Former CFTC Chairman: After the Clarity Act stalled, crypto regulation urgently needs more durable rules, and the boundaries of the crypto market may be defined by the Supreme Court.

According to Coinglass data, as Bitcoin posts a minor rebound today, funding rates on major centralized (CEX) and decentralized (DEX) crypto exchanges show the market has returned to neutral-skewed bearish territory. Specific funding rates are detailed in the attached chart. BlockBeats Note: The funding rate is a mechanism set by crypto trading platforms to align perpetual contract prices with their underlying asset prices. It acts as a fund exchange between long and short traders, with no fee collected by the platform. It adjusts the cost or return for traders holding contracts, keeping contract prices closely tied to the underlying asset. A 0.01% funding rate is the benchmark. Rates above 0.01% signal broad bullish sentiment, while rates below 0.005% indicate general bearishness.

CryptoRank data shows that in Q3 2026, decentralized finance (DeFi) total value locked (TVL) rebounded to $95 billion after three consecutive quarters of decline, marking the first quarterly growth since its 2025 peak. Monad’s TVL surged 183%, propelling it into the ranks of leading blockchain networks by TVL. Robinhood Crypto, which launched just two months ago, has seen its TVL top $1 billion.

According to monitoring by ai_9684xtpa, after Bitget enabled ETH withdrawals, its associated ETH hot wallet saw a sustained outflow of ETH, followed by inflows, and its current balance has slightly exceeded the initial level before withdrawals were opened. The observer noted that address 0xED9…18771 is suspected to be the hot wallet Bitget prepared for handling this round of ETH withdrawals. Before withdrawals went live, this address held approximately 30,000 ETH; within half an hour, it experienced continuous outflows, dipping to around 25,000 ETH at one point, after which ETH inflows resumed. ai_9684xtpa speculates that two factors drove the ETH return: first, the prior resumption of BTC withdrawals had reassured users about Bitget’s withdrawal reliability, and second, Bitget’s ETH PoolX campaign (where users lock ETH to split airdrops) likely also contributed to the ETH inflows.

Belarus has granted registration approval to its first two crypto banks, which will provide fiat currency deposits, tokenized transfers, and digital asset custody services.

Beating AI Express (from Dongcha) reports that open-source Computer Use tool Cua has added an optional Perception visual extension to its Cua Driver. The Driver originally prioritized locating buttons via web structures and system accessibility trees, but these structures may be empty when dealing with Canvas, games, remote desktops, or custom-drawn applications. Perception directly reads screenshots, splitting the screen into regions with text, type, and position for the Agent to operate on. The first version uses OmniParser to identify icons and controls, and PP-OCR to read text. The entire parsing process runs on the local CPU, no GPU required, and screenshots are not transmitted over the network. Official tests show processing a single screenshot takes roughly 2–2.5 seconds on macOS, 3.5–4 seconds on Linux, and 8–9 seconds on Windows. It acts as a fallback when page structures and accessibility trees fail, and does not run visual parsing by default for every step. Cua also restricts the reuse of old screenshots for visual operations: each visual recognition is bound to the current screenshot’s `capture_id`, and subsequent clicks must originate from the same screenshot. If the screenshot is over 60 seconds old, or an operation has already been performed, the Driver will refuse to continue using it, preventing accidental clicks with outdated coordinates after the interface changes. Cua was already capable of visual positioning via `cua-som` and OmniParser in the past; now this functionality is directly integrated into Cua Driver, allowing upper-layer Agents to obtain recognition results via a unified interface. Even if the model itself cannot process images, it can decide on the next step based on these text and region data. Perception is not a default component: the OmniParser detector uses the AGPL-3.0 license, while Cua Driver remains under the MIT license.

Beating AI Insight News Brief: Anthropic has added an automatic tuning workflow to Claude Code’s official claude-api Skill. The build-eval module first generates test sets and scoring metrics from real conversations, bugs, or manual cases; hillclimb then enables Claude Code to iteratively modify system prompts, tool descriptions, models, and inference intensity. After each adjustment, tests are rerun: improvements are retained, while regressions are rolled back. To prevent Claude from overfitting exclusively to test cases, tuning is not performed on all samples at once. A subset of cases is used to identify issues and adjust configurations, while a separate batch is reserved to validate if modifications work on unseen problems. If the primary score improves but results on the reserved test set do not, the modification may be deemed overfitting and withdrawn. Anthropic demonstrated this workflow using 44 customer support tickets. Claude Code sequentially cleaned prompts, added business rules, tested different models, and reduced inference intensity, ultimately settling on a low-effort configuration for Sonnet 5. On 14 tickets not included in tuning, accuracy rose from 78.6% with the initial configuration to 90.5%, while costs fell to roughly one-fifth of their original level.