Partner @Arb Research. Former quant trader.
Charles Dillon šø
Karma: 1,633
PacĀing the FronĀtier: A FrameĀwork & ReĀsearch Agenda
EsĀtiĀmatĀing METR Time HoriĀzons for Claude Opus 4.6 and GPT 5.3 Codex (xhigh)
An inĀterĀacĀtive toy model for exĀplorĀing AIās effect on the labour market
Comparative advantage is not at all load bearing here. The only thing that is load bearing is skepticism that AI capabilities will advance sufficiently and be sufficiently cheap.
From a post on my substack @Nathan Young linked below (now on the EA forum)
Comparative advantage alone doesnāt mean you get to eat
This is Arthur Bās point above, and mine when I pointed out that there are fewer horses these days than in the past.It is simply a claim that the market clearing price of labour does not have to be sufficient for people to get as much resources as they want. Comparative advantage tells you which actor should do which task in order to maximise output, not what anyone gets paid.
We can see this even today, without needing to imagine a future very different from our present. Around 20% of people in Africa are undernourished, with 78% (!), or one billion people, unable to afford a healthy diet. In Western countries, there is strong empirical evidence and a clear intuitive case for why e.g. some cognitively disabled individuals have market clearing wages substantially below minimum wage thresholds.
If we take sufficiently capable AI as given, we are all cognitively and physically disabled individuals for the purpose of this conversation. If AI labour is sufficiently cheap, and additional workers are sufficiently easy to spin up, the market clearing price of human labour should theoretically approach 0.
On ComĀparĀaĀtive AdĀvanĀtage & AGI
Note that Matthew responded at length on my substack, and there is further discussion there.
On neĀgoĀtiĀated setĀtleĀments vs conĀflict with misĀalĀigned AGI
If you care about animals I think it is obviously the case that the best bet is to work on finding an angle that Reform will go along with. Given their poll position, the shallowness of their bench and their relatively minimal set of policy commitments, not doing this because they are āproblematicā strikes me as putting low-consequence deontological considerations (you arenāt going to change the probability they win, it seems unlikely anybody cares about your endorsement) over a potentially huge opportunity for impact.
Itās the same thingāif I think the expected value of one thing vs another is 10x, all things considered, then that is what I think the expected value is, already factoring in whatever chance I think there is that I am various versions of wrong, which is very under specified here.
For example, letās say I do a back of the envelope calculation that says ABC is 20x as valuable as XYZ, but I see lots of people disagree with me. Then my estimate of the relative value of ABC vs XYZ will not be 20x, but probably some lower number, which could be 15x or 2x or 0.5x or 0.001x or ā2x even (if it seems ABC is harmful somehow), depending on how uncertain I am and the strength of the evidence provided. That adjustment is already attempting to take into account the possibility of my thought process being bad, my argument being wrong etc.
Personally speaking, if I say I think something is 10x as effective, I mean that as an all-things-considered statement, which includes deferring however much I think it is appropriate to the views of others.
This seems likely to be incorrect to me, at least sometimes. In particular I disagree with the suggestion that the improvement on the margin is likely to be only on the order of 5%.
Letās take someone who moves from donating to global health causes to donating to help animals. Itās very plausible that they may think the difference in effectiveness there is by a factor of 10, or even more.
They may also think that non-EA dollars are more easily persuaded to donate to global health initiatives than animal welfare ones. In this case, if a non-EA dollar is 80% likely to go to global health, and 20% to animal welfare, then by their own lights the change in use of their dollar was more than 3x as important as the introduction of the extra non-EA dollar.
I think if there were proven methods to persuade such people to give away the excess, the world would look very different.
I hope you find success in persuading those around you to give, but I donāt think the process of giving to neighbours and polling resources rather than going directly to supporting specific causes where they could have a lot of impact makes much sense.
Just had another glance at this and I think the delta vs implied vol piece is consistent with something other than a normal/ālog normal distribution. Consider: the price is $13 for the put, and the delta is 5. This implies something likeāthe option is expected to pay off a nonzero amount 5% of the time, but the average payoff when it does is $260 (despite the max payoff definitionally being 450). So it looks like this is really being priced as crash insurance, and the distribution is very non normal (i.e. circumstances where NVDA falls to that price means something weird has happened)
Generally I just wouldnāt trust numbers from Yahoo and think thatās the Occamās Razor explanation here.
Delta is the value I would use before anything else since the link to models of reality is so straightforward (stock moves $1 ā option moves $0.05 ā clearly thatās equivalent to making an extra dollar 5% of the time)
Right now the IV of June 2025 450 calls is 53.7, and of puts 50.9, per Bloomberg. Iāve no idea where your numbers are coming from, but someone is getting the calculation wrong or the input is garbage.
The spread in the above numbers is likely to do with illiquidity and bid ask spreads more than anything profound.
The IV for puts and calls at a given strike and expiry date will be identical, because one can trivially construct a put or a call from the other by trading stock, and the only frictions are the cost of carry.
The best proxy for probability an option will expire in the money is the delta of the option.
āNvidiaās implied volatility is about 60%, which means ā even assuming efficient markets ā it has about a 15% chance of falling more than 50% in a year.
And more speculatively, booms and busts seem more likely for stocks that have gone up a ton, and when new technologies are being introduced.ā
Do you think the people trading the options setting that implied volatility are unaware of this?
Seems like a rather vague collection of barely connected anecdotes haphazardly strung together.
I am not particularly concerned as I donāt see this persuading anybody.
Thanks for posting thisāon a quick read it looks pretty accurate to me and Iāll be glad to have this as a resource to point people to when they seem not to understand exactly why what FTX did was so bad.
You seem to be hung up on proving comparative advantage ānot applyingā for some reason, but there is basically no circumstance where it does not apply, in theory.
It simply doesnāt matter, because human equilibrium wages can still go to zero regardless of comparative advantage still applying, and that matters far more. Comparative advantage is a distractionāthat is the point of this post.