# Stability Related Articles

HTX News Center provides the latest articles and in-depth analysis on "Stability", covering market trends, project updates, tech developments, and regulatory policies in the crypto industry.

Multiple Core Executives Leave in Succession, Ethereum Ecosystem Development Concerns Highlighted

Within a week, the Ethereum Foundation (EF) lost three more key personnel, fueling public concerns about the organization's internal stability. Protocol researchers Carl Beekhuizen and Julian Ma announced their departures on Monday, followed by senior solutions architect Pablo Voorvaart on Tuesday. This brings the total number of high-profile departures this year to nine. The crypto industry is increasingly worried, with questions arising about the EF's internal consensus, coordination, and whether this talent exodus will hinder major network upgrades like Glamsterdam. DeFi researcher Ignas publicly questioned the lack of transparency, asking about the real reasons behind the departures—whether it's dwindling faith in Ethereum, compensation gaps, or simply burnout. Community reactions are mixed. Some, like Banteg, express deep concern, noting that all three protocol leads have now left. Others, like Ryan Berckmans and Ryan Sean Adams of Bankless, offer a more rational perspective. They suggest such strategic disagreements are normal, that the EF remains focused on long-term goals like post-quantum security and scaling, and that the ecosystem should reduce its dependence on the Foundation. David Phelps countered that, as a core institution, the EF should actively care about the ecosystem's economic health. This wave of departures follows earlier signs of turmoil. Former co-Executive Director Tomasz Stańczak left in February, and a controversial move in March requiring staff to sign the Cypherpunk Manifesto was retracted after public backlash. Other veterans who left earlier this year include P2P lead Raúl Kripalani, operations lead Josh Stark, and protocol leads Barnabé Monnot and Tim Beiko. The departing members are highly experienced. Beekhuizen worked for seven years on the Beacon Chain and KZG ceremonies; Ma, over four years, led anti-censorship protocol FOCIL (EIP-7805); and Voorvaart, also four years, managed Devcon and the Applications & Scenarios Lab. Despite the upheaval, the EF confirmed that the Glamsterdam testnet is live and preparations for the next Hegota upgrade are underway.

marsbit05/21 07:42

Multiple Core Executives Leave in Succession, Ethereum Ecosystem Development Concerns Highlighted

marsbit05/21 07:42

Turing Award Laureate Sutton's New Work: Using a Formula from 1967 to Solve a Major Flaw in Streaming Reinforcement Learning

New research titled "Intentional Updates for Streaming Reinforcement Learning" (arXiv:2604.19033v1), involving Turing Award laureate Richard Sutton, addresses a core challenge in deep reinforcement learning (RL): the "stream barrier." Current deep RL methods typically rely on replay buffers and batch training for stability, failing catastrophically when learning online from single data points (streaming). The authors propose a fundamental shift: instead of prescribing how far to move parameters (a fixed step size), their "Intentional Updates" method specifies the desired change in the function's output (e.g., a 5% reduction in value prediction error). It then calculates the step size needed to achieve that intent. This idea is inspired by the Normalized Least Mean Squares (NLMS) algorithm from 1967. Applied to value and policy learning, this yields algorithms like Intentional TD(λ) and Intentional AC. The method inherently stabilizes learning by adapting the step size based on the local gradient landscape, preventing overshooting/undershooting. In experiments on MuJoCo continuous control and Atari discrete tasks, Intentional AC achieved performance rivaling batch-based algorithms like SAC in a streaming setting (batch size=1, no replay buffer), while being ~140x more computationally efficient per update. The work demonstrates significant robustness, reducing reliance on numerous stabilization tricks. A remaining challenge is bias in policy updates due to action-dependent step sizes. Overall, this approach advances efficient, online, "learn-as-you-go" RL, enabling adaptive systems without massive data buffers or compute clusters.

marsbit05/10 06:28

Turing Award Laureate Sutton's New Work: Using a Formula from 1967 to Solve a Major Flaw in Streaming Reinforcement Learning

marsbit05/10 06:28

活动图片