I also gave a talk on the same subject.
About a year ago, the Eth1x research community embarked on an effort to bring stateless block execution to the Ethereum protocol in preparation for the Eth2 merge.
Stateless block execution is really cool. The idea is that clients no longer have to hold the state. Instead, when a new block is mined, we also create this fancy thing called a witness. The witness provides all of the state needed to execute the block. The end result is being able to build clients with the same security guarantees as we have today, but without all the cumbersome overhead of managing the state. Clients built in this model can run on low end hardware needing trivial amounts of CPU/RAM/HDD and bandwidth. Sounds amazing right?
It is. But it also isn't. Unfortunately, the only thing a client like this can do is verify blocks. Without the state, it can't serve the majority of the JSON-RPC API or send transactions, because it cannot participate in transaction gossip. It turns out that stateless block execution is only a small piece of what we really need to make broadly useful stateless clients.
Conveniently, all of the infrastructure necessary to realize functional light clients naturally extends to stateless clients as well. The scope of the Stateless Ethereum roadmap is expanding. We are going to be building new decentralized peer-to-peer networks, new clients for these networks, and we're going to need new teams of researchers and engineers to do it.
If you want to get in on the ground floor, join the next Stateless Ethereum call on February 9th at 4pm UTC. Details on joining the call can be gotten from the Ethereum R&D Discord server.
A more concrete definition of the research and work that needs to be done: https://notes.ethereum.org/mSOAdx_XT02MEqrt0f2CPA
If you are curious about the title: https://blog.x.company/tackle-the-monkey-first-90fd6223e04d