A One-Cent Payment Revealed How Coinbase's x402 Service Catalog Actually Updates

Developer tool ScrapeCheck discovered that settlements — not deployments, metadata edits, or validation runs — are what trigger updates to listings on Coinbase's x402 Bazaar, a machine-readable catalog of paid services. The finding emerged after an unknown wallet paid the service one cent on August 25, causing the catalog listing to refresh within 4.3 seconds. Through repeated instrumented observations, the team confirmed that catalog rows consistently updated within one to five seconds of a payment settling, while all other actions produced no change. The team also noted a discovery gap in the catalog: its default API surfaces roughly 110 curated rows out of over 15,000 total listings, with two providers dominating nearly half that visible shelf. One anomalous creation event took far longer than the rest, and two formal predictions made by the team failed, leaving the full indexing pipeline behavior still not entirely explained.

Alibaba Releases Qwen3.8-Flash-Next as Early Architectural Preview of Qwen4

Alibaba's Qwen team released Qwen3.8-Flash-Next on August 26, 2026, an open-weight multimodal mixture-of-experts model designed as an architectural preview of the upcoming Qwen4 family. The model has 125 billion total parameters but activates only 6 billion per token, plus a separate 51-billion-parameter N-gram embedding layer that can run in system RAM rather than GPU memory. Alibaba claims training costs were roughly one-ninth those of the 397-billion-parameter Qwen3.7-Plus model. Vendor-reported benchmarks show strong performance on agentic coding tasks, though the model trails Claude Opus 5 significantly on computer-use evaluations such as OSWorld 2.0. The release follows Alibaba's stated strategy of publishing architectural experiments with public weights before launching a full model series.

Needle Benchmark Tests Search Engines Without Allowing Memorization

Keenable AI has introduced Needle, a new benchmark designed to evaluate search engine performance in a way that prevents engines from gaming results through memorization. Traditional benchmarks can become ineffective over time as systems learn to optimize specifically for their test cases. Needle aims to address this limitation by generating evaluations that cannot be anticipated or memorized in advance. The benchmark is intended to provide a more honest and reliable measure of true search quality. Details were published on the Keenable AI blog, though the post has so far attracted limited community discussion.

US attack submarine repair delays cost billions

Corporate Political Donations Hit Record $646M Ahead of US Midterms

Corporate political donations in the United States have reached a record $646 million ahead of the 2026 midterm elections, surpassing all previous benchmarks. The figure represents contributions tallied so far, suggesting the final total could climb even higher as the election cycle continues. The surge highlights the growing financial influence of corporations in American electoral politics. Reuters reported the data, drawing attention to the scale of corporate involvement in the upcoming midterms.

How Analysts Attribute DEX Trade Origins Using On-Chain Data in Dune SQL

A technical excerpt from 'The DeFi Data Quick-Start' explains the challenge of tracing the origin of decentralized exchange trades on platforms like Dune Analytics. Every trade recorded in the dex.trades table captures chain, venue, token, and amount data, but omits which frontend or application initiated the transaction. This gap matters significantly because incentive programs and market-share claims depend on knowing whether a trade came from a human using a wallet app or an automated bot. Analysts use a method called attribution, which involves examining call data tags, router addresses, and transaction structure to infer trade origins indirectly. Some protocols like 0x and 1inch embed identifying markers in transaction data, making attribution possible through SQL joins on transaction hashes.

The terrifying mechanics of the Nepali flash flood

Article URL: https://www.economist.com/science-and-technology/2026/08/27/the-terrifying-mechanics-of-the-nepali-flash-flood Comments URL: https://news.ycombinator.com/item?id=49466188 Points: 26 # Comments: 0

Crypto for Advisors: How staking on Ethereum is changing in 2026

NASA Astronaut Jonny Kim Leaves Agency to Resume U.S. Navy Service

Astronaut Jonny Kim officially departed NASA on Thursday after nearly a decade with the agency. He completed an eight-month science mission aboard the International Space Station as part of his tenure. Kim had launched to the ISS in April 2025 aboard the Russian Soyuz MS-27 spacecraft. Following his departure, he will continue his career as a lieutenant commander in the U.S. Navy.

AgentConnect Proposes Permission Model for Multi-Agent AI Systems

A project called AgentConnect has been shared on Hacker News, arguing that sandboxing alone is insufficient as a security model for multi-agent AI systems. The repository suggests that dedicated permission frameworks are needed to govern how autonomous agents interact with each other and with external resources. The core argument is that traditional sandboxing addresses containment but does not adequately control what actions agents are authorized to perform. The project appears to propose an alternative or complementary approach to managing trust and permissions in multi-agent architectures.

Nvidia forecasts $673B in cumulative sales driven by surging AI demand

Nvidia has projected $673 billion in sales as demand for artificial intelligence infrastructure continues to accelerate. The chipmaker's forecast reflects the growing reliance on its hardware across data centers, cloud platforms, and AI development pipelines. The projection underscores Nvidia's dominant position in the AI chip market, where its GPUs have become essential for training and running large-scale AI models. The forecast signals continued investor and industry confidence in AI as a long-term growth driver for the semiconductor sector.

Cloud AI Coding Agent Founders Read Every Message From All Six Early Users

A small team building a cloud AI coding agent manually reviewed every message sent by their first six users over 21 days, uncovering critical product flaws. A key bug caused the agent to silently report "nothing happened" after spending over two minutes and 254,000 tokens reading a codebase — because the progress tracker only counted file writes and shell commands, ignoring reads and searches. A separate issue sent users a blank white screen instead of a live preview due to a missing base path in scaffolded Vite apps — a fix that had been sitting in a branch but never pushed to production. On a more positive note, one user returned unprompted nearly 13 hours later to continue building after the agent generated a substantial application skeleton, with no email nudge or onboarding prompt. The team concluded that their biggest obstacle was not agent reliability alone, but their own inability to distinguish meaningful automated work from unproductive thrashing in their logs.

Peter Cullen, Iconic Voice Behind Optimus Prime and Eeyore, Dies at 85

Peter Cullen, the veteran voice actor best known for bringing Transformers hero Optimus Prime to life, has passed away at the age of 85. Cullen originated the role of Optimus Prime in the 1984 animated Transformers series and continued voicing the character across decades of films and media. Beyond Transformers, he was also widely recognized for his portrayal of Eeyore in Winnie the Pooh productions. His distinctive, commanding voice made Optimus Prime one of the most recognizable characters in animation and pop culture history.

Developer Decompiles a Nintendo 64 Game in Just 84 Days

A developer named Chris Lewis documented the process of decompiling a Nintendo 64 game over a span of 84 days. The project involved reverse-engineering the game's compiled code back into a human-readable source form. Lewis shared the detailed journey on his personal blog, outlining the technical challenges and methods used throughout the process. The post attracted attention on Hacker News, where it was submitted as a notable technical achievement. Such decompilation efforts are significant in the retro gaming community for preservation and modding purposes.

How a Verification-First AI Harness Outperforms Bigger Models in Bug Repair

A software engineering team has published a method for integrating a language model into automated maintenance pipelines while strictly limiting its decision-making authority. The approach, derived from a production deployment managing roughly 1,400 extraction targets and 1.6 million monitored items, centers on a machine-checkable oracle that verifies every AI-proposed fix before it is applied. The team found that constraining prompts and narrowing evidence roughly doubled precision, while doubling model size yielded no measurable improvement. The method is designed for problems where artifacts are small and structured, failures are high in volume, and incorrect repairs would produce silent wrong output rather than obvious errors. The authors present the framework as domain-independent, noting it has been validated beyond web extraction in configuration migration and flaky-test repair scenarios.

Developer Tool Compares A2A Agent Cards Across AWS, GCP, and Azure Clouds

A developer tutorial published on August 25, 2026, demonstrates how to fetch and compare Agent Card metadata from A2A-protocol agents running on three major cloud platforms: AWS Bedrock AgentCore, Google Cloud Run, and Azure Container Apps. An Agent Card is a JSON document hosted at a standard path that describes an agent's name, skills, and endpoint details, and is the first thing a client reads before connecting. The project only performs discovery — it never invokes the agents, sends prompts, or incurs any token costs. A key finding is that only AWS Bedrock AgentCore separates card-discovery permissions from invocation permissions via a distinct IAM action, while Cloud Run and Azure Container Apps bundle both behind a single access role. The accompanying open-source repository includes injectable test transports to simulate edge cases such as authentication denials and legacy card paths without requiring live cloud environments.

Startup Raises $26M to Build Underground Infrastructure Mapping Platform

A startup founded by a former PG&E engineer is developing a mapping platform designed to chart underground infrastructure, often described as a 'Google Maps for the underground.' The company has secured $26 million in a Series A funding round to fuel its growth. The capital will be used to expand its customer base across utility and construction sectors. The platform aims to cut through bureaucratic delays and red tape that typically slow down underground utility and construction work.

Ripple Recommends Withdrawing XChainBridge Amendment XLS-38 After Low Adoption

Ripple has recommended that the XRP Ledger community withdraw the XChainBridge amendment (XLS-38) and the related fixXChainRewardRounding amendment, citing insufficient developer interest over a 12–15 month evaluation window. XLS-38 was originally designed to enable native cross-chain bridges on the XRPL, allowing assets to move between mainnet and custom sidechains via a decentralized network of witness servers. Ripple concluded that the witness server model posed difficult trade-offs between security, decentralization, and governance, particularly as bridge value scales. The company chose Axelar — a purpose-built bridging network with 75+ validators and experience across 55+ blockchains — to serve the XRPL EVM Sidechain instead, a decision announced in June 2024. With the EVM Sidechain bridge needs met by Axelar and no significant private sidechain projects emerging that required XLS-38, Ripple determined the amendment no longer warrants activation.

Gill confident India can still qualify for WTC final despite early struggles

Indian captain Shubman Gill has acknowledged the challenges of adjusting to his new leadership role. Despite the frustrations, Gill remains optimistic following a series win that yielded several positives for the team. He believes India still have a realistic chance of qualifying for the World Test Championship final. Gill is targeting strong performances in the upcoming series against New Zealand and the Border-Gavaskar Trophy against Australia early next year.

Hawaii launches multi-year drive to remove invasive red mangroves from Kailua wetlands

Environmental teams in Hawaii are planning a large-scale effort to eradicate invasive red mangroves across nearly two miles of waterways in Kailua. The project aims to protect endangered native waterbirds and reverse environmental degradation caused by the non-native species. Community volunteers and local organizations will play a key role in carrying out the extensive removal work. The initiative is backed by national foundation funding and is scheduled to begin in September 2026.