Quick Scan
What matters today
The AI story at CES 2025 was not a new model or a benchmark. It was a shift in what AI is asked to do.
Format Weekly newsletter issue
Audience Executives using AI at work
Time 11 min read
Topic Beyond Chatbots: AI Agents Now Execute Work for You
Key points
- Microsoft Copilot Gets Pattern Recognition in Excel (Jan 4): Copilot in Excel now auto-detects data patterns and suggests complex formulas, cutting time spent on manual spreadsheet formatting.
- Meta AI Now Delivers Real-Time News Inside WhatsApp (Jan 5): Meta AI on WhatsApp now pulls live news from Reuters and CNN partnerships, delivering current event summaries without leaving the app.
- NVIDIA Releases Private Local AI Model for Windows PCs (Jan 6): Llama Nemotron-3 8B runs locally on NVIDIA RTX GPUs, providing full AI inference without any data leaving the device.
- Perplexity Deprecates Legacy Sonar Models (Jan 1): Perplexity officially retired its Llama-3.1-sonar model variants, migrating all users to the new Sonar and Sonar Pro endpoints.
- NVIDIA Open-Sources Cosmos World Models for Physical AI (Jan 6): NVIDIA released Cosmos, a family of foundation models designed to help AI systems understand physical laws and simulate real-world environments.
Welcome
This issue gives executives a clean scan of Beyond Chatbots: AI Agents Now Execute Work for You. Start with the quick hits, then use the companion guides for the workflows worth testing this week.
Quick Hits
Top AI Updates
1. NVIDIA Just Put an Autonomous AI Agent on Your Desktop
How the NVIDIA RTX 50-series agentic layer automates complex administrative tasks without sending data to the cloud.
2. Microsoft Copilot Now Reads Only the Paragraph You Ask About
The technical mechanics of the Scoped Chat update and how it eliminates context drift in complex documents.
3. Mastering Chain-of-Thought Prompting
Most users treat AI like a search engine, expecting an instant answer. But for complex logic, math, or strategy, you need to change your approach. Chain-of-Thought (CoT) prompting is the single most effective way to reduce hallucinations and improve output quality.
Pro Tip
Modern LLMs now support context windows of 128k tokens or more. But just because you can dump an entire book into a prompt does not mean you should. Without structure, the model suffers from the "lost in the middle" phenomenon.
Productivity Gem
How to activate the Scoped Chat feature in Microsoft Word to analyze specific document sections.
Health Tip
How to extract raw sleep metrics from Oura, Apple Health, and Garmin for external analysis.
This information is for educational purposes only and is not medical advice. Please consult a qualified healthcare professional before making changes to your health routine.
Kids Tip
Kids create a hands-on AI project: The "What's Wrong With This Answer?" Game Teaches Kids to Think Critically About AI. A parent or educator helps them build, test, and explain what the AI tool gets right and wrong.
Wrap Up
Free subscriber access
Keep reading with your free PromptHacker subscription.
Enter the email on your PromptHacker subscription. If access is available, a secure link will arrive in that inbox.
New to PromptHacker? Join the free weekly briefing first.
PB
About the author
Pierre Bradshaw Founder, PromptHacker.ai
Pierre has spent 25+ years building growth systems across fintech, real estate, lending, campaigns, and AI workflows, with $1.5B+ in client value delivered.
Email us