# Horizon AI Briefing — 2026-07-13 ## Top stories - **Anthropic Expands Aggressively: Claude Design, Creative Tools, Robotics, India Pricing, and Sydney Office** (https://www.anthropic.com/news/claude-design-anthropic-labs) — Anthropic launched Claude Design (a visual collaboration tool from Anthropic Labs), published robotics benchmarking research, began localizing pricing for India in rupees, and opened a Sydney office — all within 24 hours. This signals a multi-front expansion into creative workflows, physical AI, and emerging markets that positions Anthropic well beyond its safety-research origins. - **Christopher Nolan: Public 'Disdains' AI Despite Wall Street Embrace** (https://bsky.app/profile/theguardian.com/post/3mqik3rqjfv2q) — Director Christopher Nolan argues that AI has been enthusiastically adopted by investors and tech firms but met with broad public rejection, calling the idea that AI will replace humans 'nonsense.' The remarks from a culturally influential figure highlight the widening gap between industry optimism and public sentiment, which could accelerate regulatory and reputational pressure on AI companies. - **Richard Sutton Launches Oak Lab, Targeting a 20-Watt Trillion-Parameter AGI Agent** (https://oaklab.ai/mission) — Turing Award winner and RL pioneer Richard Sutton has founded Oak Lab, aimed at building a trillion-parameter agent that learns and plans in real-time using only 20 watts — inspired by his OaK (Options and Knowledge) architecture for AGI. This is a significant signal that foundational AI researchers are moving from academia to dedicated labs with explicit AGI mandates, echoing the early days of DeepMind and Anthropic. - **Claude Fable Solves 6-Month Physics Impasse for Leading Theoretical Physicist** (https://www.reddit.com/gallery/1uv399n) — Yuji Tachikawa, a world-leading theoretical physicist, publicly reported that Anthropic's Claude Fable resolved a problem that had stumped his team for six months. This is a high-profile, credible use-case validation for frontier AI in hard science, and raises the bar for what 'useful AI' means in research contexts. - **Chinese Tech Giants Race to Build Their Own Coding Agents** (https://www.recodechinaai.com/p/chinese-ai-companies-are-building) — Alibaba, Tencent, ByteDance, Moonshot, and Z.ai are all shipping coding agents modeled on Claude Code, driven by both revenue and deeper strategic goals around developer ecosystem lock-in. The rapid Chinese response to Western agentic coding tools signals that AI coding infrastructure is becoming a competitive battleground with geopolitical dimensions. - **Helsing Raises $1.8B Series E at $18B Valuation, Cementing European Defence AI** (https://tech.eu/2026/07/13/european-defence-ai-leader-helsing-secures-18b-series-e-at-18b-valuation/) — European defence AI company Helsing closed a $1.8 billion Series E, reaching an $18 billion valuation and becoming Europe's most valuable defence tech firm. The raise reflects surging institutional appetite for AI applied to defence, particularly in a European context where governments are rapidly increasing defence spending. - **Xi Jinping to Attend World AI Conference for First Time, Signaling China's Escalating AI Ambitions** (https://www.scmp.com/tech/article/3360404/xi-jinping-attend-world-ai-conference-first-time-china-elevates-tech-push?utm_source=rss_feed) — President Xi Jinping will personally attend and keynote the 2026 World Artificial Intelligence Conference in Shanghai — his first appearance since the event launched in 2018. The move signals that AI has risen to the highest levels of Chinese state priority, with implications for global AI governance, standards-setting, and geopolitical competition. - **Apple's Lawsuit Is Already Damaging OpenAI Before Trial** (https://thenextweb.com/news/apple-openai-lawsuit-disrupts-hardware-ipo) — Apple's trade-secrets lawsuit against OpenAI is producing reputational and operational damage well ahead of any courtroom verdict, with damaging internal communications already surfacing publicly. The ongoing legal cloud could complicate OpenAI's enterprise sales, partnerships, and IPO ambitions even if Apple ultimately loses. - **AI 'Never-Skilling' Threat: Junior Workers May Never Develop Core Competencies** (https://thenextweb.com/news/ai-never-skilling-critical-thinking-research) — New research formalizes the concept of 'never-skilling' — where junior employees who rely on AI from the start never build foundational expertise, unlike 'deskilling' which affects experienced workers. This is an emerging workforce risk that organizations deploying AI broadly need to actively manage through structured learning programs. - **Zhipu AI Pivots to AGI Research Over Short-Term Commercialization** (https://technode.com/2026/07/13/zhipu-ai-founder-outlines-touch-high-plan-for-agi-research/) — Chinese AI lab Zhipu AI announced its 'Touch High' plan, deprioritizing near-term revenue in favor of AGI research across four technical pillars, even as rival MiniMax's stock dropped nearly 20%. The move mirrors Western lab strategies and suggests Chinese frontier AI is maturing beyond the race for quick commercial wins. ## Emerging signals - **AI 'Penalty' at Work: Employers Credit AI, Not Employees, for AI-Assisted Output** (https://thenextweb.com/news/ai-penalty-workers-hiding-ai-use-promotions) — Workers report losing promotions and raises when managers attribute AI-assisted work to the tool rather than the human, a phenomenon researchers are calling the 'AI penalty.' As AI becomes mandatory in more workplaces, this attribution problem could trigger significant labor relations and HR policy battles. - **Defensive Prompt Injection as a Security Tool** (https://arstechnica.com/security/2026/07/now-defenders-are-embracing-the-prompt-injection-too/) — Security researchers are now exploring 'context bombing' — using prompt injection offensively against hacking agents to make them shut down before causing harm. This inversion of a known attack vector into a defensive technique is an early signal of a maturing AI security discipline. - **AI Memory Market Crunch Won't Ease Until 2028** (https://thenextweb.com/news/ai-memory-crunch-boom-bust-2028) — The AI boom has inverted normal DRAM/NAND price dynamics, with memory prices rising instead of falling, and analysts say the crunch will persist until 2028. This supply constraint is a structural bottleneck that will affect the pace of AI infrastructure buildout globally. - **Chinese Internet Giants Sign Self-Regulatory AI Agent Data Pact** (https://technode.com/2026/07/13/chinese-internet-firms-sign-ai-agent-data-protection-pact/) — Baidu, Tencent, Alibaba, Volcengine, and 27 others signed a China Internet Association pact on personal data protection for AI agents. Self-regulatory frameworks ahead of formal regulation are a pattern worth watching as they often set de facto standards that shape later government rules. - **Native AI Agent Smartphones Debut at WAIC 2026** (https://pandaily.com/ai-agent-phones-nubia-bytedance-waic-jul2026) — Nubia unveiled what it claims is the world's first native AI agent smartphone at WAIC 2026, moving beyond AI feature add-ons to fully autonomous agent systems that operate across apps. If this form factor gains traction, it could reshape the smartphone industry the way app stores did in 2008. ## New entrants - **Oak Lab** (company) — Richard Sutton's new AI research lab targeting a trillion-parameter, 20-watt real-time learning agent based on the OaK (Options and Knowledge) AGI architecture. Founded by one of the most influential figures in reinforcement learning. - **Claude Design** (tool) — An Anthropic Labs product enabling users to collaborate with Claude to create visual designs, prototypes, slides, and one-pagers — marking Anthropic's entry into the visual/creative workflow market. - **Claude Fable** (model) — An Anthropic model (likely a frontier research variant) that resolved a six-month-old unsolved theoretical physics problem for a leading physicist, suggesting strong scientific reasoning capabilities beyond prior Claude releases. - **SingGuard-NSFA** (model) — Ant Group's open-sourced safety guardrail model for autonomous AI agents, designed to detect prompt injection, data theft, malicious code execution, and permission misuse before agents act. - **Nubia AI Agent Smartphone** (tool) — Claimed world-first native AI agent smartphone debuted at WAIC 2026, featuring autonomous agent systems that understand, execute, and remember tasks across apps — going beyond feature-level AI integration. ## Biggest movers this week - **OpenAI** (company) — 539 mentions this week, ↑297 vs the prior week - **ChatGPT** (model) — 172 mentions this week, ↑92 vs the prior week - **Apple** (company) — 114 mentions this week, ↑86 vs the prior week - **GPT-5.6** (model) — 64 mentions this week, ↑58 vs the prior week - **NVIDIA** (company) — 77 mentions this week, ↑40 vs the prior week - **Google** (company) — 157 mentions this week, ↑36 vs the prior week ## China & East-Asia AI - **🖥️Chinese AI Companies Are Building Their Own Claude Code** (https://www.recodechinaai.com/p/chinese-ai-companies-are-building) — rss - **Nvidia’s future challenger? Chinese start-up reveals aggressive AI chip road map** (https://www.scmp.com/tech/tech-war/article/3360415/nvidias-future-challenger-chinese-start-reveals-aggressive-ai-chip-road-map?utm_source=rss_feed) — rss - **China works on AI safety benchmark as regulators target large model risks** (https://www.scmp.com/tech/article/3360399/china-works-ai-safety-benchmark-regulators-target-large-model-risks?utm_source=rss_feed) — rss - **Xi Jinping to attend World AI Conference for first time as China elevates tech push** (https://www.scmp.com/tech/article/3360404/xi-jinping-attend-world-ai-conference-first-time-china-elevates-tech-push?utm_source=rss_feed) — rss - **China’s drug industry pivots to AI-powered candidates to drive next wave of deals** (https://www.scmp.com/business/china-business/article/3360395/chinas-drug-industry-pivots-ai-powered-candidates-drive-next-wave-deals?utm_source=rss_feed) — rss ## Europe (EU) AI - **The next creator economy won’t be built around entertainment** (https://www.eu-startups.com/2026/07/the-next-creator-economy-wont-be-built-around-entertainment/) — rss - **Anthropic recruits Monzo cofounder Tom Blomfield for AI compute team** (https://sifted.eu/articles/anthropic-monzo-founder-tom-blomfield/) — rss - **Austria’s Sodex Innovations secures €4 million to scale automated construction-site surveying** (https://www.eu-startups.com/2026/07/austrias-sodex-innovations-secures-e4-million-to-scale-automated-construction-site-surveying/) — rss - **European defencetech leader Helsing secures $1.8B Series E at $18B valuation** (https://tech.eu/2026/07/13/european-defence-ai-leader-helsing-secures-18b-series-e-at-18b-valuation/) — rss - **'increasing use of artificial intelligence in research requires the sector and policymakers to take action, including supporting ways to assess the “p** (https://bsky.app/profile/eicathomefinn.bsky.social/post/3mqjg5ojjb22z) — bluesky ## Regulation updates - [US State] **Public schools; universities; AI policies** — Proposed. Tracked - [US State] **Prohibits using artificial intelligence to create child sexual abuse materials. (8/1/26)** — Passed. Tracked - [US State] **WARP Act of 2026 Warfighter Artificial Intelligence Readiness and Preparedness Act of 2026** — Proposed. Tracked - [US State] **An Act Requiring The Department Of Economic And Community Development To Develop A Plan To Establish An Artificial Intelligence Small Business Program.** — Passed. Tracked - [US State] **Prohibits public contracts with entities owned or controlled by foreign adversaries for the provision of artificial intelligence technology (OR NO IMPACT See Note)** — Proposed. Tracked - [US State] **Senate Study Committee on the Impact of Artificial Intelligence; create** — Passed. Tracked - [US State] **Providing for disclosures and safeguards relating to the use of artificial intelligence; and imposing duties on the Attorney General.** — Floor Action. Tracked - [US State] **Artificial intelligence; content verification** — Floor Action. Tracked - [US State] **Artificial Intelligence; state agencies required to perform quarterly AI-assisted review of rules** — Proposed. Tracked - [US State] **Artificial Intelligence; state agencies required to perform quarterly AI-assisted review of rules** — Proposed. Tracked - [US State] **Social media platforms: artificial intelligence models.** — Proposed. Tracked - [US State] **Requires the owner, licensee or operator of a generative artificial intelligence system to conspicuously display a notice on the system's user interface that is reasonably calculated to consistently apprise the user that the outputs of the generative artificial intelligence system may be inaccurate.** — Floor Action. Tracked - [US State] **An act relating to liability standards for developers and deployers of artificial intelligence systems** — Proposed. Tracked - [US] **AI for Main Street Act** — Floor Action. Mr. Williams (TX) moved to suspend the rules and pass the bill, as amended. - [US] **AI–WISE Act** — Floor Action. Motion to reconsider laid on the table Agreed to without objection. --- Source: Horizon (https://horizon.alchemylab.sh) — aggregated, LLM-scored AI intelligence; each item also lists its own primary source. Cite both — a ready-to-paste citation is in provider.citation.