Sam Altman’s OpenAI unveils ‘EVMbench’ to test whether AI can keep crypto’s smart contracts safe
AXIOM EMPLOYEE ACCUSED OF ALLEGEDLY INSIDER TRAINING BY ZACHXBT: Blockchain sleuth ZachXBT said a senior employee at onchain trading platform Axiom Exchange allegedly abused internal access to user data to track private wallets and potentially trade memecoins using inside information. In a thread posted to X, ZachXBT said Broox Bauer, a New York-based senior business development employee at Axiom, used internal dashboards to look up sensitive user information — including linked wallet addresses — and shared that data with a small group that mapped trades of prominent crypto influencers. Axiom, founded in 2024 by Mist and Cal and a member of Y Combinator’s Winter 2025 cohort, has generated more than $390 million in revenue to date, according to the investigator. ZachXBT said he was retained to investigate allegations that internal tools were being misused. He did not say who retained him. In audio clips shared in the thread, a person said to be Bauer allegedly claims he can track “any Axiom user” by referral code, wallet address or UID and “find out anything to do with that person.” In the same recording, he describes initially researching 10-20 wallets and gradually increasing activity “so it does not look that suspicious.” — Oliver Knight Read more.
EF ‘STRAWMAP’ ROADMAP RELEASED: Ethereum Foundation published a roadmap that reads like it’s building for the next decade, not surviving the current quarter. The document, called the “strawmap” and released Wednesday by EF researcher Justin Drake, lays out a plan for seven hard forks through 2029. Hard forks are network-wide software upgrades that every node must implement or get left behind, making them the highest-stakes type of change Ethereum can make. The plan is organized around five goals described as “north stars.” These include a faster layer 1 with transaction finality in seconds; dramatically higher layer-1 throughput capable of around 10,000 transactions per second (referred to as “gigagas” scale); Layer-2 networks reaching “teragas” levels of throughput, or roughly 10 million TPS; post-quantum cryptography and built-in privacy through shielded ETH transfers. — Shaurya Malwa Read more.
ROBINHOOD CHAIN TESTNET UPDATE: Robinhood’s (HOOD) testnet logged 4 million transactions in the first week its testnet chain went live, the investment platform’s CEO Vlad Tenev said on X. The Robinhood Chain, which focuses on tokenization and trading, comes as centralized exchanges are looking to build their own blockchain infrastructure even as the broader Ethereum ecosystem debates its future. “Developers are already building on our L2, designed for tokenized real world assets and onchain financial services,” Tenev wrote. Testnets are risk-free environments for developers to test code and experimental features before the mainnet goes live. The two stages of a network’s development can be compared to a flight simulator and a commercial flight. The Robinhood Chain’s testnet has arrived against the backdrop of a larger reckoning in the Ethereum world. Earlier this month, Ethereum co-founder Vitalik Buterin declared that the protocol’s long-held layer-2 (L2) rollup-centric roadmap “no longer makes sense,” arguing that many rollups have fallen short of full decentralization and that Ethereum’s base layer is scaling faster than expected. — Margaux Nijkerk Read more.
OPENAI DIPS ITS TOES INTO SMART CONTRACTS: OpenAI is stepping deeper into crypto security with the debut of EVMbench, a testing framework designed to measure how well artificial intelligence can understand and potentially secure smart contracts on Ethereum and similar blockchains. Smart contracts, the self-executing code deployed on blockchains like Ethereum, underpin decentralized exchanges, lending protocols and a wide range of onchain financial applications. Because these contracts are typically immutable once deployed, vulnerabilities can be serious. EVMbench is OpenAI’s attempt to see whether modern AI systems are up to the task of helping prevent those issues. Built in collaboration with crypto investment firm Paradigm, the benchmark draws on real-world smart contract vulnerabilities already uncovered through audits and security competitions. The system measures performance across three core abilities: identifying security bugs, exploiting those bugs in a controlled environment and fixing the vulnerable code without breaking the contracts. OpenAI says the goal is to establish a clear standard for evaluating AI systems in blockchain security, especially as decentralized finance continues to secure billions of dollars in user funds. The stakes for smart contracts are only rising. — Margaux Nijkerk Read more.
