AI for
Everyone

By combining five Claude Co-work capabilities, you can build an automated, high-level marketing and operations system. Here's what we covered:

• Trend Intelligence with X & Grok — Use X and Grok to gather real-time trend data scored by custom metrics like recency and velocity, so you're always ahead of the curve.
• Competitive Intelligence via Apify — Scrape competitor social media performance and market pricing using Apify's custom connectors — no expensive API subscriptions needed.
• NotebookLM as Your AI's Second Brain — Install the NotebookLM skill to give Claude near-infinite memory and deep research capability without burning through tokens.
• Building the NotebookLM Skill with Claude Code — Because Claude Co-work runs in a sandboxed environment, you use Claude Code to extract, optimize, and embed your browser cookies into a specialized file Co-work can read.
• Mission Control Dashboard — Aggregate all your intelligence and business numbers — website traffic, revenue attribution, and more — into a centralized, AI-powered dashboard that updates daily to reduce cognitive load and sharpen decision-making.

Taking a week off after a packed session on April 1. Classes are resuming April 15 with a new, more structured format — specific topics announced each week. Stay tuned and register below to get notified.

Exciting session with business owners in the room. We covered how to use AI to create a full website, manage daily tasks and projects, and build a brand — logo, tone, messaging, and marketing copy — with AI doing the heavy lifting. Great energy, real conversations about putting this to work immediately.

This week's class covered how to turn AI skills into real income — freelance coding, prompt engineering, and automation services that businesses will pay for. Meanwhile in AI news: OpenAI quietly discontinues the Sora public API after 30 days notice, citing unsustainable video generation costs at scale.

Anthropic accidentally exposed its most capable model yet — insiders call it "a step change" in reasoning; an official release is weeks away. Google ships Gemini 3.1 Deep Think to Ultra subscribers and opens early API access. Google also drops Lyria 3 and Lyria 3 Pro — AI music generation models. And the Model Context Protocol crosses 97 million installs, cementing it as the backbone of agentic AI infrastructure. OpenAI adds Walmart shopping directly inside ChatGPT.

Really interesting session this week. We covered how anyone can move beyond basic AI prompting into cinematic prompting — crafting detailed, scene-level instructions that produce stunning images and video — and schematic prompting, where you describe what you want to build and AI generates working applications and code. You don't need to know how to code first. It helps, but you can absolutely learn as you go.

Google's Veo 3.1 launches in public preview — generate 4–8 second clips from text or up to three images. OpenAI drops Sora directly into ChatGPT for millions of users. Open-source strikes back: LTX 2.3 (22B params) generates 4K video at 50 FPS with synchronized audio in a single pass — free for commercial use.

Anthropic rolls out persistent memory to all Claude users — even the free tier. Claude now remembers your name, writing style, and ongoing projects across sessions. Also new: Computer Use on macOS lets Claude control apps, browsers, and files directly on your behalf. Off-peak usage limits double through March 27.

OpenAI releases GPT-5.4 — its most capable model yet — in three variants (Standard, Thinking, Pro) with a 1.05M token context window. Perplexity launches "Computer," an AI agent that autonomously executes complex workflows using 19 different models and can spawn subagents. Google Gemini adds "Fill with Gemini" to Docs and Sheets, generating table content, categorizing data, and pulling live info from Search.

The AI research community heavily pivots toward "model merging" — combining the capabilities of different specialized models without massive retraining — as the next major strategy to bypass the ceiling of traditional data scaling.

Anthropic's Claude Code gains the ability to translate legacy COBOL — powering 95% of global ATM transactions — into modern code at 99.9% accuracy. IBM stock drops 13% in one day. Meanwhile, legal battles intensify over always-listening wearables and real-time facial recognition.

Google DeepMind releases Gemini 3.1 Pro, Anthropic drops Claude Sonnet 4.6, and xAI releases Grok 4.20. Researchers at the University of New Hampshire use AI to discover 25 new high-temperature magnetic materials for rare-earth-free electric vehicles.

AI models identify signals for next-generation solid-state battery ion conduction. A University of Michigan team uses AI to compress cancer tumor stress-testing from 22 days down to under 5 minutes.

Anthropic releases Claude Opus 4.6 for complex multi-step autonomous logic. OpenAI drops GPT-5.3 Codex, achieving unprecedented scores on software engineering benchmarks and further automating high-level programming.

Moonshot AI releases Kimi K2.5 — a 1-trillion parameter open-source model that outperforms proprietary giants. Google rolls out Chrome's "Auto Browse" AI agent and Gemini 3's "Agentic Vision." Apple acquires Q.ai to improve whispered speech recognition.

OpenAI introduces Prism for automated scientific writing. NVIDIA launches Earth-2, the first open AI weather forecasting system. Microsoft debuts the Maia 200 accelerator to slash the cost of generating AI responses at scale.

Anthropic's tools begin handling autonomous project management, sparking debates about software engineering's future. EPFL researchers unveil software allowing organizations to run powerful AI models locally without cloud reliance.

OpenAI launches "ChatGPT Health," allowing users to securely connect medical records for personalized support. Amazon debuts a web-based AI assistant for Alexa, moving beyond voice-only smart speakers.

Enterprise leaders finalize 2026 budgets, marking an industry-wide shift from experimental GenAI chat tools to autonomous multi-agent systems designed to execute complex, multi-step workflows without human supervision.