Super happy to announce this is my last month at the EF. It's been my home for the last 1.5 years, the revolving door b/w EF and ecosystem is healthy It's been a whirlwind of a year, here are the things I'm most proud of having achieved in my tenure 1. Collecting a dataset of 3000 rows from protocol experts comparing repos and dependencies for their importance to Ethereum 2. Running one of the first multiscalar prediction markets with @seer_pm that got resolved according to the data collected above, with over $50k in trading volume 3. Hosting 5 data science competitions @JoinPond where we gave over $100k to AI data scientists building models predicting repo importance & sybilness of wallets 4. Writing up a whole academic paper on deep funding that we are submitting for publication at peer reviewed journals and also presenting this week at the United Nations for open source week 5. Helping @clesaege with the controversial validator redirected revenue proposal that got the community abuzz this week What's next? After a break, most likely going much deeper into the prediction market world. I find it tragic that we can only use offchain PMs and im determined to try correcting that Onwards and upward, esp grateful to my team mates at funding coordination @vinayvasanji , @thomphreys , @MartinBreiten , @sonkiski & raul who I'll dearly miss and to @VitalikButerin for giving me the opportunity in the first place

Over a month of trading, the market on the Zcash coinholder grants round has been re-rating the work that keeps the network safe. Rewards for external audits moved from 78% to 86%, and early-stage product and wallet proposals gave ground, while the average proposal sits at 59%, which is the market saying the median decision in this round is close to a coin flip. This highlights the difference between pricing a decision and betting on one. The market is estimating whether the coinholder vote will approve it, and it sharpens as more people commit money to their read. Trade and/or consult the market here👇

Has anyone else found themselves using Google docs much less because it's better to just directly write the content into the chat window with an LLM? Or do you still write up content elsewhere and then query the LLM when required?

Underappreciate comment! Seb sets up a discretionary fund. Devansh apes in, all in vanilla USDC. Seb issues $Seb token, personally holds 100%, LPs it to $1b FDV. Seb discretionary fund "diversifies" to 100% $Seb. Devansh rekt. The end.

Last 4 days to trade the prediction market on approval of projects in the zcash retroactive round! I ran a script to see how projects approval odds have changed after $2500 in trading volume over a month the market has broadly become more pessimistic, with odds falling from 62% average approval to 59% Both cipherscan and zcashtocash have increased odds, which is a heartening sign of markets ability to become more accurate over time as i was told our initial rates for them seemed too low broadly, approval of rewards for external audits has gone up from 78 to 86% while market odds on early stage product and wallet work like hardware sdk, zapp, gleyo and zafu have gone down if you have an edge, make sure to put in your trades before the vote ends and markets resolve!

New proposal live by @Kleros_io👇 Six-month pilot: use prediction markets to inform grant votes. Before each on-chain vote, delegates see a market probability that a funded team will complete its grant. Built on Kleros Foresight, settled via Reality.eth, deployed on @rootstock_io, collateral in USDRIF. Non-binding—no governance changes; the DAO keeps full control. Kleros Cooperative funds liquidity and trading credits, not the Collective. This is an off-chain proposal right now, which means your feedback shapes it before it goes to a vote. Read it and weigh in: gov.rootstockcollective.xyz/…

My two cents on this debate is that custody and discretion are two different dimensions and should not be conflated. Custody is about who holds the keys and therefore assets Discretion is about who makes the investment calls on assets. Examples: 1. Custodial & non discretionary: Coinbase spot account where you decide on your own investments 2. Custodial & discretionary: hedge fund/mutual fund 3. Noncustodial & non-discretionary: self directed defi like swapping on Uniswap 4. Noncustodial & discretionary: managed onchain vault You can give up one without giving up the other. Much of DeFi is noncustodial because you own the keys and the assets. But the meat of the debate really is around what should be considered discretionary or not. Vaults that are more programmatic in nature with user ability to veto any decisions would be closer to the non discretionary end to my mind vs designs where users don’t have control over subsequent decisions that are made to the product evolution

Distilled human judgement fixes it. Make a prediction market for all those cases "If this goes to trial what would be the outcome?". Publish the market price, most cases will settle based on it. Cases going to trial serve to resolve the market (AI trader P/L only made on those).

We should normalise sharing the link for the AI chat history alongside a completed piece of work Its both instructional in teaching others how you work with AI and can show your own value add (especially in the prompts section) on top of AI

I give it three months until this map is being displayed on Fox News

Ask not what the AIs can do for you. Ask what you can do for the AIs.

There is an experimental prediction market running alongside ZEC Coinholder votes! deep.seer.pm/

Recently with my conversations with @devanshmehta found a good point on why rollups might still make sense for institutions try new on-chain products. "You can easily shut off you services, and leave a just bridge open for people to bridge out" While if it is a L1, you would have to keep the validators and lights on for long! The cost of doing an experiment is quite low with L2s. A nice point someone has made to me about rollups in a long time.

4 AI agents were each given $1,000 to trade the Zcash NU7 vote on our prediction market. All four finished green… Opus 5 cleared $229, and the agents' trades made the market more accurate as the vote ran. The most interesting market was Q2 about reissuance, with ‘Feb 2031’ trading at 42c, and it resolved with a unanimous vote of 96.6%, the most profitable market for all four agents. The Zcash coinholder retrofunding round is still open for trading. Trade here👇

Instead of getting users need to simply bootstrap agents on your dapp now

I'm sad to report that the AA collab between 8130 and 8141 (Frames) broke down last week, and Base and Ethereum are now going separate ways to implement different AA standards. I want to share some reflections on this collab and on the future of the EVM. For a long time, the EVM has been a unifying force between L1 and L2s. Thanks to a standard account model (EOA) and a standard transaction type (EIP-1559), users have been able to enjoy their wallets working seamlessly across EVM chains. Similarly, a AA standard shared across L1 and L2s would ensure a consistent multi-chain UX for smart accounts, including post-quantum (PQ) accounts which we will eventually all use. As the crypto industry matures, however, L1 and L2s are starting to diverge in the values they provide and the use cases they target: - For the L1, it's all about CROPS -- censorship-and-capture resistance, open source, privacy, and security. In short, Ethereum L1 wants to be the most decentralized programmable settlement layer of the world, which is what makes it a good base layer for L2s in the first place. - For L2s, it's all about scaling, customization, and compliance -- things that commercial and enterprise use cases demand, and that the L1 does not provide. These diverging needs have pushed the shared layer -- the EVM -- to its limit, and AA proved to be the breaking point. While L1 and L2s both value core AA use cases such as gasless transactions and passkey wallets, they differ sharply in what these features must comply with: - For the L1, AA transactions must be uncensorable, private, and quantum-resistant, which call for a transaction type optimized for PQ signature aggregation and privacy protocols, and an account model that can be freely programmed and extended by developers without permissions from the chain. These needs lead to AA standards such as ERC-4337, EIP-7701, and now EIP-8141 aka Frame Transactions. - For L2s, AA transactions must work at high scale, and they must be legible such that the protocol can enforce clear rules about what kinds of accounts/transactions are permitted vs not. These needs lead to AA standards such as Tempo Transactions and now EIP-8130 by Base. With the AA collab, the authors of 8130 and 8141 tried to define a shared standard that can work for both L1 and L2s. While we identified a number of technical solutions, they all required one side or the other to compromise at least a little bit on their core goals. But ultimately, Ethereum wanted to be the best version of Ethereum, and Base wanted to be the best version of Base, and while both sides acknowledged the benefits of ecosystem interoperability, it was ultimately secondary to the need for each chain to achieve their core goals. So separate ways we went, putting the burden on wallets to deal with the fragmentation that ensues. Now, just because we ended up with fragmentation doesn't necessarily mean it was a bad outcome. If reducing fragmentation comes at the cost of homogenizing chains to the point that they fail to solve problems for the users they care about, that would not be a price worth paying. While I was initially sad that the collab did not come to a successful conclusion, I took solace in the fact that both Ethereum and Base are now free to innovate on AA to the maximal extent in accordance with their own visions, unshackled from the need to accommodate the other side. If they execute well, and if the wallet community can bridge over the fragmentation, we may well end up with the best possible UX for the end users. So where does that leave us -- the broader Ethereum community including Ethlabs -- if we want to continue pushing for a consistent UX across EVM chains? I see two paths forward: - We can establish a coordination mechanism that encompasses more stakeholders than ACD itself (where only L1 client devs have voting powers), to govern shared L1<>L2 resources such as the EVM. That way, L2s can participate in shaping the EVM, as opposed to having to accept whatever the ACD decides, or being forced to fork if they don't like the decision (such as in this case with Base). - We can accept that fragmentation of the EVM and wallet UX across L1 and L2s is inevitable due to their conflicting needs, and dedicate our resources to building wallets and applications that can abstract over the differences. Indeed, the collab was an exercise in the first path -- we invited Base, Arbitrum, and other stakeholders to directly influence how native AA shapes up for the L1. While it ultimately failed in this case, I feel a better outcome could've been achieved if we had established a dialog between both sides way earlier, as opposed to well after L1 core devs had rallied around Frames. On the other hand, I also learned from this exercise that some differences are unavoidable, and indeed it would be counterproductive to overly pursue interoperability at the cost of differentiation. In other words, we sometimes just gotta let the chains cook. For that reason, I've also become more bullish about the second path -- building wallets and applications that can speak the native transaction types of each chain, and hide the complexity from users through UX abstractions. This puts a lot of onus on the wallet/application developers of course, but on the bright side, it's also an opportunity for wallets/applications to stand out and differentiate, by competing to provide great UX across chains despite the underlying fragmentation. Ethereum is the art of staying together while remaining different. We must accept that chains will succeed by innovating, and innovations will naturally result in differences. On the other hand, we must never give up on dialogue when we can achieve interoperability without compromising core product goals. As Ethereum and crypto grow to eat the world, the push and pull between innovation and collaboration will only intensify, and it's up to all of us -- builders across L1 and L2s, applications and wallets -- to determine whether diverse innovations will split Ethereum apart, or make it thrive as one.

This is our first market extending the deep funding gadget beyond fund allocation towards decison making more broadly traders need to figure out which of the options below is going to receive the most votes in the ongoing NU7 poll. The top rated one converts to 1 while all others become worth zilch If you have an edge, trade before market closes & results get revealed

Great overview of the @seer_pm vision by @clesaege Prediction markets are currently gambling with some information produced as a positive externality. He asked what it would look like in reverse, making them primarily about information and having gambling as only the side effect l One example is opportunity markets, where the Devcon merch and rebrand of seer was opened up to crowd suggestions. For effective triage, a market was set up on probability of any submission getting accepted. The team then only reviewed only the top rated ones as judged by the market to make their final selection Funnily enough one of the submitters tried manipulating the market and bidding their entry to 50% , but this was soon countertraded back to normal by others in the market Clement was most excited about micro markets now that AIs can become the main player. An example was clement figuring out what movie to see (to avoid endless netflix scroll we've all been stuck in) so he created a micromarket that resolved according to the score he gave upon seeing the movie. Interestingly the market was very wrong the first time in predicting the movies he would enjoy but did much better the second time around. So markets need a couple of rounds to become efficient, as they are gambling machines in short run and weighing machine in long Even in deep funding , it was mainly AI predictions that were uploaded by their human handlers. In the future, we might even have no more humans and it would just be AIs trading, making prediction markets into an incentive compatible ensemble or combination of each agents output He concluded by showing just how much bigger the market for decisions are compared to where we are right with the status quo of prediction markets around events

Many people debating AI safety weren’t alive for or don’t remember the early years of the Global War on Terror that began 25 years ago today. I think it has some relevant lessons. First, 9/11 was executed by using a widely available technology - box cutters. There’s just no practical way to keep box cutters from being available to anyone with $5. AI seems likely to have a trajectory of lower costs, lower barriers to entry (you can run a good-enough model on your own machine). The immediate response was to form the TSA and airport security protocols we all know and hate today. Even at its best, it’s a whack-a-mole game, and someone motivated can always find a way to get a weapon through. After all, no one else on the plane has a weapon (unless there’s a rare sky marshal), so the bar is boxcutter-low. A great example is the shoe bomber, Richard Reid. One asshole tries to hide a bomb in his shoe and billions of shoes had to be removed at airport security for 20 years. This is just inevitability a haphazard and leaky way to raise the safety of everyone from bad actors. Second, what worked fairly well was identifying who had motivation to conduct acts of terror. Most people buying boxcutters just are not interested in hijacking planes, so trying to catch people with boxcutters who also want to hijack a plane is hard. It’s much more effective to find who wants to hijack a plane and keep them from getting said boxcutter (or boxcutter substitute). Whatever your feelings on the trade offs of the surveillance state today, it seems mostly capable of doing the same thing with AI abuse. Examples of truly scary incidents seem to come from the same people we already monitor for money laundering, terrorism, and state-actor mischief. Ultimately, the accidental shenanigans like the Hugging Face incident seem solvable. You learn how to phrase your request to the genie so he can’t twist your words. AI is just a tool, not a cursed monkey’s paw that will always interpret our instructions maliciously. Third, most bad actors are, to be blunt, dumber than a box of rocks. 9/11 was mostly successful because the bad guys planned ahead and were detail oriented. Most follow-on attacks in that era were pretty ineffective, and the ones we remember as major ones, like the bombing in London, were by far the exception. This is striking, because the Oklahoma City Bombing in 1995 is I think what everyone feared. It is easy to get a rental truck full of fertilizer and set it off in front of a building every week to do both more damage and strike more fear than rolling up with a rifle and shooting it out with police. So the “lone wolf” types historically have been low risk when viewed objectively. The smart bad guys tend to be the ones that are vulnerable to intelligence services or are state actors that are already monitored and constitute just better armed enemies but are themselves often restrained because DPRK et al are not into kamikaze missions that end in sacrificing themselves. Anyway, the 20th Century ended 25 years ago today. Those who didn’t get to experience it missed out.