something profitable and getting what they asked for, albeit in a regrettable and illegal way. This is a case of management pushing for something profitable but difficult, and the workers pushing back with something worthless but easy.
Is this very different from founding a pharmaceutical company?
Critch wrote a related paper:
Existing multi-objective reinforcement learning (MORL) algorithms do not account for objectives that arise from players with differing beliefs.Concretely, consider two players with different beliefs and utility functions who may cooperate to build a machine that takes actions on their behalf. A representation is needed for how much the machine’s policy will prioritize each player’s interests over time. Assuming the players have reached common knowledge of their situation, this paper derives a recursion that any Pareto optimal policy must satisfy. Two qualitative observations can be made from the recursion: the machine must (1) use each player’s own beliefs in evaluating how well an action will serve that player’s utility function, and (2) shift the relative priority it assigns to each player’s expected utilities over time, by a factor proportional to how well that player’s beliefs predict the machine’s inputs. Observation (2) represents a substantial divergence from naive linear utility aggregation (as in Harsanyi’s utilitarian theorem, and existing MORL algorithms), which is shown here to be inadequate for Pareto optimal sequential decision-making on behalf of players with different beliefs.
War only happens if two agents don’t have common knowledge about who would win (otherwise they’d agree to skip the costs of war).
They might also have poorly aligned incentives, like a war between two countries that allows both governments to gain power and prestige, at the cost of destruction that is borne by the ordinary people of both countries. But this sort of principle-agent problem also seems like something AIs should be better at dealing with.
In light of this:
Build over-communication into the process.
In particular, don’t let silence carry information. Silence can be interpreted a million different ways (Cramton 2001).
Thanks for writing this! I found it very interesting, and I like the style. I particularly hadn't properly appreciated how semi-distributed was worth than either extreme. It's disappointing to hear, but seemingly obvious in retrospect and good to know.