May 16, 2026 • by Daniel Osei • Ethereum
For most of its life, Ethereum has had a fundamental UX problem that most people in the industry quietly accepted as the cost of doing business. To interact with the network you needed an externally owned account — a private key that, if lost, meant your funds were gone with no recourse. No recovery phrase logic baked into the protocol, no spending limits, no way to set a trusted contact to help you out if things went wrong. You just had to not lose the key. This worked fine for a narrow audience of technically comfortable users but made mass adoption a difficult proposition.
ERC-4337 — account abstraction — changes the underlying model. Instead of all user accounts being dumb key-controlled addresses, the standard allows smart contracts to act as accounts. This opens a surprisingly wide surface area. Wallets can now offer social recovery, where a set of trusted contacts can collectively restore access if you lose your credentials. They can enforce daily spending caps so a compromised device can't drain everything at once. They can let you pay gas fees in whatever token you happen to hold, rather than requiring ETH specifically. And they can bundle multiple operations into a single transaction, which alone removes a lot of the friction that makes DeFi feel clunky compared to traditional finance.
The change is not cosmetic. It shifts where application logic can live, who bears responsibility for transaction validity, and how wallets can differentiate themselves. Whether this actually translates into meaningfully better products depends entirely on how wallet developers and applications choose to use the new primitives — but the primitives are now there, and they are solid.
Read more →
May 12, 2026 • by Daniel Osei • Layer 2
A few years ago the conversation around Ethereum scaling was still largely theoretical. Today there are live rollup networks processing more transactions per day than the Ethereum base layer itself, with fees that are orders of magnitude cheaper. That's a genuine engineering achievement and worth acknowledging before getting into the nuances. But the nuances matter, because "Layer 2" has become a catch-all label applied to things that are meaningfully different from one another, and that difference matters when you're deciding where to build or where to keep assets.
Optimistic rollups — Optimism, Arbitrum, and their descendants — batch transactions and post them to Ethereum, assuming they're valid unless someone raises a fraud proof within a challenge window that typically lasts seven days. This works well and the ecosystem around both networks has matured considerably. The seven-day withdrawal delay is a real inconvenience for users who want to move funds back to mainnet, though bridge solutions have emerged to absorb that wait at a cost. ZK rollups take a different approach, generating a cryptographic proof of correctness for every batch. Verification is faster and withdrawals don't require the same waiting period, but generating proofs is computationally intensive and the engineering complexity has historically been high. That gap is closing faster than most people expected.
What's less discussed is the diversity of security models across networks that all call themselves rollups. Some post full transaction data on-chain. Others post only compressed calldata. A few post only state roots, which makes them more like validiums than rollups in any strict sense. These are not equivalent choices — they represent real trade-offs between cost, throughput, and the guarantees users actually get if things go wrong. The label matters less than understanding what the network actually does.
Read more →
May 7, 2026 • by Daniel Osei • Identity
Most discussions about Web3 revolve around financial applications — tokens, trading, lending, and the various ways value can move between addresses. This makes sense given where most of the activity and attention has been concentrated. But there is a parallel layer being built that is less visible and arguably more foundational: decentralised identity. The ideas here are old, but the tools to implement them properly are only now becoming practical.
The core concept is straightforward. Today your identity on the internet is fragmented across dozens of platforms, each holding a slice of your data, each controlling your access, each able to revoke or suspend your presence at will. Log in with Google, and Google knows every service you use that way. Build a reputation on a platform, and that reputation disappears if the platform does. Decentralised identity flips this by making credentials something you hold rather than something platforms issue and control. W3C's Decentralised Identifiers standard and the Verifiable Credentials framework provide the scaffolding. Projects like Ceramic and Lens Protocol have been building application layers on top.
The practical implications are substantial. A developer who builds a reputation contributing to open-source projects could carry that record across platforms without any one of them being able to erase it. A borrower in DeFi could demonstrate creditworthiness using verified off-chain data without exposing their entire financial history. A community could gate access to spaces based on verified credentials rather than token holdings, which creates richer and more nuanced membership structures. None of this is fully deployed at scale yet, but the pieces are assembling and the direction is clear.
Read more →
April 30, 2026 • by Daniel Osei • Governance
The theory behind decentralised autonomous organisations is genuinely compelling. Replace corporate hierarchy with on-chain governance. Let token holders vote on decisions. Make the rules transparent and auditable by anyone. Remove the need to trust a small group of insiders to act in the interest of the broader community. In principle this addresses real problems that have frustrated people in traditional organisations for a long time.
In practice, most DAOs have struggled with a predictable set of issues. Voter participation is chronically low — the majority of token holders in most active DAOs never vote on most proposals, which means decisions effectively get made by a small active minority anyway. Plutocratic weighting, where voting power scales directly with token holdings, means large holders have disproportionate influence over outcomes regardless of how legitimate their interests are relative to smaller participants. And on-chain voting is slow and expensive enough that many organisations have moved to off-chain signalling tools like Snapshot, which reintroduces trust assumptions the whole structure was supposed to avoid.
None of this means DAOs are a failed concept — it means the first generation of implementations made understandable simplifications that are now being revised. Newer governance designs are experimenting with delegated voting, reputation-weighted systems that don't reduce directly to token balances, staged proposal processes that reduce noise at the voting stage, and time-locks that give communities a window to respond to decisions before they execute. The governance design space is genuinely interesting and moving quickly. Getting it right matters because the organisations that figure this out will be more resilient and more legitimate than anything that came before them.
Read more →
April 23, 2026 • by Daniel Osei • ZK
Zero-knowledge proofs have gone from a theoretical cryptography curiosity to a core piece of blockchain infrastructure in a relatively short period of time. They are now central to how several major networks achieve scalability, and they are beginning to show up in privacy applications, identity systems, and compliance tooling. The mathematics behind them is genuinely complex, but the intuition is accessible and worth understanding because the technology is going to matter in many adjacent fields over the next decade.
The basic idea is this: a zero-knowledge proof allows one party — the prover — to convince another party — the verifier — that a statement is true without revealing anything beyond the truth of the statement itself. The classic informal example involves proving you know the solution to a maze without showing the path. More practically, a ZK proof can verify that a computation was carried out correctly without requiring the verifier to re-run the computation themselves, and without the prover needing to reveal the inputs. This property — succinct verification of computation — is what makes ZK rollups viable. A network can process thousands of transactions off-chain, generate a single compact proof that all of them were valid, post that proof to Ethereum, and have it verified cheaply. The verifier doesn't need to see the individual transactions.
Beyond scaling, the privacy properties are where things get genuinely interesting. A user could prove they are over a certain age to access a service without revealing their date of birth. A borrower could prove their income exceeds a threshold without disclosing the actual figure. An organisation could demonstrate regulatory compliance without opening its books. These are not hypothetical — implementations of each exist today at varying stages of maturity. The hard part is building the tooling and standards that make these proofs easy to generate, verify, and integrate into existing systems. That work is ongoing, and it is progressing faster than most observers expected even two years ago.
Read more →