Hey there,

AI is now reshaping email, research, and even court decisions—Phi-4 compresses serious reasoning into a small model, NotebookLM turns source-heavy work into clear videos, and judges are deciding exactly what AI can and cannot be.

Pick one of these tools to test this week and track what actually changes.

shortwave

Visit website here »

Shortwave is an AI-native inbox that turns email into an action list, summarizing threads, drafting replies in your voice, and keeping you at inbox zero without the grind.

Best for

• Operators and founders who live in email and need to clear hundreds of messages without losing key deals or decisions.

• Teams that rely on shared threads, read receipts, and fast internal comments to move work forward.

How to use it

• Use AI bundles and filters so promos, newsletters, and noise auto‑group, leaving only real work in your main view.

• Turn important emails into todos you can assign, snooze, or clear with one-click AI-drafted replies in your voice.

• Connect Tasklet so specific emails automatically trigger replies, team comments, or updates in tools like Slack and your CRM.

When not to use it
Skip it if your company blocks third‑party email tools, or if you only handle a few emails a day and don’t need advanced triage or team features.

Pro tip
Start by automating one painful pattern, like weekly status chasers or invoice replies—then expand rules as soon as you’re comfortable hitting “send without editing.

🏭 Microsoft’s Phi-4 Vision — Today’s Efficient AI Frontier

Microsoft's Phi-4-reasoning-vision-15B delivers elite multimodal reasoning on a fraction of rival compute, mastering math, charts, UIs, and high-res screenshots while knowing when to skip wasteful thinking. This 15B-parameter open-weight model, trained on just 200B tokens, directly challenges trillion-token giants for edge devices, agents, and real-world enterprise deployments.

Read the full article here »

Key Takeaways:

• 🔍 Multimodal Mastery: Solves math, reads charts/docs, navigates UIs, and handles everyday visual tasks.

• 🧠 Selective Reasoning: Thinks step-by-step only when needed, staying fast on simple perception tasks.

• ⚡ Efficient by Design: Matches larger rivals using far less data and compute, optimized for speed and cost.

• 🚀 Built for Deployment: Open-weights on Foundry, Hugging Face, and GitHub for edge, agents, and robots.

🚀 NotebookLM: Research Notes to AI-Directed Documentaries

Civil leaders argue for human-in-the-loop delivery. The goal is fewer defects, safer sites, and faster approvals without handing final say to software. The lever is pairing model checks with licensed sign-off, turning flashy demos into measurable gains on rework, RFIs, and delay claims.

Read the full article here »

Key Takeaways:

• 🎬 Cinematic Summaries: Turns your sources into immersive, documentary-style videos.

• 🧠 Three-Model Stack: Gemini 3 directs, Nano Banana Pro illustrates, Veo 3 animates.

• 💼 Ultra-Tier Only: Limited to 18+ Google AI Ultra users at about $250/month.

• 📚 Pro-Level Use: Geared toward researchers, educators, and content teams at scale.

🔍 Judges Say: AI Isn’t an Author, Lawyer, or Legal Rights Holder

Two major U.S. rulings draw a clear line between people and AI, finding that image generators and chatbots can’t be treated as authors, attorneys, or rights‑holders. From rejecting copyright for AI‑made art to denying attorney‑client privilege for Claude chats, courts are already rewriting the rules for AI.

Read the full article here »

Key Takeaways:

• ⚖️ Human-Only Copyright: AI-made works can’t be copyrighted without a human author.

• 🧑‍⚖️ No Legal Privilege: Chats with Claude don’t qualify for attorney‑client protection.

• 🧠 Intelligence Redefined: Judges see AI as tools, not independent intelligent actors.

• 🏛️ Courts Set Boundaries: Courts are now drawing the first firm legal lines around AI.

Rapid Fire Resources

📊 Take This Edition’s Poll:

Until next time,