The global payments sector saw several major shifts during the week of Aug. 31 to Sept. 6, with stablecoins, bank charters, cross-border settlement rails and dealmaking all moving at once. A group of 21 financial institutions, including 17 global systemically important banks, said they would form a joint stablecoin company in the second half of 2026 and issue a U.S. dollar stablecoin in the first half of 2027, with a euro version to follow. In the U.S., the Office of the Comptroller of the Currency granted conditional preliminary approval on the same day to Revolut and Andreessen Horowitz-backed OpenReserve, and both outlined plans tied to issuing or distributing stablecoins. Bloomberg also reported that total stablecoin supply has fallen by roughly $15 billion from its mid-May peak, even as capital keeps flowing into payment and banking infrastructure. Elsewhere, Citi, FAB, OCBC and DBS completed live transactions on Swift’s blockchain-based shared ledger, including a weekend tokenized deposit payment from Singapore to New York that settled in minutes. The week’s developments pointed to a market where public-chain stablecoins, bank-led stablecoins and tokenized deposits are all competing for a larger role in payments and settlement.

YZi Labs on Aug. 25 unveiled the 24 startups selected for Season 4 of EASY Residency, its incubation program for early founders in Web3, AI and biotech. Each company will receive a $500,000 investment. In coverage of the cohort, Odaily identified nine names that already offer some form of participation path for users, ranging from points campaigns and waitlist applications to product trials and onchain deposits. The list includes Neutral Trade, Primus, Roostoo, SmartX, Alloco, XHunt, Nara, ElfomoFi and Liquorice. Among them, Neutral Trade has a live points system tied to its Autopilot Deposit campaign, while Primus has launched a Kaito Warm Up content campaign with an $8,000 reward pool. Other projects are at earlier stages, such as waitlist sign-ups, Telegram community access, browser extension usage or deposit-only product access without a points program. Odaily said the projects were ordered by the current number of tasks users can complete.

Shopify CEO Tobi Lütke said he is considering banning Claude Code inside the company, not because he thinks the tool is bad, but because it insists on using its own CLAUDE.md file instead of sharing the AGENTS.md setup used by other agents. Shopify runs multiple coding agents, including Claude Code and Codex, and Lütke argued that maintaining two parallel rule files becomes a real operational problem in large repositories where recursive file loading can create sync issues. Shopify has already built automated checks to catch those mismatches, a step Lütke sees as unnecessary overhead. Claude Code team member Thariq replied that the team will make AGENTS.md easier to use, but he also defended Claude-specific prompt structures, Skills, and CLAUDE.md on the grounds that different models benefit from different optimizations. The near-term workaround, he said, is still to reference AGENTS.md from within CLAUDE.md. That did not address Lütke’s core complaint. The discussion also drew support from figures including DHH, Jesse Pollak, Chamath Palihapitiya, and Kun Chen, while Grok said a review of previously leaked Claude Code source code suggested the CLAUDE.md filename is hardcoded and could be swapped to AGENTS.md without a major architectural rewrite.

Matt Palmer, head of developer experience and education at Cursor, outlined how Grok Bot works in an early beta write-up, describing it as an AI agent with its own always-on cloud computer. Rather than acting only inside a chat window, the system can open webpages and apps, access files, take screenshots, reuse browser sessions, and hand control back to the user when it hits login walls such as SSO, 2FA, captchas, or payments. Palmer said the setup felt closer to what a personal agent should actually be, compared with earlier tools that required too much deployment work, account management, and constant maintenance. The article breaks the product into several parts: persistent cloud execution on a Linux VM, layered memory across users, agents, and projects, trigger-based automation through messages, schedules, Slack, and git events, plus integrations with the same connectors and skills used by Cursor, including Notion, Slack, GitHub, MCP servers, Cursor Cloud Agents, and GitHub MCP. Palmer also described guardrails such as natural-language rules, a separate review agent, and allow/block lists. He gave examples ranging from demo generation and internal content monitoring to grocery comparison across Instacart and Amazon delivery, and DoorDash group-order navigation. After several weeks of use, he said the system had gradually earned his trust.

OpenAI has expanded Codex’s migration tools to pull in developer assets from rival AI coding products, including Claude Code, Claude Cowork, and Cursor. On Aug. 11, the company consolidated its external agent import documentation and surfaced new import options inside the ChatGPT desktop app and the Codex CLI. Users can bring over local configuration files, skills, plugins, project data, MCP server settings, and chats from the past 30 days without changing or deleting the original agent setup in the source tool. The import map covers a broad range of artifacts. CLAUDE.md becomes AGENTS.md, settings.json is converted into config.toml, slash commands are turned into skills, and local project memory is migrated into Codex memories. Imported conversations are also handled with an automatic compression step when they are too long to fit the available context window. The larger limitation is directionality. OpenAI’s documentation includes import workflows, but no equivalent export path from Codex. Changes made in Codex do not sync back to Claude Code. Several categories also remain hard to migrate cleanly, including fine-grained permissions, complex hooks, Anthropic-native model behavior, and web-based chat history. The update shows how competition in AI coding is shifting from raw model performance toward control over the accumulated workflows and local assets developers build over time.

A joint report from non-custodial crypto platform ChangeNOW and digital asset management platform CoinRabbit argues that crypto privacy tools provide essential protective functions rather than existing only as vehicles for illicit finance. Drawing on data from TRM Labs, Chainalysis, RAND Corporation and the firms’ own research, the paper says public blockchain transparency can expose ordinary users, companies and wealthy holders to risks that would normally remain private in traditional finance. The report points to sanctions-related financial exclusion, corporate wallet surveillance and the rise of wrench attacks as examples of where privacy can shield users from harm. It also cites CertiK data showing 52 verified wrench attacks in the first half of 2026, with $124.1 million exposed, and notes that France accounted for 33 of those cases. At the same time, the authors acknowledge that privacy tools can be abused, referencing TRM Labs estimates of $158 billion in illicit crypto inflows in 2025 and $75 billion in pig-butchering losses between 2020 and 2024. Even so, the report argues that enforcement leverage sits mainly at fiat off-ramps, not in upstream privacy infrastructure, a view echoed by AQ Forensics founder Albert Quehenberger.

Axis Robotics said on July 27 that it had raised a $12 million seed round led by Hack VC, with participation from Nomad Capital, Pi Network Ventures and 10K Ventures. The company positions itself as infrastructure for physical AI rather than a model or hardware builder, using browser-based teleoperation to let users collect robot interaction trajectories without owning robots. Its system pairs remote simulation tasks with a Web3-based attribution layer on Base. Each approved trajectory receives a Data ID tied to a contributor wallet, submission time, quality score and task metadata, while the full data stays off-chain. Axis says this structure is meant to turn contributor output into verifiable records that can later support reward distribution, licensing or governance. The incentive stack now includes POINTS, possible fiat revenue sharing from enterprise task packages, and a token that has not yet launched. The company has disclosed community experiments, academic benchmarks and a list of partners, but key business details such as revenue-sharing ratios, contract values and repayment data remain undisclosed. That leaves a central question unresolved: whether enterprise demand can become recurring enough to support the broader incentive loop.

OpenAI engineers said AI agents generated about 1 million lines of code in five months, with humans writing none directly. The approach, called Harness Engineering, focuses on context, constraints, and feedback systems around the model.