Hey, I'm Owen. I think rationality is pretty rad.


Instrumental Rationality

Wiki Contributions


Grammar: I think there's a few words missing in:

there is real value on the line here and this is a real trust-building exercise that [is not?] undertaken lightly by either LessWrong or the EA Forum.

As a defi user / Ethereum dev, there are two things that seem relevant:

  1. Many of the use-cases for which Kleros seems good for (e.g. arbitrating social contracts, prediction markets, etc.) is not where most of the capital is on Ethereum. Most capital is engaged in various overcollateralized lending/borrowing, LP, derivatives, synthetics, etc. So until we see additional applications which run on the social layer, I don't think the utility of Kleros will be made that apparent.

  2. Partially related to the above, the lack of social-based applications I think has to do with a combination of generally poor UX and high latency for Ethereum. We're seeing some exciting updates on this front with rollup solutions like Optimism and zkSync which can provide ~10x increases to start, but again, the ecosystem hasn't really caught up.

More broadly, I think that attention in crypto markets shifts often, and valuations are kind of all over the place. I think I believe that scalable protocols for dispute arbitration if done well can be worth several multiples of Kleros' market cap, but also it's unclear if Kleros is that specific protocol? I personally need to dive into this more.

ah yes, the proof of stake bridge is faster.

i guess it depends if you're running this strategy with size. e.g. for over $100,000, 10% returns means you'd earn back gas fees in ~3 days.

fyi you can get around half these returns on aave on ethereum mainnet without having to mess with matic at all.

while i don't think the matic team is untrustworthy, it's worth pointing out their entire network is currently secured by an upgradeable multisig wallet.

there is also a ~1 week period to move back from matic to ethereum mainnet which can be irksome if you e.g. want to sell quickly back to fiat via some centralized exchange.

Just chiming in here to say that I completely forgot about Intercom during this entire series of events, and I wish I had remembered/used it earlier.

(I disabled the button a long time ago, and it has been literal years since I used it last.)

Hi Rohin! Thanks for this summary of our post. I think one other sub-field that has seen a lot of progress is in creating somewhat competitive models that are inherently more interpretable (i.e. a lot of the augmented/approximate decision tree models), as well as some of the decision set stuff. Otherwise, I think it's a fair assessment, will also link this comment to Peter so he can chime in with any suggested clarifications of our opinions, if any. Cheers, Owen

Ah, I didn't mean to ask about the designing part, but moreso about how you use the word optimize in your definition when it comes to 'optimizing from scratch', which might get a little recursive.

Your definition of optimizer uses "optimizing that function from scratch" which might need some more unpacking.

You may be interested in this prior discussion on optimization which shares some things with your definition but takes a more control theory / systems perspective.

I have not read the book, perhaps Peter has.

A quick look at the table of contents suggests that it's focused more on model-agnostic methods. I think you'd get a different overview of the field compared to the papers we've summarized here, as an fyi.

I think one large area you'd miss out on from reading the book is the recent work on making neural nets more interpretable, or designing more interpretable neural net architectures (e.g. NBDT).

Load More