128: Today in AI: AGI by 2027 Predicted, Claude Code Goes Mainstream, Self-Improving AI Loops Arrive

Free AI App Building Guide: https://aiwithkyle.com/ai-app-opportunity

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 22 January 2026

Summary: At the World Economic Forum in Davos, we got a proper sit-down between two of the most important people in AI right now: Demis Hassabis from Google DeepMind and Dario Amodei from Anthropic. These are the guys building what are arguably the best models in the world, so when they talk about AGI timelines, it's worth paying attention. Dario's going aggressive, predicting we'll hit Nobel laureate-level AI by 2026 or 2027 — which, if you've noticed, means potentially this year. His reasoning centres on self-improving loops, where AI can now code and improve AI, creating this recursive cycle. Demis is more cautious but still gave it a 50/50 chance by the end of the decade, noting we're still missing some capabilities around scientific creativity and physical experimentation. Either way, both reckon it's coming sooner than most people think. Meanwhile, Claude Code is properly breaking into the mainstream. The Wall Street Journal ran a piece on what they're calling "getting Claude-pilled" — that moment when people first use it and realize just how capable it actually is. Even Jensen Huang at Nvidia picked it up at Davos, saying Anthropic made a huge leap in coding and reasoning, and that every software company needs to use it. I've only had that watershed feeling a few times in AI — when ChatGPT first dropped, when Gemini's NotebookLM nailed images with text, and now with Claude Code. It's genuinely that significant, and it's brilliant seeing it get the recognition it deserves.

Watch full edited live stream: https://youtu.be/hYluEimWNMc

—— Time Stamps ——

0:00 Introduction: Demis Hassabis & Dario Amodei on AGI 2:32 AGI Timeline Predictions & Definitions Explained 3:00 Dario's Aggressive Timeline: AGI by 2026-2027 3:49 Demis's Cautious View: 50/50 Chance by 2030 4:48 Self-Improving AI Loops & Physical World Limitations 5:36 AlphaGo & The Thinking Game Documentary Recommendation 6:31 Labor Market Impact: Entry-Level Jobs Under Threat 8:24 How to Prepare for AI-Driven Job Disruption 9:28 Geopolitical Strategy: Should We Sell Chips to China? 10:56 Chip Wars: The Technology Race Explained 11:29 International Cooperation vs US-China Competition 12:42 Existential Risk: Will AGI Wipe Us Out? 14:34 Commercial Pressure vs Ethical AI Development 15:48 AI Safety Movement & The Slowdown Debate 17:27 Prisoner's Dilemma: Can AI Labs Coordinate? 19:03 Trump's Take: Energy is the New AI Battleground 20:41 China's Massive Energy Production Advantage 21:45 Claude's New Constitution Released 23:28 Google's Revolutionary Educational AI Tools 24:35 Free SAT Practice Exams Inside Gemini 25:39 Khan Academy Partnership: AI Writing Coach 27:40 Companion Community Reacts to Claude Changes 28:41 Claude Code Breakthrough: Viral AI Coding Agent 29:17 Wall Street Journal: Getting "Claude-Pilled" 30:33 Jensen Huang Praises Claude Code at Davos 31:12 How to Install & Access Claude Code 33:00 Using AI to Solve Technical Problems 33:31 Claude Code Fixes Notion Bug in One Shot 34:30 Claude Code vs GitHub Copilot Comparison 34:46 Outro: Join Newsletter & Community

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/