←
Akaki Mamageishvili🇺🇦644 posts
Akaki Mamageishvili🇺🇦 avatar✓

Akaki Mamageishvili🇺🇦 ✓

@kakia1989

Research at @Offchain | prev: research @ETH_en | prev: Ph.D. in CS @ETH_en

◷ Joined December 2012▧ 8 Photos and videos
Akaki Mamageishvili🇺🇦 avatar
Akaki Mamageishvili🇺🇦@kakia1989·Nov 10

In a joint work with @pranaytej and @EdFelten (both from @Offchain), we study the effects of block time reduction from 2s to 200ms due to the introduction of Flashblocks, on the Base chain: arxiv.org/abs/2610.03922. We classify searcher and non-searcher addresses, and using difference-in-differences design estimate the fall of effective priority fee per gas. Interesting observation is that searching activity is not uniformly distributed across (Flash-)blocks and 30% of it is concentrated in the first (Flash-)block. Reduced priority fees can be attributed to the less available value per time unit due to shorter effective times. Beyond this, we also develop an auction theoretic framework that qualitatively explains reduced priority fees, in which searchers may bid less hoping that only their transaction makes it in the current (Flash-)block. For a more comprehensive discussion, check out research post by @pranaytej: research.arbitrum.io/t/prior…. Any feedback is welcome.

Whitepaper Reading Club avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Whitepaper Reading Club@WPReadingClub·Nov 17

📺 Why Did Arbitrum Start Selling Speed? (4/8) As @arbitrum moves to PGA, we look back at why TimeBoost sold a 200ms advantage through 60-second auctions. 📝 Main points The speed race: Traders spend heavily on faster servers to beat others by milliseconds. Most of that money never reaches the blockchain. Selling speed: TimeBoost lets traders bid for priority, sending auction revenue to the Arbitrum DAO. The trade-off: Traders still need fast infrastructure. Can auctions reduce the speed race while keeping transaction ordering fast? 🧠 We spoke with @kakia1989 , Senior Research Scientist at @Offchain and co-author of TimeBoost. YouTube: youtu.be/V-6SlbNzNSI #Arbitrum #TimeBoost #MEV #BlockchainEconomics

7:20
Whitepaper Reading Club avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Whitepaper Reading Club@WPReadingClub·Nov 12

📺 Does Paying for Speed Mean You Go First? (3/8) As @arbitrum moves to PGA, we look back at TimeBoost and why its 200ms advantage doesn’t guarantee going first. 📝 Main points 1.⁠ ⁠A head start, not 1st place: Earlier transactions can still go ahead of the winner. 2.⁠ ⁠How it works: The winner keeps their real arrival time. Other TXs get +200ms before both queues are merged. 3.⁠ ⁠Why timing matters: Arbitrum produces blocks every ~250ms. The winner must react within 200ms to beat others chasing the same trade. 🧠 We spoke with @kakia1989 , Senior Research Scientist at @Offchain and co-author of TimeBoost, in January 2026. YouTube: youtu.be/vO-WiAgwAfA #Arbitrum #TimeBoost #TransactionOrdering #MEV #Ethereum

4:05
Ed Felten avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Ed Felten@EdFelten·Nov 3

Cross River is building with @Offchain Enterprise Services Very excited to support their business as one of the biggest banking as a service companies out there. Our team will be working very closely with theirs for the build out of a proprietary tokenization engine, all with settlement to @Arbitrum and wallet infra powered by @ZeroDev The programmable economy is growing, and Offchain promises to be at the center of it.

Post attachment
Whitepaper Reading Club avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Whitepaper Reading Club@WPReadingClub·Oct 28

📺 Why Does the Highest Bidder Pay Less? (2/8) As @arbitrum moves to PGA we look back at TimeBoost, and how they the 60-second auction works. 📝 Main points 1. Pay less than you bid: Bid $100, win, but pay $80 if that’s the 2nd-highest bid (subject to a minimum price). 2. How it works: Traders bid without seeing each other’s offers. Bids close 15 seconds before the next minute starts. 3. Why this design? 2nd-price auctions encourage honest bids when traders value things differently. But what if everyone is chasing the same profits? 🧠 We spoke with @kakia1989 🇺🇦, Senior Research Scientist at @Offchain and co-author of TimeBoost, in January 2026. YouTube: youtu.be/vO-WiAgwAfA Linkedin: linkedin.com/posts/whitepape… #Arbitrum #TimeBoost #AuctionDesign #MEV

2:51
Akaki Mamageishvili🇺🇦 avatar
Akaki Mamageishvili🇺🇦@kakia1989·Oct 11

We study a question of incentivizing attention in optimistic rollup environment, which can also be generalized to wider ecosystems: arxiv.org/abs/2610.12220 There is a natural separation of lightweight (attention) and heavyweight (validation) roles, which is modeled as a double Tullock contest with corresponding reward functions. Players can choose their roles and how many Sybils they want to deploy for each role, each Sybil costing them extra constant depending on the role. For non-decreasing total rewards for attention, we obtain a necessary condition that allows having a target number of attention players. Decreasing total rewards allow arbitrary number of target attention players, but sounds otherwise controversial. The paper is co-authored with @pranaytej,@BenBerg07141321 and Matteo Campanelli (all from @Offchain).

𝗵𝘂𝗻𝘁𝗲𝗿 avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
𝗵𝘂𝗻𝘁𝗲𝗿@bfresh·Sep 28

untapped opportunity here btw it only costs $17 to get early viewing access to @Arbitrum txs before the regular feed sees them currently 16 entities have subscribed, wherein viewing access lasts for ~24 hours (not incl a buffer inbetween the subscription and the actual viewing) blocks are 250ms long and split into two 125ms PGA rounds, so fast feed pushes out txs during those rounds instead of making you wait for the whole block to finish so if you're a searcher, propAMM, or a market maker, this is for you *IMPORTANT NOTE* fast feed doesn't give you access to > viewing the mempool > state changes/partial blocks > track chain state or follow the chain

Post attachment
Fahad Saleh avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Fahad Saleh@cryptoeconprof·Sep 23

🚨Blockchain conference on Oct 29 - 30 @Columbia, Co-organized by CBER Forum and @CUSEAS 🚨 Speakers include @HarryDCrane, @RuizheJia, @_julianma, @malleshpai, @casparschwa, @kakia1989, @EdFelten, @AlvaroCartea, @sui414, @financeUTM, @MaxResnick, @camharvey and @taylor_lindman Topics include prediction markets, stablecoins, quantum and tokenization. Full program and registration are at cglink.me/21U/r376457

danning avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
danning@sui414·Aug 19

Our paper with @kakia1989 and Christoph on Searcher Competition just got accepted to ACM TWEB! This is also my first peer-reviewed journal paper :) The takeaway from studying the Block Building market back then — shown both theoretically and empirically — keeps showing up in every new market structure, including Compute: Competition squeezes margins thin and transfers the profit to the rent collector. dl.acm.org/doi/10.1145/38403…

Akaki Mamageishvili🇺🇦 avatar
Akaki Mamageishvili🇺🇦@kakia1989·Aug 11

4(!) papers co-authored by me were accepted at the Advances in Financial Technologies conference, that will take place at London School of Economics, later this year, October 6-9 (aft.ifca.ai/aft26/index.html). In the first paper, arxiv.org/abs/2606.13555, co-authored with my colleague at @Offchain, @pranaytej, we look into the price elasticity of the blockspace demand for the Ethereum and Arbitrum chains. In the second, arxiv.org/abs/2603.20175, co-authored with @boez95, Christoph Schlegel (from Flashbots) and @AliTslm(from @EntropyAdvisors), we look into the performance of the TimeBoost auction. In the third, arxiv.org/abs/2505.24440, co-authored with Benny Sudakov (from @ETH_en), we study capital efficiency of restaking protocols compared to regular Proof-of-Stake protocols. In the fourth, arxiv.org/abs/2407.14485, co-authored with @0xBrMazoRoig, Christoph Schlegel (from Flashbots) and Minghao Pan (from @Caltech), we look into Sybil proof mechanisms and show that the only dominant strategy incentive compatible and Sybil proof mechanism for selling a single item is the 2nd price auction.

Akaki Mamageishvili🇺🇦 avatar
Akaki Mamageishvili🇺🇦@kakia1989·Jul 10

Our paper on axiomatization of automated market makers and prediction markets has been published by Mathematical Finance journal. It is available online here: onlinelibrary.wiley.com/doi/…. The paper is co-authored with Christoph Schlegel (Flashbots) and Mateusz Kwasnicki (Wroclaw University of Science and Technology). We obtain the constant elasticity market makers to be the class that satisfies two natural axioms. The main axiom is independence, for more than two assets, in which the trade between two assets does not depend on the inventories of other assets. For two assets, we assume separability of a trading curve in functions of asset inventories. The second axiom is scale invariance. We also obtain Logarithmic Market Scoring Rules (LMSR), widely used in prediction markets, as a class of scoring rules that satisfy the same independence/separability axiom and a slightly modified axiom of translation invariance, instead of scale invariance. The main new result since the EC'2023 version of the paper is the optimality of product market makers for the traders, among 2 asset constant elasticity market makers.

Ed Felten avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Ed Felten@EdFelten·Jul 7

The best onchain products will hide complexity from users without giving up control for developers. That is the promise of programmable accounts. ZeroDev Wallet brings embedded wallet UX and smart account infrastructure into one platform. Proud to help bring this next phase of ZeroDev to market.

0:56
Ed Felten avatar
⟳ Akaki Mamageishvili🇺🇦 retweeted
Ed Felten@EdFelten·Jul 1

Very happy to see Robinhood Chain Mainnet now live. One of the biggest strengths of @Arbitrum’s infrastructure is that it lets enterprises build whatever best fits their business. Robinhood's journey illustrates that well: launch tokenized stocks on Arbitrum One, validate with real users, then move to dedicated infrastructure as requirements around performance, security, and compliance become more specialized. This is why the Arbitrum Platform is so powerful. Enterprises can customize and evolve their infrastructure as their products and requirements evolve. As more financial products move into the programmable economy, different categories will need different infrastructure. Our goal is to give enterprises that flexibility while they scale the next generation of finance. Looking forward to seeing what developers build on the Robinhood Chain.

Post attachment