Ethereum’s second Glamsterdam hard fork testnet was cancelled after ChainSecurity identified vulnerabilities in gas scaling proposals EIP-8037 and EIP-8038. These flaws could have disrupted at least 13 DeFi contracts by breaking logic that relies on current gas behavior when calls run out of fuel, potentially allowing attacker exploitation. Derek Chiang from Eth Labs noted the discovery occurred late in the process, prompting a rapid response. To address the issue, researchers from Eth Labs and the Ethereum Foundation deployed AI agents to scan for vulnerabilities. The teams coordinated their efforts via a GitHub repository message board, a method inspired by recent rogue agent tactics but repurposed for security defense. This approach proved effective enough to trigger a second initiative examining EIP-8198 for the upcoming Hegota fork.
In other developments, Solana activated its Alpenglow consensus upgrade on mainnet, reducing transaction finality time from approximately 12.8 seconds to 150 milliseconds. The first phase, Votor, replaces voting mechanisms with direct validator voting and certificates, expected to cut headline transaction counts by 60%. Meanwhile, Starknet community governance voted overwhelmingly (93.52%) to support research into becoming an independent, post-quantum layer-1 blockchain, though this requires a separate binding vote for implementation. Senator Richard Blumenthal requested information from Tether CEO Paolo Ardoino regarding a $6.3 million marketing deal linked to former White House adviser Bo Hines, citing potential ethics violations. Bitcoin Core also updated its disclosure policy, shortening embargo periods because large language models can now rapidly rediscover software flaws.
The cancellation of the Glamsterdam testnet highlights the increasing complexity of Ethereum's roadmap and the critical role of automated security auditing. Using AI agents to identify deep-seated protocol vulnerabilities represents a shift in how major blockchains manage risk, moving beyond traditional manual audits to scalable, machine-driven surveillance. This incident underscores that even well-tested upgrades carry significant operational risks if edge cases in gas mechanics are not fully modeled before deployment.
Simultaneously, the regulatory scrutiny facing Tether and the strategic pivot by Starknet illustrate diverging pressures on crypto infrastructure. While Ethereum focuses on technical robustness through AI-assisted development, other entities face external compliance challenges or internal governance shifts toward independence. The rapid adoption of AI tools for both offense and defense suggests that future network stability will depend heavily on the ability to detect and patch vulnerabilities faster than malicious actors can exploit them.
