Shawn’s AI new marketing breakdown so you don’t have to frantically follow everything in the AI news space. Why this newsletter exists As a CMO who runs more dashboards than ad sets, I’ve learned that the real moat isn’t “ideas”—it’s systems. The tools below are the gears I’m bolting onto that moat so our brand (and yours) can move faster than the algorithm shifts. 1. Kimi-Researcher schools Gemini — and hints at self-optimising funnels Moonshot AI’s new agent hammered the notoriously brutal Humanity’s Last Exam with a 26.9 % Pass@1—triple its previous score and enough to nose ahead of Google’s Gemini. The kicker: it needed 200+ reasoning steps and scraped 23+ web pages per task to get there. That is, it planned, researched and iterated before writing a single word. Marketing take: If an LLM can chain its own thinking, it can A/B test copy, bids and audiences without human babysitting. Expect the first “agent-native” campaign optimisers (think: Adobe Sensei on steroids) by Q4. Start logging every conversion touch-point now—agents will need the data. 2. DeepSeek’s export-ban dodge: a brand-safety headache in waiting U.S. officials claim DeepSeek routed American user data through China Mobile networks and quietly bought Nvidia H100s via shell firms—tying the model to PLA research. Marketing take: Your compliance team will ask whether any Chinese-trained AI you use is “clean.” If you’re pitching Fortune-500 clients, be ready with an AI supply-chain map and a fallback vendor—the same way you prep for privacy audits today. Don’t get caught rewriting privacy policies mid-campaign. 3. Anthropic’s half-victory on copyright: fair-use yes, pirated data no A federal judge said Claude’s training on legally purchased books is “spectacularly transformative,” but kept a December trial over 7 million pirated titles. Marketing take: Fair-use breathing room means you can keep using models that learn from licensed stock libraries—great for on-demand creatives. But the minute your vendor’s corpus turns out to be “too good to be true,” you inherit that legal tail risk. Ask for a data provenance letter before you sign the MSA. 4. Reid Hoffman funds an AI ultrasound helmet — neurotech’s soft sell Hoffman led $12 M into Sanmai’s in-home ultrasound headset for anxiety and depression, calling it a safer Neuralink alternative. FDA trials are underway. Marketing take: Wearables that treat rather than track health issues reshape DTC funnels: think medical-grade influencer testimonials, HIPAA-compliant remarketing, and a legal mandate to prove outcomes. If you’re in health & wellness, start scoping creative that explains “AI-guided therapy” in 15 seconds or less. Tools I’m stress-testing this week Kimi-Researcher API (beta) - Autonomous desk-research for competitor tear-downs; feeds insights straight into your slide deck. Gumloop - Zapier-style automations with an AI brain—great for routing UGC into personalised email flows. Einstein X - Salesforce’s agentic upgrade; lets you set “pipeline velocity” as a goal and auto-tunes touch-points. Demandbase Autopilot - Spins up account-level ad creative using live intent signals; cuts hours off ABM sprints. My take Tool fatigue is real, but so is the compounding effect of picking one agentic platform early and feeding it clean, structured data. This week’s stories—whether breakthrough, scandal or lawsuit—all point to the same mandate: own your data, audit your vendors, and let the agents do the busywork. Today's Top Tools: • 🗣️ 11ai – ElevenLabs' new voice assistaKimi-Researcher Just Topped The World’s Toughest AI Examnt with MCP integrations • 💻 Mu – Microsoft’s fast local AI for Copilot + PCs • 🎆 Imagen 4 Ultra – Google's flagship image model, now API-accessible Quick News: • ChatGPT iOS downloads hit 29.6M in 28 days — nearly overtaking TikTok, Instagram, Facebook combined (32.6M) • Sam Altman called the ‘io’ lawsuit “silly and wrong,” accusing founder Jason Rugolo of pushing for acquisition • Mira Murati’s Thinking Machines Lab to develop profit-focused custom models • Google launches Gemini Robotics On-Device: fast, offline VLA model for robotics • Andy Konwinski (Databricks/Perplexity) launched Laude Institute, pledging $100M for applied CS breakthroughs • XBOW’s autonomous AI beats all humans on HackerOne; raised $75M Series B Looking for a community of like-minded individuals who are interested in AI and Entrepreneurship? Join our free community here to get started: The AI Advantage Community Thank you for reading! -Shawn
