

Breaking moves, market context, and crypto explained for actual humans. Newest first, always no-shill.
Showing 1–34 of 34 stories
Clear filters ×
The inflation cut is real, but the first binding vote also showed how much one large validator can matter.

AI found real bugs, but operators are being asked to run signed fixes before the full source story is public.

Pasteur puts two quiet assumptions in public: who can validate bridge messages, and who gets to build blocks before validators sign.

Urgent L1 releases can be necessary, but binary-only candidates move the risk from code review to operator trust.

This is not a wallet hack. It is a phishing list with names, emails, phone numbers and home addresses.

BIP-110 needed miners to signal during its mandatory window. Instead, the chain enforcing that rule is showing what weak support looks like in public.

The proposal's real test is no longer the data limit. It's whether enforcing nodes can make miners signal without walking onto a minority chain.

This wasn't a private-key story first. It was a server credential story, and that's where Bitcoin payment security is moving.

The latest releases are less about shiny features and more about making bad operator states fail loudly.

This is not an exchange hack. It is a wallet safety story, and the affected-version list is still moving.

The upgrade admits the old Orchard pool can't stay open while still trying to preserve private money.

Two hidden patches this close together don't look like routine cleanup. They look like the first fix missed something.

The patch is live and the advisory isn't. Every operator who hasn't upgraded is now sitting between those two facts.

The Powerloom chain goes dark at 6 AM UTC on July 21. After that, the Arbitrum bridge stops working and any assets still on-chain are permanently inaccessible.

Blockchain security researcher Specter flagged the drain. Hedera hasn't said a word yet.

This is not a bad app bug. It is a poisoned developer package that can steal secrets before your code even runs.

Ironwood is technically ready. The thing that delays L1 launch windows isn't bugs. It's exchanges.

Japan's SBI Group is pulling its pool after five years without giving a reason. The real question is where those miners go next.

Governance just moved from forum post to onchain vote. The 100,000 SOL proposal threshold already tells you who runs Solana.

Karst finishes the op-geth sunset, but the more important move is Optimism wiring ZK proving deeper into its fault-proof stack.

A two-hour halt on one of Ethereum's busiest networks ended with a promise and no root cause. The post-mortem is the story.

Apple's moderation system flagged the developer who was fighting fake Bitcoin wallet apps. If his appeal fails before June 30, Sparrow for macOS stops updating.

Two fork-choice bugs in Ethereum's consensus layer could have made validators lock in the wrong chain. Sigma Prime is keeping the details hidden until beacon node operators catch up.

Multi-prover systems are supposed to eliminate single points of failure. Key management is a different problem.

Base patched a live bug alongside a required hardfork upgrade, and the bug is the more important of the two.

CryptoBandits doesn't need you to approve a transaction. It waits for a copied address, swaps it, then moves the real theft over Tor.

Both co-executive directors are gone, a board member is holding the interim role, and the EF still hasn't named anyone permanent.

Oman's government can now see every licensed miner's revenue, energy use, and fresh bitcoin without banning a single machine. The template has two countries behind it now.

The difficulty algorithm is doing exactly what it was built to do. The question is how many miners it's adjusting around.

The v9 testnet line is moving from feature work into stabilization, and the alarm is aimed at validators, not traders.

A transaction that fails gracefully and a transaction that crashes a node are two different problems. Sui just fixed the second kind.

Base Azul already dropped every client except two. Pruned node operators still on `base-consensus` have until tomorrow to catch up.

A wrong state trie in a proof client produces wrong proofs. Multi-client security is a real upgrade, but it comes with a multi-client maintenance bill.

Validators are being asked to upgrade to code no one outside Aptos Labs can inspect, with no root cause disclosed and no timeline for opening the book.