Network News
NEW SCALING PLANS FOR ETHEREUM: Ethereum co-founder Vitalik Buterin published a blog post on X outlining his latest vision for scaling the blockchain, arguing the network can boost capacity in the near term while laying the groundwork for a longer-term shift to advanced cryptography and>in which much of the ecosystem’s scaling strategy centered on layer-2 rollups. The plan comes on the heels of the Ethereum Foundation publishing a ‘strawmap’ aimed at making the network more efficient in the long term. In the short term, Buterin says Ethereum can safely increase throughput by making blocks easier and faster to check. Upcoming upgrades will allow the computers that run Ethereum to review different parts of a block simultaneously, rather than processing everything step by step. At the same time, changes to how blocks are built will let the network use more of each 12-second processing window, rather than finishing early out of caution (known as ePBS, and will be implemented in the Glamsterdam upgrade). The result: Ethereum should be able to fit more transactions into each block without increasing the risk of errors or instability. Another major piece of the plan involves rethinking how transaction fees — known as “gas” — are calculated. Buterin argues that not all activity on Ethereum puts the same strain on the network. There’s a big difference between using computing power temporarily and permanently adding new data that every Ethereum computer, or node, must store forever. — Margaux Nijkerk Read more.
OKX DABBLES WITH AI AGENTS: OKX rolled out an AI-focused upgrade to OnchainOS, its developer platform, pitching it as infrastructure for autonomous crypto trading agents. The AI layer builds on familiar components such as wallet infrastructure, liquidity routing and onchain data feeds, combining them into a unified execution framework aimed at AI agents operating across chains. Rather than wiring price feeds, token approvals, gas estimation and swap routing manually, developers can connect an agent and issue a high-level instruction, such as swapping $ETH for USDC below a certain price. OnchainOS handles the workflow behind the scenes, from monitoring markets to sourcing liquidity and confirming settlement. The intersection between crypto and AI has grown exponentially in the past 12 months — the blockchain AI market projected to rise from $6 billion in 2024 to $50 billion by 2030 — and traders are using the technology to their advantage. One recent example occurred when a group of retail traders used AI to find "glitches" on platforms like Polymarket before instructing AI to trade on its behalf. — Sam Reynolds Read more.
$NEAR FOUNDER ON THE FUTURE USERS OF BLOCKCHAIN: For years, the crypto industry has searched for its next breakout moment — something on the scale of DeFi summer or the NFT boom. Meanwhile, artificial intelligence (AI) has quietly become embedded in daily life. Developers use ChatGPT as a co-pilot. Consumers rely on AI assistants to draft emails, plan travel and, increasingly, manage workflows. Crypto, by comparison, still feels infrastructural. Illia Polosukhin, a co-founder of $NEAR, believes the divide is about to collapse, but not in the way many expect. “The users of blockchain will be AI agents,” Polosukhin said in an interview. “AI is going to be on the front end, and blockchain is going to be the back end.” His framing cuts against much of crypto’s recent experimentation with AI, which has centered on speculative tokens, memecoins and agent-themed trading bots. Instead, Polosukhin argues that AI will become the primary interface layer for everything online, including crypto, abstracting away wallets, explorers and transaction hashes. “The goal is to make your AI hide all the blockchain,” he said. “The fact that we have [blockchain] explorers is effectively a failure, because we don’t abstract the technology.” In this view, blockchain doesn’t disappear, it recedes. AI agents interact with protocols directly, executing payments, managing assets, coordinating services and even voting in governance systems. Humans, meanwhile, interact with the AI. — Margaux Nijkerk Read more.
coindesk.com