Theo Diamandis @theo_diamandis Twitter profile

Pinned Tweet

Theo Diamandis

1 month

What do optimal power flow, wireless network resource allocation, market clearing, and exchange order routing all have in common? These are all convex flow problems that can be solved very quickly! (New paper with @GuilleAngeris and Alan Edelman)

4

11

100

Last Seen Profiles

@Kakhu4

@ArkingKyle

@alittlealittle

@WonderTheMovie

@WilliamBriar

@pplwht

@Farokh36

@pjuvilla

@_prodkv

@EsMrGhost

@hazymelancholia

@ArtWanism

@thomasmkane11

@FaqeehFatma

@djubes60

@SimplyZhouYezi

@Minogue_Alyssa7

@martingommel

@runshowcase

@Leo_Noa_

@ProfRafia

@STI_Bret

@_novariani

@badhipgaga

@Serbehell

@koroku6_

@cameron_cates

@conradodess

@DSaddiqui4300

@igorll2

@BigPichuFan

@NOOKDRAWS

@tsf_customsigns

@RPMComo

@letsgowa

@tCXJivYyhg71964

Theo Diamandis

@theo_diamandis

2 years

How should we allocate and price scarce resources on public blockchains? @GuilleAngeris , @alexhevans , @tarunchitra , & I formalize the multi-dimensional resource pricing problem in blockchains and propose a simple price-setting mechanism!

91

162

436

Theo Diamandis

@theo_diamandis

1 year

Just finished teaching a 4 week convex optimization course during MIT's IAP. All lecture notes are available online.

5

27

211

Theo Diamandis

@theo_diamandis

1 year

How do we implement on-chain routing? In a new paper, @MaxResnick1 , @tarunchitra , @GuilleAngeris , and I provide an efficient algorithm and a number of new insights!

16

31

210

Theo Diamandis

@theo_diamandis

2 years

I hear some of you like routing, so I'm excited to share something @GuilleAngeris and I hacked together!

Introducing CFMMRouter.jl

We created CFMMRouter.jl for convex optimization enthusiasts, twitter anons, and Tarun Chitra to easily find the optimal way to execute trades across a network of decentralized exchanges. In general…

baincapitalcrypto.com

6

28

198

Theo Diamandis

@theo_diamandis

2 years

Convex optimization provides a nice framework to construct custom probability distributions.

4

5

97

Theo Diamandis

@theo_diamandis

2 years

👀

33

7

74

Theo Diamandis

@theo_diamandis

2 years

Frantically googling “how to erase commit history” before my repo goes public

6

4

72

Theo Diamandis

@theo_diamandis

2 years

This didn't make it onto the cvx exam, so @GuilleAngeris and I are putting it here for your mathematical pleasure

4

6

65

Theo Diamandis

@theo_diamandis

15 days

Have been working on network flows which led me down a rabbit hole tracing the history of the max flow problem. The original formulation of this problem wasn’t declassified until 1999 (!), almost 45 years after its formulation

5

10

61

Theo Diamandis

@theo_diamandis

1 year

@GuilleAngeris

3

5

54

Theo Diamandis

@theo_diamandis

2 years

Didn’t feel like reading 40 pages of convex analysis? Check out this blog post summarizing our work on resource pricing

Multi-dimensional On-chain Resource Pricing

Public blockchains allow any user to submit transactions which modify the shared state of the network. These transactions are independently verified and executed by a decentralized network of full…

baincapitalcrypto.com

Theo Diamandis

@theo_diamandis

2 years

How should we allocate and price scarce resources on public blockchains? @GuilleAngeris , @alexhevans , @tarunchitra , & I formalize the multi-dimensional resource pricing problem in blockchains and propose a simple price-setting mechanism!

91

162

436

8

13

48

Theo Diamandis

@theo_diamandis

9 months

Spoiler: “Fair” ordering policies are unfair

A Note on the Welfare Gap in Fair Ordering

Public blockchains group submitted transactions into batches, called blocks. A natural question is how to determine which transactions are included in these batches. In this note, we show a gap...

arxiv.org

Sarah Allen

@sarahalle_

9 months

Kicking off the first session, @theo_diamandis is speaking about ordering policies and welfare

1

2

11

1

4

44

Theo Diamandis

@theo_diamandis

1 year

@GuilleAngeris , @tarunchitra , and I are preparing a tutorial on CFMMs for FC this year! We’ll present a unified view of CFMMs and all their properties through convex geometry.

CFMM'23 tutorial

Tutorial on Constant Function Market Makers at FC 23

fc23.ifca.ai

5

6

43

Theo Diamandis

@theo_diamandis

2 years

We used a neat duality trick to create a super fast routing algorithm.

1

5

43

Theo Diamandis

@theo_diamandis

2 years

All questions will be answered by @hdevalence or my alt @GuilleAngeris . Pls make then difficult.

Tina.Is.Not.Aligned⚡🤖

@tzhen

2 years

#darkmode 3.5/ Convex optimization for fun and profit @theo_diamandis Agenda: Reply w/ your questions in advance!

1

16

1

7

40

Theo Diamandis

@theo_diamandis

2 years

While our work is mostly theoretical, we show that, even in a toy model with two resources (e.g., compute and storage), pricing them separately improves throughput!

4

3

37

Theo Diamandis

@theo_diamandis

2 years

Of course, this utility is unknown (and likely unmeasurable) by the designer, so the problem isn’t useful as stated. Fortunately, duality allows us to decompose this problem, which captures the “ideal end state", into two others, coupled by the resource prices, p

1

0

18

Theo Diamandis

@theo_diamandis

6 months

Our pricing framework includes per-contract pricing. Lots of additional modifications can be made to accommodate system designer preferences For details of per-contract fees, check out the paper!

Integrated Kyle e/acc

@KyleSamani

6 months

Fee markets. The most important innovation in Solana (fee markets are downstream of parallelism) In this talk, @tarunchitra expands on the opportunities to make fee markets even more multi-dimensional than they are today

1

4

46

2

5

33

Theo Diamandis

@theo_diamandis

1 year

1

5

31

Theo Diamandis

@theo_diamandis

1 year

Turns out you can get something pretty similar to @VitalikButerin 's proposed multidimensional EIP1559 rule () using our fees framework. Check out Appendix C in the new version on arXiv!

Theo Diamandis

@theo_diamandis

2 years

How should we allocate and price scarce resources on public blockchains? @GuilleAngeris , @alexhevans , @tarunchitra , & I formalize the multi-dimensional resource pricing problem in blockchains and propose a simple price-setting mechanism!

91

162

436

3

7

31

Theo Diamandis

@theo_diamandis

2 years

We formalize a mechanism for multi-dimensional fee markets, which are a key part of both EIP-4844 and Solana’s upcoming fee redesign and have been explored by @jadler0 , @VitalikButerin , @aeyakovenko , among others

2

1

27

Theo Diamandis

@theo_diamandis

2 years

I compiled a few resources that I've found helpful for writing fast Julia code. Hope this provides a good starting point for others.

0

2

29

Theo Diamandis

@theo_diamandis

2 years

Most blockchains implement 1D fee markets for resources (e.g., the Ethereum gas market). But shoehorning everything into one market doesn’t allow prices to change relative to each other! This limits scalability and enables DoS attacks when resources (e.g., opcodes) are mispriced

2

1

26

Theo Diamandis

@theo_diamandis

9 months

Gotta minimize your Loss Via Reordering

Sarah Allen

@sarahalle_

9 months

@theo_diamandis @kartik1507 @KushalBabel @phildaian @_julianma @malleshpai @0xQuintus Next up @tarunchitra is presenting about quantifying MEV and economic (un)fairness

1

15

0

5

27

Theo Diamandis

@theo_diamandis

2 years

We start with an omniscient network designer who gets to choose which transactions are placed in which block. This decision is a convex optimization problem: maximize the utility of the included transactions, minus the loss to the network, subject to constraints.

1

24

Theo Diamandis

@theo_diamandis

2 years

Check out our paper for the details! And stay tuned for more…

2

1

24

Theo Diamandis

@theo_diamandis

2 years

And of course it's all open source. Happy routing!

GitHub - bcc-research/CFMMRouter.jl: Convex optimization for fun and profit. (Now in Julia!)

Convex optimization for fun and profit. (Now in Julia!) - bcc-research/CFMMRouter.jl

github.com

2

0

27

Theo Diamandis

@theo_diamandis

2 years

Many fractional performance metrics (e.g., the Sharpe ratio = μ / σ) are nonconvex but can still be maximized with convex optimization after a simple transformation

2

27

Theo Diamandis

@theo_diamandis

3 months

Another neat result from this paper: the EIP-1559 controller (as implemented in 4844) is an online mirror descent algorithm

guille

@GuilleAngeris

3 months

Releasing our newest paper on blockchain resource pricing! Work done with @theo_diamandis and @ciamac PDF here:

11

25

134

3

9

26

Theo Diamandis

@theo_diamandis

1 year

CFMMs? I only see convex analysis... Check out our ( @GuilleAngeris , @tarunchitra ) upcoming tutorial at @IFCA_Conference

Theo Diamandis

@theo_diamandis

1 year

@GuilleAngeris , @tarunchitra , and I are preparing a tutorial on CFMMs for FC this year! We’ll present a unified view of CFMMs and all their properties through convex geometry.

5

6

43

0

6

24

Theo Diamandis

@theo_diamandis

4 months

Blockspace allocation is tricky. The supply is roughly fixed. Demand fluctuates wildly. And user sophistication varies quite a bit.

Umbra Research

@UmbraResearch

4 months

TOWARD MULTIDIMENSIONAL SOLANA FEES In this piece, @theo_diamandis , @tarunchitra , and @0xShitTrader explore how multidimensional fee markets can elicit demand signals from the market, and how multidimensional fees lead to better pricing, blockspace utilization, and UX on Solana.

11

36

169

6

8

21

Theo Diamandis

@theo_diamandis

2 years

Maybe we can nerd snipe some others into thinking about this with us? *cough* @MatheusVXF , @barnabemonnot , @Tim_Roughgarden *cough*

2

0

19

Theo Diamandis

@theo_diamandis

2 months

For better or worse, in optimization you usually get what you ask for (what's specified by the objective function). imo, it's important to analyze a mechanism in terms of what it's actually trying to optimize & the algorithm it's using

Péter Szilágyi (karalabe.eth)

@peter_szilagyi

2 months

Apologies for the jab, but this is why "just push the gas up" is IMO not a clear cut scaling solution. Cheaper gas enables dumber shit that will just consume it away *(e.g. my BurntPix thing). IMO we need capacity, then apps, then capacity, then apps. Do it in lockstep.

33

20

259

2

21

Theo Diamandis

@theo_diamandis

2 months

@MIT_CSAIL 2012: Julia (at @mit itself)

Julia (programming language) - Wikipedia

en.wikipedia.org

1

19

Theo Diamandis

@theo_diamandis

8 days

Orthogonal resources should be priced separately (and we have a framework to do so!)

vitalik.eth

@VitalikButerin

8 days

Multidimensional gas pricing:

1K

3K

1

4

20

Theo Diamandis

@theo_diamandis

1 year

TLDR: FIFO inclusion of txns in blocks can lead to a large welfare gap, especially when the txn utility distribution has a heavy tail (very common for MEV opportunities like liquidations!)

guille

@GuilleAngeris

1 year

Today @theo_diamandis and I are releasing a quick note which might be interesting to L1 buildoooors (and fair ordering connoisseurs): "A Note on the Welfare Gap in Fair Ordering"

5

8

43

0

2

20

Theo Diamandis

@theo_diamandis

2 years

And this is only a start! Plenty of additional work to do, including a game theoretic analysis of incentives (we group the miners and users into one entity), an analysis of dynamical behavior, and a more thorough numerical evaluation.

1

0

18

Theo Diamandis

@theo_diamandis

1 year

Last year, @GuilleAngeris and I released CFMMRouter.jl, a basic router implementation for constant product and weighted geometric mean pools. Back then, we left a full exploration of the algorithm for future work.

GitHub - bcc-research/CFMMRouter.jl: Convex optimization for fun and profit. (Now in Julia!)

Convex optimization for fun and profit. (Now in Julia!) - bcc-research/CFMMRouter.jl

github.com

1

0

19

Theo Diamandis

@theo_diamandis

2 years

Curious how to design multi-dimensional fee markets? Check out the talk I gave at Devcon! Video: Slides:

Theo Diamandis - A Mechanism for Pricing Non-fungible Resources: Toward Multi-dimensional Fee...

Pricing Non-fungible Resources Toward Multi-dimensional Fee Markets Theo Diamandis Bain Capital Crypto & MIT Joint work with Alex Evans, Tarun Chitra, Guillermo Angeris

docs.google.com

Theo Diamandis

@theo_diamandis

2 years

How should we allocate and price scarce resources on public blockchains? @GuilleAngeris , @alexhevans , @tarunchitra , & I formalize the multi-dimensional resource pricing problem in blockchains and propose a simple price-setting mechanism!

91

162

436

1

18

Theo Diamandis

@theo_diamandis

9 months

@FrankieIsLost ppl inventing new vocabulary for "optimization problem"

2

18

Theo Diamandis

@theo_diamandis

1 year

Possibly helpful for anyone who wants to know wtf @GuilleAngeris is talking about most of the time.

1

0

17

Theo Diamandis

@theo_diamandis

6 months

Very excited to announce that the DeFi Workshop at Financial Cryptography ( @IFCA_Conference ) will be back! More info at

DeFi'24 workshop

The 4th Workshop on Decentralized Finance (DeFi) 2024

fc24.ifca.ai

DeFi Workshop @Financial Cryptography

@DeFi_FC

6 months

📣 DeFi researchers: submit your work to the 4th Decentralized Finance Workshop @IFCA_Conference by 6 Dec! This year's workshop in Curaçao continues one of the most rigorous venues focused on DeFi research spanning security, economics, and game theory.

0

7

15

1

6

18

Theo Diamandis

@theo_diamandis

2 years

In these markets, each resource can be metered, priced, and constrained separately, providing more fine-grained control. But how do we choose the prices?

1

0

14

Theo Diamandis

@theo_diamandis

10 months

Turns out a market is just a convex set

guille

@GuilleAngeris

10 months

New paper with @theo_diamandis , @tarunchitra , @ks_kulk , and @alexhevans on the geometry of constant function market makers is (finally) out! It’s even got pretty pictures and everything! Check it out at:

7

38

171

3

2

17

Theo Diamandis

@theo_diamandis

15 days

This fun bit of history was brought to my attention by Alexander Schrijver, who requested the declassification of the report and wrote a great historical account of these problems.

1

0

17

Theo Diamandis

@theo_diamandis

2 years

New results on MEV in CFMMs with @ks_kulk and @tarunchitra !! In routing, we show that, counterintuitively, sandwiches can actually improve user welfare in certain cases (an inverse Braess's paradox!). Check out the paper for more

Tarun Chitra

@tarunchitra

2 years

⏰⏰⏰ New Paper Alert ⚠️⚠️⚠️ MEV is always bad for users… right? Not always! @ks_kulk , @theo_diamandis , & I formalize the loss from sandwich attacks and show that sandwichers act as decentralized highway traffic controllers & *improve* network welfare

24

62

304

2

16

Theo Diamandis

@theo_diamandis

2 years

The first problem captures the cost to the network and is simple enough to be solved on chain. The second is a maximal utility problem that miners and users implicitly solve when creating and including transactions in a given block. They do all the work for the network!

1

0

14

Theo Diamandis

@theo_diamandis

1 year

This approach also helps us easily incorporate “aggregate CFMMs,” like @Uniswap v3, into our framework and write an efficient arbitrage function.

2

0

16

Theo Diamandis

@theo_diamandis

2 years

Correctly setting the resource prices aligns incentives such that the resource costs to the network are exactly balanced by the utility gained by the users and miners. This, in turn, leads to block allocations which solve the original “ideal” problem, on average.

2

0

14

Theo Diamandis

@theo_diamandis

1 month

Very nice summary of this work by @ks_kulk , @malleshpai , @tarunchitra , and me

Francesco 🌊

@fulminmaxi

1 month

I finally read the intents paper by @tarunchitra @theo_diamandis (and others) Here's the breakdown👇

5

30

168

0

1

15

Theo Diamandis

@theo_diamandis

2 months

This, of course, formalizes previous observations (h/t @0x94305 )

Tarun Chitra

@tarunchitra

2 months

Three items are behind a wall and a solver is going to get one of them for you Do you get a goat or a car, anon? @malleshpai , @ks_kulk , @theo_diamandis and I show you that if solvers have to do more work to deliver the item to you, they're not going to show up to the auction

31

96

532

1

3

14

Theo Diamandis

@theo_diamandis

1 year

Our new paper has an in-depth derivation of the dual decomposition algorithm we use. Importantly, all you need to route through a CFMM is a black-box "arbitrage" function.

2

0

14

Theo Diamandis

@theo_diamandis

1 year

For more, check out the paper, which we’ll be presenting at @IFCA_Conference in May. Would love to see many of you there!

0

14

Theo Diamandis

@theo_diamandis

1 month

We think there’s a lot more to do here on both the algorithm and theory. Check out our paper for more :)

Convex Network Flows

We introduce a general framework for flow problems over hypergraphs. In our problem formulation, which we call the convex flow problem, we have a concave utility function for the net flow at every...

arxiv.org

1

14

Theo Diamandis

@theo_diamandis

2 years

Amazing team pushing the limits of what DEXs can do. Excited to see @dualityxyz usher in the future of on-chain markets!

Duality Labs

@dualityxyz

2 years

Today, we unveil Duality, the DEX that gives Liquidity Providers superpowers. The next generation exchange is now coming to Cosmos.

58

171

620

0

1

13

Theo Diamandis

@theo_diamandis

9 months

More info on the Gas Inflation Reduction Act of 2023 at

Dynamic Pricing for Non-fungible Resources: Designing...

Public blockchains implement a fee mechanism to allocate scarce computational resources across competing transactions. Most existing fee market designs utilize a joint, fungible unit of account...

arxiv.org

Patrick O'Grady 🔺

@_patrickogrady

9 months

1/ #Avalanche HyperSDK Research Deep Dive: 𝑴𝒖𝒍𝒕𝒊𝒅𝒊𝒎𝒆𝒏𝒔𝒊𝒐𝒏𝒂𝒍 𝑭𝒆𝒆𝒔 Why pay 1 type of fee when you could pay 5 types of fees?

7

75

216

0

3

12

Theo Diamandis

@theo_diamandis

7 months

@tarunchitra @carolinecapital True to strategy, he doubled down on this position the next year

2

1

12

Theo Diamandis

@theo_diamandis

1 year

And, of course, convex analysis is a good approach for Everything (tm)

Tarun Chitra

@tarunchitra

2 years

The conformal bootstrap (which I almost did a PhD on) was invented to try to formalize a solution to the 3D Ising model using geometry — but now the Convex Analysis overlords have taken the beautiful geometry and made a (less aesthetic) SDP my alt @GuilleAngeris always wins :/

3

1

28

2

0

13

Theo Diamandis

@theo_diamandis

15 days

The goal: determine the Soviet Union’s ability to move around soldiers and supplies. They formulated this problem as finding a maximum flow over a graph.

3

1

13

Theo Diamandis

@theo_diamandis

1 year

This presentation gives simplifies the arbitrage problem to a single variable root finding problem.

1

0

12

Theo Diamandis

@theo_diamandis

1 year

We also dive into the mathematics of two asset ''swap pools, which are by far the most common CFMMs today. We write all results in terms of the "forward exchange function."

1

0

12

Theo Diamandis

@theo_diamandis

1 year

@snoopy_mev @typeofemale @GuilleAngeris Like I stopped at real analysis. I just do “engineering math” (heresy, some might say)

2

0

11

Theo Diamandis

@theo_diamandis

2 years

probability bounds ⊆ convex optimization

1

2

10

Theo Diamandis

@theo_diamandis

1 month

and optimal order routing (route courtesy of @bidflood ).

2

11

Theo Diamandis

@theo_diamandis

7 months

“It’s ironic that FL does more for greenhouse gas emissions [with high-speed rail] than people using paper straws” — @tarunchitra

0

2

11

Theo Diamandis

@theo_diamandis

1 year

And it provides an easy-to-check condition for if we use a particular CFMM or not while routing (the "no-trade condition").

1

0

11

Theo Diamandis

@theo_diamandis

2 years

Absolutely love working with and learning from this incredible group of people

guille

@GuilleAngeris

2 years

It's rare that you get to work with people you like so much, but I absolutely adore the @BainCapCrypto team. I wake up every day looking forward to chatting with @alexhevans , @stefancoh , @_weidai , @theo_diamandis , and @chofmann88 about weird and interesting ideas

6

0

38

0

1

11

Theo Diamandis

@theo_diamandis

5 months

DeFi workshop submissions close tomorrow. Hope to see many of you in Curaçao next year!

DeFi Workshop @Financial Cryptography

@DeFi_FC

5 months

Update: the workshop has an extended deadline. Authors have until December 10, 2023 (17.00 UTC) to submit work.

1

2

5

0

2

10

Theo Diamandis

@theo_diamandis

15 days

These bottlenecks were of particular interest to the US, and you can imagine how a graph cut might be implemented by the air force, practically speaking.

1

0

10

Theo Diamandis

@theo_diamandis

1 month

And yes, this is a real transaction executed on chain (h/t @fulminmaxi )

1

0

10

Theo Diamandis

@theo_diamandis

15 days

They used the (heuristic) 'flooding technique' from Boldyreff: greedily push as much flow as possible through the network. This technique naturally found a bottleneck, suggesting a 'minimum cut' (later, the famous max-flow min-cut theorem was proved).

1

0

9

Theo Diamandis

@theo_diamandis

2 years

Important disclosure: I am not a statistician.

1

0

9

Theo Diamandis

@theo_diamandis

1 year

@GuilleAngeris From Dantzig's 1981 report on the history of LP

0

9

Theo Diamandis

@theo_diamandis

7 months

@matt_levine When do traders learn about portfolio theory?

guille

@GuilleAngeris

3 years

@spreekaway @ArthurB @danrobinson @SBF_Alameda @tarunchitra @alexhevans I guess the point here is that there is usually an explicit downside risk constraint in most basic portfolio optimization problems; not to mention there are usually leverage constraints, among many other things, which of course aren't captured by just vanilla EV.

1

3

0

1

9

Theo Diamandis

@theo_diamandis

15 days

The US Air Force tasked T.E. Harris and General F.S. Ross with studying Soviet railway network capacities.

1

0

9

Theo Diamandis

@theo_diamandis

2 years

Concentration bounds = applied duality theory?

Theo Diamandis

@theo_diamandis

2 years

probability bounds ⊆ convex optimization

1

2

10

1

3

8

Theo Diamandis

@theo_diamandis

2 months

For the other online optimization nerds here, we'll write a post on 4844 as follow the regularized leader soon :)

guille

@GuilleAngeris

2 months

hell yeah, happy blob day ! fun fact: it turns out blob pricing is essentially optimal! (thread, of course)

3

7

32

1

8

Theo Diamandis

@theo_diamandis

15 days

The original RAND report from Harris and Ross is here

1

0

8

Theo Diamandis

@theo_diamandis

2 years

Wondering why we bother teaching Chebyshev's inequality anymore

2

1

8

Theo Diamandis

@theo_diamandis

2 years

Counterintuitive results in CFMM routing MEV (now with pictures!)

K Kulkarni

@ks_kulk

2 years

The Inverse Braess Paradox for CFMM MEV: a simple conceptual example (without all the hairiness of CFMM math), for those reading

2

22

127

0

9

Theo Diamandis

@theo_diamandis

2 years

Tons of great talks plus someone babbling about optimization towards the end

@bertcmiller ⚡️🤖

@bertcmiller

2 years

mev-day dark mode recording up: light mode recording still coming

9

22

137

0

1

9

Theo Diamandis

@theo_diamandis

1 year

Excited to talk about multi-dimensional fee markets @cryptoeconlab this afternoon!

Theo Diamandis

@theo_diamandis

2 years

How should we allocate and price scarce resources on public blockchains? @GuilleAngeris , @alexhevans , @tarunchitra , & I formalize the multi-dimensional resource pricing problem in blockchains and propose a simple price-setting mechanism!

91

162

436

0

3

8

Theo Diamandis

@theo_diamandis

2 years

@tarunchitra @llllvvuu @tarunchitra leaking all the SBC alpha

1

0

9

Theo Diamandis

@theo_diamandis

1 year

When a large number of swap pools are available, our router dramatically outperforms a state-of-the-art commercial solver.

1

0

9

Theo Diamandis

@theo_diamandis

15 days

Recently, @GuilleAngeris , Alan, and I formulated a similar framework and algorithm for the nonlinear 'convex flow problem'. This problem is convex (so generic algorithms exist), but exploiting network structure leads to simpler models and a much faster algorithm.

1

0

8

Theo Diamandis

@theo_diamandis

2 years

@GuilleAngeris @laurence_e_day @hdevalence Does minting more phds devalue your degree?

2

0

9

Theo Diamandis

@theo_diamandis

1 year

And, using real data from Uniswap, we show that routing can yield a significant price improvement for large trades.

1

0

8

Theo Diamandis

@theo_diamandis

2 months

and I agree that "just pushing the gas up" is not a clear cut soln

1

0

8

Theo Diamandis

@theo_diamandis

1 month

Network flow problems are everywhere in CS. The flow models are almost always linear, but much of the real world is nonlinear. We thought it was time to revisit the max flow problem (and friends).

1

0

7

Theo Diamandis

@theo_diamandis

1 month

market clearing and Nash-bargaining problems;

1

0

8

Theo Diamandis

@theo_diamandis

2 years

Finally getting around to showing my work... h/t @tarunchitra

1

0

8

Theo Diamandis

@theo_diamandis

3 months

Another day, another front run by a 15 year old Bertsekas technical report

guille

@GuilleAngeris

4 years

Between Bertsimas, Nesterov, and Rockafellar, I'm scooped like twice a week 15 years ago

1

0

10

1

8

Theo Diamandis

@theo_diamandis

9 months

“Fair ordering” was good marketing, I guess

0

2

8

Theo Diamandis

@theo_diamandis

2 years

@hdevalence @GuilleAngeris Are alts an equivalence relation?

1

7

Theo Diamandis

@theo_diamandis

2 years

@tarunchitra @llllvvuu I don’t think there’s a website. Idk. Someone showed it to me on explorecourses

1

0

8

Theo Diamandis

@theo_diamandis

2 years

Enjoyed talking about RandomizedPreconditioners.jl at @iccopt2022 & @JuliaConOrg this past week! If you're interested using or developing these methods, please reach out!

GitHub - tjdiamandis/RandomizedPreconditioners.jl

Contribute to tjdiamandis/RandomizedPreconditioners.jl development by creating an account on GitHub.

github.com

1

0

7

Theo Diamandis

@theo_diamandis

1 month

Each step of the algorithm requires solving subproblems which can be parallelized and often have closed form solutions. These subproblems only require evaluating the Fenchel conjugates of the objective terms and the support functions of the sets of allowable edge flows.

1

0

7

Theo Diamandis

@theo_diamandis

1 month

Of course, you can exchange the nodes for the edges and vice versa in hypergraphs.

1

0

7

Theo Diamandis

@theo_diamandis

1 year

@tarunchitra @GuilleAngeris This is roughly what I took from an undergrad degree in EE

0

7