Theo Diamandis

@tjd

22 Following

79 Followers

Tarun Chitra pfp

Paper is out! https://arxiv.org/abs/2403.02525

1 reply

15 recasts

61 reactions

Theo Diamandis pfp

+1: when we cease to understand the world was my favorite read of 2023. (enjoyed the maniac as well)

1 reply

0 recast

2 reactions

Theo Diamandis pfp

The true convex optimizers use concave utility functions

0 reply

1 recast

2 reactions

Theo Diamandis pfp

Convex optimization has concave gains; the losses are convex. Maximizing concave gains is associated with serious side effects and is not recommended by most medical professionals

1 reply

0 recast

2 reactions

Theo Diamandis pfp

Looks like it's up now https://mitbitcoinexpo.github.io/

0 reply

0 recast

2 reactions

Theo Diamandis pfp

Check out the paper for more https://angeris.github.io/papers/oco-fees.pdf

0 reply

0 recast

1 reaction

Theo Diamandis pfp

These results follow from standard analysis techniques in online convex optimization. We suspect there is a lot of interesting future work to be done here.

1 reply

0 recast

1 reaction

Theo Diamandis pfp

Furthermore, we show a corresponding lower bound: in the presence of these adversaries, you cannot achieve better than O(1/sqrt(T)) regret.

1 reply

0 recast

1 reaction

Theo Diamandis pfp

This bound holds _even in the presence of adversaries_ who are only limited by a fixed (but perhaps very large) budget. Market clearing prices need not exist!

1 reply

0 recast

1 reaction

Theo Diamandis pfp

But how good of an idea is this type of simple algorithm? We show that, on average, there is little difference in welfare between “correctly” setting fixed fees with oracular knowledge of all future transactions, and using these basic gradient descent-like algorithms (which only depend on observable quantities).

1 reply

0 recast

1 reaction

Theo Diamandis pfp

In fact, the EIP-1559 style update (as implemented in 4844) is an online mirror descent algorithm.

1 reply

0 recast

1 reaction

Theo Diamandis pfp

Importantly, we can compute the gradient of this function. This means, we can update fees with a simple first-order algorithm like gradient descent, it’s regularized counterpart, or mirror descent.

1 reply

0 recast

1 reaction

Theo Diamandis pfp

While this problem is unsolvable in practice, we can construct an equivalent (dual) problem: find the fees which minimize the difference between the blockspace “supply” of the network and the blockspace “demand” of the users.

1 reply

0 recast

1 reaction

Theo Diamandis pfp

This problem is to maximize the utility of included transactions, minus the loss incurred by the network, subject to transaction constraints (e.g., gas limits, conflicting txns, etc).

1 reply

0 recast

1 reaction

Theo Diamandis pfp

In previous work (also with @pinged, @alexevans) we show that most pricing rules, including EIP-1559, are implicitly solving an optimization problem.

1 reply

0 recast

2 reactions

Theo Diamandis pfp

Simple blockspace pricing mechanisms are a good idea! In a new paper, @gaa, @ciamac, and I show that simple pricing rules are optimal against worst-case adversaries.

1 reply

6 recasts

10 reactions

Mallesh Pai pfp

I'd say more but @pinged already said everything I wanted to say. So for now I'll just add: my casting debut! Paper coming v soon ™️

0 reply

3 recasts

4 reactions

Tarun Chitra pfp

Going to give a *sneak peak* about this paper w/ @ksk and others to FC before Twitter! There's been a surge of interest in intent-based systems like @uniswapX or CoW Swap — their claim has been to help improve the acquisition of off-chain price data on-chain But do they work?

9 replies

22 recasts

88 reactions

Theo Diamandis pfp

Wondering the same thing tbh

0 reply

0 recast

0 reaction

Francesco 🌊 pfp

gm Ever got an arbitrage from a Frame? Now you can 😛 Here's how it works: - Flood detects arbitrages and shows them in the frame - The first to click the button gets the arbitrage to their wallet (on Arbitrum) once the Flood takes it We hacked it in a few hours, so won't be perfect etc... https://frame.flood.bid/

0 reply

6 recasts

13 reactions