i believe most engineering teams working on core ethereum will disappear over the next 3 years. there is no real money to be made there, and nowadays there isn't even that much status to gain from doing it. AI changed the landscape pretty fast because implementing a client is simply not as difficult as it used to be. i actually think this is good for ethereum users and stakers. most crypto engineering teams outside ethereum will disappear even earlier. lambda's biggest strategy is to just keep doing what we are doing while most people eventually vanish from crypto. there aren't many teams with lambda's combination of expertise in cryptography, compilers, security, consensus and performance, together with the connections we have across corporations and startups. we're finding different ways of making money from the connections and expertise we have across aligned and lambda. i also don't think there are going to be many fundamentally new use cases beyond the ones we already know. what will happen instead is that ethereum will become the financial backend for a large part of international payments and fx. there are almost no economic incentives to be a core dev, and i don't think donations from ETFs or similar mechanisms will fundamentally change that in the long run. the reason is that it doesn't make sense to keep pouring money into large groups of people that don't produce that much. most of the work that actually needs to be done is researching new ideas, testing them and implementing them well. the ethereum spec will likely end up being written in lean, and lean may eventually become one of the main implementations too. there won't be a need for so many independent engineering teams, apart from having enough competition to keep progress moving quickly. the problem today is that many of the existing engineering teams have very little incentive for the system to change too much or too fast. most people on CT believe the EF is the main thing blocking progress in ethereum. it's true that for a few years after the merge, i think research progress stagnated and we weren't bold enough outside of ZK research. most of my criticism towards the EF was related to its size and the fact that it never prioritized making the UX of ethereum better or spreading the gospel to corporations and governments. now there are multiple new organizations doing this, so we're fine. nowadays i believe engineering teams are the biggest bottleneck preventing ethereum from becoming faster and better. the EF is actually the one pushing things forward. most engineering teams have almost zero economic incentive for things to change dramatically. what ethereum needs is to accelerate the natural consolidation of these engineering teams so that we can push for bigger changes much more aggressively. i hope the EF and CT start realizing this. ethereum won't have a real competitor thanks to: - current network effects - formal verification - minimalism - post-quantum security - privacy - zk becoming stupidly fast, with proofs verifiable on basically any piece of hardware - lean ethereum: new consensus, faster p2p and faster finality, while supporting even more stakers and maximizing decentralization - propamms solving many of the problems around pricing assets and trading on a slower chain like ethereum i don't think solana, tempo, hyperliquid or robinhood have a chance against Ethereum if we just focus on the important things and do the inevitable: remove the bottlenecks that slow progress.

i believe most engineering teams working on core ethereum will disappear over the next 3 years. there is no real money to be made there, and nowadays there isn't even that much status to gain from doing it. AI changed the landscape pretty fast because implementing a client is simply not as difficult as it used to be. i actually think this is good for ethereum users and stakers. most crypto engineering teams outside ethereum will disappear even earlier. lambda's biggest strategy is to just keep doing what we are doing while most people eventually vanish from crypto. there aren't many teams with lambda's combination of expertise in cryptography, compilers, security, consensus and performance, together with the connections we have across corporations and startups. we're finding different ways of making money from the connections and expertise we have across aligned and lambda. i also don't think there are going to be many fundamentally new use cases beyond the ones we already know. what will happen instead is that ethereum will become the financial backend for a large part of international payments and fx. there are almost no economic incentives to be a core dev, and i don't think donations from ETFs or similar mechanisms will fundamentally change that in the long run. the reason is that it doesn't make sense to keep pouring money into large groups of people that don't produce that much. most of the work that actually needs to be done is researching new ideas, testing them and implementing them well. the ethereum spec will likely end up being written in lean, and lean may eventually become one of the main implementations too. there won't be a need for so many independent engineering teams, apart from having enough competition to keep progress moving quickly. the problem today is that many of the existing engineering teams have very little incentive for the system to change too much or too fast. most people on CT believe the EF is the main thing blocking progress in ethereum. it's true that for a few years after the merge, i think research progress stagnated and we weren't bold enough outside of ZK research. most of my criticism towards the EF was related to its size and the fact that it never prioritized making the UX of ethereum better or spreading the gospel to corporations and governments. now there are multiple new organizations doing this, so we're fine. nowadays i believe engineering teams are the biggest bottleneck preventing ethereum from becoming faster and better. the EF is actually the one pushing things forward. most engineering teams have almost zero economic incentive for things to change dramatically. what ethereum needs is to accelerate the natural consolidation of these engineering teams so that we can push for bigger changes much more aggressively. i hope the EF and CT start realizing this. ethereum won't have a real competitor thanks to: - current network effects - formal verification - minimalism - post-quantum security - privacy - zk becoming stupidly fast, with proofs verifiable on basically any piece of hardware - lean ethereum: new consensus, faster p2p and faster finality, while supporting even more stakers and maximizing decentralization - propamms solving many of the problems around pricing assets and trading on a slower chain like ethereum i don't think solana, tempo, hyperliquid or robinhood have a chance against Ethereum if we just focus on the important things and do the inevitable: remove the bottlenecks that slow progress.

We've been working on something similar with @LambdaClass, @alignedlayer and @polfinance_ since @EthCC, great work by @kubimensah and @titanbuilderxyz. This will change things very fast. The bid ask spread is the hidden tax every trader pays on every trade. Tighter spreads mean better prices, more volume, and deeper liquidity. Right now Uniswap owns DeFi, Binance owns CEX, and Solana ate part of Ethereum's lunch. If PropAMM brings spread compression to Ethereum onchain, things change entirely and Ethereum will become the best place to trade. Traders follow liquidity, liquidity follows better prices. Better prices come from tighter spreads. All of it gets contested: Uniswap's dominance, Solana's momentum, Binance's volume. PropAMM on Ethereum could redraw the entire power map of this ecosystem. As always, LambdaClass will work with anybody trying to make Ethereum win. As always, LambdaClass will work with anybody trying to make Ethereum win. We've already delivered @ethrex_client. We're about to deliver an open source Wallet-as-a-Service and LambdaVM, a RISC-V zkVM built with @alignedlayer. We're building @ethlambda_lean, a Lean Ethereum client. We're also launching stablecoins on Ethereum for LATAM with @lemonapp_ar. We've been helping build a new coalition in the block building world, and now we're entering DeFi.

In Cannes, Blockspace Forum held its second workshop, attended by 32 teams representing >95% of out-of-protocol blocks. We’ve summarized the key outcomes and concepts that surfaced. Thank you to everyone who contributed. The ticker is ETH!

Justin (@drakefjustin) is one of the main reasons I love working in Ethereum and why I believe it has a promising future. During @EthCC, I hosted a fireside with him covering Bitcoin, post-quantum security, @LeanEthereum, @eth_proofs / ZK, Ethereum transaction rails, block building, MEV, and a few other fun topics. Hope you enjoy it as much as I did. Thanks @blockspaceforum for the space to explore so many interesting topics.

LLMs now make critical decisions in hospitals, defense, banks, and governments. Yet nobody can verify which model actually ran, or whether the output was tampered with. A provider or middleman can swap weights, silently requantize the model, alter decoding, inject hidden prompts, do supply chain attacks, or change the deployment surface without the user knowing. This problem is already serious. It will become critical. We think this needs a practical solution, not just a theoretically clean one. CommitLLM is designed to be deployable on existing serving stacks now: the provider keeps the normal GPU serving path, does not need a proving circuit, does not need a kernel rewrite, and does not generate a heavy proof for every response. In practice, two families of approaches dominated the conversation before this work: fingerprinting, which can be gamed, and proof-based systems, which are theoretically strong but too expensive for production inference. We built CommitLLM to target the middle ground. The core idea is to keep the verification discipline of proof systems, but specialize it to open weight LLM inference. The cryptographic core is simple: Freivalds style randomized checks for the large linear layers, plus Merkle commitments for the traced execution. Then a lot of engineering work is needed to make that line up with real GPU inference. The key trick is this. A provider claims `z = W × x` for a massive weight matrix. Normally you would verify that by redoing the multiply. Instead, the verifier samples a secret random vector `r`, precomputes `v = rᵀ × W`, and later checks whether `v · x = rᵀ · z`. Two dot products instead of a full matrix multiply. In the current implementation, a wrong result passes with probability at most `1 / (2^32 - 5)` per check. A full matrix multiply, audited with two dot products. Most of the transformer can then be checked exactly or canonically from committed openings. Nonlinear operations such as activations and layer norms are canonically re executed by the CPU verifier. The one honest caveat is attention: native FP16/BF16 attention is not bit reproducible across hardware. CommitLLM verifies the shell around attention exactly, then independently replays attention and checks that the committed post attention output stays within a measured INT8 corridor. So attention is bounded and audited, not proved exactly. That means the protocol already gives very strong exact guarantees on the parts that matter operationally most. If an audited response used the wrong model, the wrong quantization/configuration, or a tampered input/deployment surface, the audit catches that exactly. That includes things like model swaps, silent requantization, and provider side prompt or system prompt injection. Today the implementation and measurements are strongest on Qwen and Llama. But the protocol itself is not meant to be Qwen or Llama specific: we expect it to generalize across open weight decoder only families. What still has to be done is the engineering work to integrate and validate more families explicitly, and we are already working on that. On the measured path, online generation overhead is about 12 to 14% with the provider staying on the normal GPU serving path. The heavier receipt finalization cost is separate and can be deferred off the user facing path. The main systems costs are RAM and bandwidth, not proof generation. The full response is always committed, but only a random fraction of responses are opened for audit. Individual audits are much larger, roughly 4 MB to 100 MB depending on audit depth. The important number is the amortized one: under a reasonable audit policy, the added bandwidth averages to roughly 300 KB per response. After too many weeks without sleep, I’m proud to show what I built with @diego_aligned: CommitLLM. Thanks Diego for your patience. I've been calling you at random hours. The code and paper still need some cleaning and formalization. We’re already in talks with multiple providers and teams that have cryptography related ideas on how to improve it even more. We’re really excited about this and we will continue doubling down on building products in AI, cryptography and security with my company @class_lambda. If governments, hospitals, defense and financial systems are going to run on LLMs, verifiable inference is not optional. It is infrastructure. I will be explaining this in more details in the days to come and I will show how to test it and run it.

With @class_lambda we interviewed @DrewVdW from @Commit_Boost. I believe Commit Boost is the most important public goods that Ethereum should embrace. It already has a 35% of marketshare. I recommend you to read the interview and to use it.