Looks very exciting! It always seemed likely that the optimal model architecture would be one forward pass per logical step. LLMs are quite wasteful on one end of the spectrum at one forward pass per token. This looks more like one forward pass per prompt? Would be very interesting if one could somehow chain this together now into multi-step (i.e. multi forward pass) reasoning, and optionally in the end unpack cheaply with small LLM.

Binji joining Ethlabs and ETH up 10% the same day - coincidence? 🤔 But jokes aside, super excited to have Binji on the team - Ethereum needs more definite optimism, and I believe there are few people better able to help shape Ethereum’s vision and narrative than him.

Very interesting article by @benthompson on the financial engineering around the AI infrastructure buildout boom: A prompt: This is clearly a place with a lot of financial innovation right now. What productive involvement could crypto have?

Welcome to Ethlabs, @decentrek! I'm incredibly excited to have Derek join us - he is one of the rare people that combine the ability to effortlessly dive deep into any new topic with a strong sense of urgency for shipping. Exactly who want on your team to make Ethereum win!

For a less difficult topic: Our @ethlabs_org week 7 update is out! I am particularly excited about the EIP for faster slots authored by our former EF colleague @CarlBeek and now championed by @barnabemonnot for Hegota inclusion.

For a less difficult topic: Our @ethlabs_org week 7 update is out! I am particularly excited about the EIP for faster slots authored by our former EF colleague @CarlBeek and now championed by @barnabemonnot for Hegota inclusion.

My thoughts on issuance: 1. the problem is real 2. the direction of the EIP is valid 3. but: changing issuance also has major downsides 4. most importantly: the decision must be with the community I said a few words on the ACD call on Thursday, and wanted to expand on my thinking - both as Ethlabs co-founder and ACD moderator. In general, being able to have difficult conversations in public has always been a core strength of Ethereum. 1. The problem is real - Slashing is core to Ethereum’s security. Not all attacks are automatically slashable. E.g. a majority of validators could censor the chain through malicious attesting. It needs to always be credible to slash such attackers. The larger portion of ETH is staked, and the tighter staked ETH is integrated into DeFi, the more credibility of slashing is at risk. - Worse dilution for stakers: at low stake %, most rewards are real income. At high stake %, most rewards just offset dilution. - Economies of scale: The competitiveness gap between centralized staking providers and both solo stakers and more decentralized staking providers widens at higher stake rates. - Narrative confusion around ETH: ETH is a “productive asset” due to protocol revenue (fees & MEV), not staking yield. The staking yield also risks crowding out the emergence of other (productive) yield opportunities for raw ETH. - Worse dilution for raw ETH holders: In general, harder assets are more attractive. BTC as the only $1T+ digital store-of-value explicitly centers zero long-term supply growth as its value proposition. Counter point: Gold with 1-2% yearly supply growth has a $30T store-of-value market cap. 2. The direction of the EIP is valid - If one wants to limit the stake rate, the issuance curve needs to bend downwards for high stake %. The EIP is a specific instance of such a curve. - The EIP also includes a further issuance reduction as a second phase. This introduces minimum viable issuance (MVI) as a secondary, monetary objective of the EIP. I personally think that is reasonable, but these two aspects should be discussed and reasoned about separately. 3. Changing issuance also has major downsides - The decentralization of the staking set is crucial for Ethereum’s health. Many solo stakers are less economically competitive than large operators. A change that results in a significant drop in staking yield risks having these solo stakers disproportionately leave the staking set. - Staked ETH is tightly integrated into DeFi today, both directly though LSTs, and through raw ETH lent out for the purpose of staking. Any major reduction in staking yield thus risks disrupting a core part of DeFi. This effect is worse the more rapid and more significant the yield drop is. - Any change to Ethereum’s monetary policy resets the “monetary policy ossification clock”. Predictability of supply matters for ETH’s attractiveness as store-of-value. So far, changes are only ever towards issuance reduction, but will investors trust that that will hold true forever? - Ethereum is overall at a pivotal moment. Given the large social cost of any months-long debate on issuance, it is reasonable to argue that this should wait until Ethereum overall is in calm waters, even if it makes a decision at a later time more painful. 4. The decision must be with community - Most hard fork decisions are made by the allcoredevs (ACD) governance process. Conceptually, this is “delegated authority” by the community to the core devs. This delegation is very effective on most technical topics. For the occasional hard fork decision that is not primarily technical, this setup is however less well suited. - Issuance in particular is not a technical decision. The decision still needs to be anchored to the ACD process - core devs need to coordinate on what to implement - but the actual decision forming needs to happen on the community level. - This is easier said than done. The ACD process is flawed but well defined. A “community decision” is much fuzzier and harder to operationalize. Discussion venues like X are important, but also limited. How to have high quality community-wide discussions remains an important challenge to solve. As we are getting off the ground at Ethlabs, we are trying to find the right path for how to contribute to both Ethereum and ETH. Issuance is a tricky topic for us: Caspar and I have argued for issuance changes in the past, and for now continue to have conviction in that path. The Ethlabs team as a whole has a wider spectrum of opinions, which has been healthy for internal discourse. And of course, we also see the broad community skepticism on this topic. The ability to have difficult conversations has always been a core strength of Ethereum. Our role in that should be to productively contribute to and engage with such conversations, not to blindly run in one direction. We will iterate to get that balance right, and genuinely appreciate feedback as we do.

Happy 11th birthday, Ethereum! The Ethlabs team worked together with @jackbutcher, @jalilwahdat, and @yougogirl_eth, to put together a small art project to celebrate the occasion. Click through to mint your own celebration NFT and become a node in the birthday graph!

Imagine being on the sidelines of this growth. Billions of people don’t have to imagine - they have no access to invest in American companies. The S&P500 is up 75% in the last 5 years, and the majority of the world just had to watch from the sidelines. Tokenized stocks are changing this, and finally making financial access more open. Just a phone and an internet connection can let you own a piece of the world’s best companies.

Another important step in Ethereum's reconfiguration for the next phase of adoption. Welcome, Ethereum Institutional! We have been closely coordinating with the team and will share more about Ethlabs <> Ethereum Institutional collaboration plans soon. Exciting times ahead!

“There are two crypto assets worth more than $100 billion: BTC and ETH. In many ways, Ethereum is in a unique position. Ethereum has had a lot of success. The problem is, it had that success five years ago… so the market is saying “you won round one, but what’s round two?” So internally, our prompt has been “what does it take for ETH to become a trillion dollar asset?” There aren’t many individual trillion dollar assets on earth, so what does it take for ETH to be a trillion dollar asset? For Ethlabs, that will be one of our north stars. We’d be a bit full of ourselves if we came out and said that we had the answer today; it’s going to be a huge project for the entire community. But Ethlabs can act as the bridge. We can be the entity that goes and talks to the people that want to take on that mission. Then we can go and talk to the people that help steward Ethereum and work together to make that outcome happen.” — @adietrichs talks @ethlabs_org and a trillion dollar Ethereum on @Bankless.

very excited to start @ethlabs_org!! some thoughts: * ethereum the protocol is decentralized and now ethereum's ecosystem is accelerating its decentralization. it's like client diversity: more costly and coordination friction but ultimately the principled and resilient path. the multi-node future decentralizes both the tech and the culture. * I strongly believe principles and pragmatism are not at odds with each other. the core properties of ethereum, ultimately its credible neutrality, is precisely what gives ethereum a shot at becoming the root of trust for the world. trust doesn't scale well but ethereum's credible neutrality does. this is unique and what I'm most excited about pushing for.