Today on The Anvil, we are watching the formal collapse of the diplomatic track in the Middle East, as the US-Iran conflict escalates into a multi-country exchange and a renewed naval blockade. Concurrently, the competition to capture the enterprise AI developer is accelerating, with Oracle, Figma, and GitHub simultaneously deploying tools that embed autonomous agents directly into core engineering workflows.
Oracle is extending its Fusion Agentic Applications to professional developers, allowing them to use existing tools like VS Code, Git, and AI coding agents (including Claude Code, Codex, and Gemini) to build deeply integrated applications. These applications run natively within Oracle Fusion Cloud Applications, inheriting its security, governance, and data models. The move follows an earlier announcement on Tuesday of a new AI-native builder experience for creating these 'Fusion Agentic Applications'.
Why it matters
This move bridges the gap between no-code AI builders and pro-code enterprise development. By allowing professional developers to use their preferred toolchains to create autonomous agents that operate securely within the Oracle ecosystem, Oracle is creating a powerful, governed platform for automating complex business processes across finance, HR, and supply chain. For a product builder, this signals a significant trend of embedding agentic AI directly into core enterprise systems rather than using it as a peripheral tool.
Researchers have unveiled AIDE², a system that marks the first experimental evidence of consistent recursive self-improvement (RSI) in AI. The system autonomously optimized its own code, discovering a more efficient search algorithm, shrinking its prompt size by 16x, and adding defenses against reward hacking, all within eight days. Its final performance surpassed a version that had been manually tuned by engineers for over two years.
Why it matters
This is a significant milestone in AI research, moving beyond systems that are simply powerful to ones that can accelerate their own development. The ability of an AI to improve its own architecture and efficiency more effectively than human engineers suggests a potential paradigm shift, where the pace of AI advancement could begin to increase non-linearly.
Independent research from MIT, Tencent, and Huawei has converged on 'continual learning' as the next critical bottleneck for frontier AI. The research highlights that current models are static and suffer from 'catastrophic forgetting'—losing old knowledge when trained on new data. The consensus is that architectural solutions for models that can evolve are now more important than just scaling up model size.
Why it matters
This marks a significant intellectual pivot in the AI research community. For product builders, it signals that future AI systems will need to be designed for continuous adaptation in real-world environments. The challenge shifts from deploying a fixed model to architecting a system that can learn over time without degrading, a much harder but more valuable proposition.
Coinbase executives reported on Wednesday that over 95% of the company's code is now written by or with large language models, a dramatic increase from 40% in February. The rapid adoption follows layoffs in May, with the company linking the shift to increased development speed and a reorganization toward smaller, more senior teams managed by AI agents. Separately, Microsoft CEO Satya Nadella said AI now generates 20-30% of code in the company's repositories.
Why it matters
This marks one of the most aggressive adoption rates for AI in coding reported by a major public company, suggesting a dramatic acceleration in the trend of AI-integrated software development. This real-world data point provides a concrete look at how companies are restructuring engineering teams and workflows around AI agents, shifting the role of human developers toward oversight and architectural guidance.
Building on the recent free release of the Copilot standalone desktop app, GitHub has launched a public preview of a `/security-review` slash command within the client. The feature provides AI-driven vulnerability scanning for uncommitted code changes, offering actionable suggestions to catch flaws like injection and cross-site scripting before they reach a repository.
Why it matters
This is a significant 'shift left' for security, integrating vulnerability scanning directly into the developer's immediate workflow. By catching security flaws in the 'flight code' before a commit, it addresses a major concern with AI-generated code—that it can accelerate the introduction of vulnerabilities. For a product builder, this tool helps institutionalize security as part of the core development loop, not just a later-stage check.
Following its recent acquisition of AI coding agent startup Bud, Figma is repositioning itself as a 'full-stack creation platform.' A new 'code layers' feature allows developers to handle code directly on the design canvas, aiming to unify design, code, and AI into a single workflow.
Why it matters
This represents a significant step towards blurring the lines between design and development, a core challenge for design engineers. By bringing functional code into the design environment, Figma aims to create a true single source of truth and eliminate the costly, time-consuming handoff process. For product builders, this could enable faster, more collaborative iteration cycles and reduce the drift between design intent and the final shipped product.
Meta has open-sourced Astryx, its internal design system built on React and StyleX. The system, which has been used to build over 13,000 internal apps over eight years, provides over 150 accessible components, brand-level theming with dark mode, and a CLI. Meta positions it as a bridge between human-led and AI-assisted development.
Why it matters
The release of a mature, battle-tested design system from a company at Meta's scale is a significant event for the design engineering community. Astryx offers a robust, component-based foundation for building scalable web applications. Its explicit design for both human and AI developers makes it a noteworthy tool for teams looking to integrate generative UI workflows.
Dave Bozeman, CEO of logistics giant C.H. Robinson, reports a 45% increase in employee productivity since 2022, attributing it to the deployment of in-house AI agents. According to the company, these systems, built by a team of 450 engineers using open-source models, automate tasks like generating customer quotes, reducing processing time from 20 minutes to just 31 seconds. The firm also credits the AI deployment for double-digit EPS growth since 2023.
Why it matters
This is a powerful case study demonstrating the tangible ROI from a deep, internal investment in AI within a legacy industry. C.H. Robinson's success in building its own solutions, rather than just buying off-the-shelf products, provides a playbook for how large enterprises can leverage AI to drive significant productivity gains, margin expansion, and revenue growth in core operations.
Following up on the July 4th 'TikTok takeover' that led to hundreds of arrests, Newport Beach officials are advancing specific preventative measures in public meetings. Proposals include enhanced curfews for juveniles and stricter rules for short-term lodging, while neighboring cities like Huntington Beach implement their own zero-tolerance countermeasures.
Why it matters
The regional response signals a critical moment for Orange County coastal cities as they grapple with the public safety challenges of social media-fueled mass gatherings. The potential changes to short-term rental policies and public space regulations could significantly impact the local tourism economy, real estate investors, and community life.
Spokane area officials and housing experts are working to streamline the permitting process to encourage the development of 'middle housing' like duplexes and townhomes. The effort aims to address the region's housing affordability crisis by adapting regulations to allow more diverse housing options within existing neighborhoods as land for new single-family homes becomes scarcer and more expensive.
Why it matters
This represents a significant policy shift in Spokane's urban planning, moving away from outward sprawl and toward infill development. Addressing the 'missing middle' of housing is critical for the region's economic health and could offer a template for how similar mid-sized cities handle growth and affordability.
As the reinstated naval blockade and airstrikes enter their fourth consecutive night, Iran has expanded its retaliatory targets to US facilities in Bahrain, Kuwait, and Jordan. Tehran has officially declared the previously signed memorandum of understanding 'dismantled,' while the IRGC threatens to counter the blockade by halting all Middle East energy exports, directly targeting the Bab el-Mandeb Strait.
Why it matters
The formal collapse of the MoU confirms the death of the fragile diplomatic track we followed earlier this summer. The IRGC's threat to expand operations beyond the Strait of Hormuz to the Bab el-Mandeb Strait introduces a severe new risk vector for global energy markets and supply chain stability.
Enterprise AI Moves Deeper into Developer Workflow Major platform players like Oracle, Figma, and GitHub are embedding agentic AI directly into existing pro-code and design systems. Rather than just offering standalone tools, they are creating integrated environments where AI assists with everything from code generation and security review to chip design and platform engineering governance.
US-Iran Conflict Enters Fourth Day of Escalation The conflict has intensified for a fourth consecutive day. The US has reimposed a naval blockade and is conducting strikes, while Iran has responded by attacking US assets in multiple allied countries and threatening key global shipping lanes. The previously signed memorandum of understanding has been officially declared 'dismantled' by Tehran.
AI Coding Adoption Reaches New Heights New reports show enterprise adoption of AI for coding is accelerating dramatically. Coinbase now says over 95% of its code is written with AI assistance, while Microsoft's Satya Nadella states AI generates up to 30% of their code. This trend is driving the development of more advanced, integrated agentic coding platforms.
The New Reality of Local Public Safety Following the chaotic July 4th 'takeover,' Newport Beach and other Orange County cities are now actively debating stricter public safety measures. The proposals include enhanced curfews and new rules for short-term rentals, reflecting a growing regional challenge in managing social media-fueled mass gatherings.
AI Safety and Governance Debates Solidify As AI capabilities advance with systems demonstrating recursive self-improvement, the focus on safety is hardening. Google DeepMind's Demis Hassabis has proposed a detailed framework for an independent standards body, while new research identifies 'continual learning' as the next major architectural challenge to prevent knowledge decay in models.
What to Expect
2026-07-17—Google's delayed launch of Gemini 3.5 Pro is now targeted for this Friday following an architectural rebuild.
2026-07-19—The EU's ban on destroying unsold apparel and footwear for large companies is set to take effect this Sunday.
2026-09-23—The 8th Annual Keller Logistics Summit will be held in Defiance, Ohio, focusing on warehousing, freight, and AI.
How We Built This Briefing
Every story, researched.
Every story verified across multiple sources before publication.
🔍
Scanned
Across multiple search engines and news databases
470
📖
Read in full
Every article opened, read, and evaluated
176
⭐
Published today
Ranked by importance and verified across sources
11
— The Anvil
🎙 Listen as a podcast
Subscribe in your favorite podcast app to get each new briefing delivered automatically as audio.
Apple Podcasts
Library tab → ••• menu → Follow a Show by URL → paste