Have you heard these exciting AI news? - March 06, 2026 AI Updates Weekly

Have you heard these exciting AI news? - March 06, 2026 AI Updates Weekly Presented by Lev Selector Slides on GDrive: https://drive.google.com/drive/folders/1dmj00iilSP-JYmriMUUaHCKz5X54e3UX?ths=true Slides on GitHub - https://github.com/lselector/seminar (click on pptx file, then on "raw" or download button on the right) ------------------------------------------------------- Question to you: Which "Claw" or other desktop "Digital Employee" you recommend? ------------------------------------------------------- Need AI Consulting ? - Enterprise AI Solutions - https://EAIS.ai - Linkedin - https://www.linkedin.com/in/levselector - GitHub - https://github.com/lselector ------------------------------------------------------- Our books: "Artificial Intelligence In Business: Strategies That Work" https://www.amazon.com/dp/B0FKZ3364G/ "AI Product Manager: Building Tomorrow" https://www.amazon.com/dp/B0FMS9NFK9/ "How to Learn with AI?: A Practical Guide for Students, Professionals, and Self-Learners" https://www.amazon.com/dp/B0FWLNVQMJ/ "Professional + AI: Future-Proofing Your Career" https://www.amazon.com/dp/B0GBXLCTZQ --------- Contents of today's video: 00:00 Introduction 00:07 AI is a 'wrecking ball' destroying SaaS 00:28 "Skills instead of SaaS" 00:45 "Your Job is to Onboard AI Agents" - Peter Yang 01:07 Crowd-sourced "LM Arena" Leaderboard 02:51 Mercury 2 by Inception Labs 03:55 OpenAI GPT-5.4 & GPT-5.3 Instant 04:58 Google Gemini 3.1 Flash-Lite & Pro 05:24 Gemini Nano Banana 2 05:52 Qwen3-Coder-Next 06:13 Claude Co-Work Can Schedule Tasks 06:41 Anthropic Skill Creator tool (skill) 07:42 Claude Code MEMORY.md 08:19 Claude Code voice mode feature 09:01 Alibaba's CoPaw, ModelScope 10:20 Claude Code vs OpenClaw 10:42 Claude Code supports Remote Control 11:08 Claude now remembers your work habits 12:17 From Apps & SaaS to Generic Agent 14:46 ClawX - FREE Desktop App for OpenClaw AI Agents! 15:13 OpenClaw + Simflow AI Phone Voice Agent 15:20 Accomplish (Openwork) - AI Desktop Agent 15:42 NullClaw against OpenClaw 15:58 Harness Engineering 17:42 MaxClaw - Minimax's agent 18:16 Qwen3.5-9B - text, images, short videos 19:47 OpenClaw Tips 21:14 Tools That Make OpenClaw Actually Useful 21:43 Toggle - browser agent skill for OpenClaw 22:00 SHARE + Sidecar - Model That Learns 22:32 Kane AI (KaneAI) software testing assistant 22:40 Claude Code + Telegram bot 22:58 Apple Updates (M5, Neo) 24:02 AI Agent Making Money and Doing Marketing 24:42 The 2028 Global Intelligence Crisis report 25:38 AT&T cuts costs 90% by using multi-agent architecture 26:10 Perplexity Computer with Max Plan $200/mo 26:50 "Ilya Didn't Leave OpenAI for Nothing" 28:05 LFM2-24B-A2B Open Model from Liquid AI 28:33 CodeBuff, a new open-source AI coding agent 28:52 Humanize Text, Remove AI Accent 30:18 Antropic vs US Dept of Defense 33:42 Moonshots 34:00 Interview with Andrew Ng 34:37 SaaS Collapse into Personal AI Agents 36:40 Jobs, Layoffs, demand for AI Engineers ---- Takeaways: 1. The software industry is undergoing a massive shift from specialized SaaS applications to generic AI agent platforms with skills. Instead of buying separate apps, users will interact with one central agent and add skills or plugins for specific tasks, making traditional SaaS obsolete. 2. Harness engineering matters more than the model itself. Top frontier models only complete 24 percent of real-world tasks despite scoring over 90 percent on benchmarks. The infrastructure wrapped around models, including tool management, error recovery, and context tracking, is what drives real results. 3. Anthropic refused Pentagon contract terms requiring mass surveillance and autonomous weapons, sparking a major controversy. OpenAI signed but faced backlash and renegotiated. Meanwhile, Anthropic abandoned its safety-first pledge, stating competitive pressure destroys voluntary restraint. ---- Quotes: "AI is not a bubble but a wrecking ball destroying traditional software as a service business model." "The pledge does not work if competitors are racing ahead. Competitive pressure destroys voluntary restraint. The company that waits is the company that loses." "Top frontier models only complete 24 percent of real world professional tasks despite scoring over 90 percent on benchmarks." ---- Summary: Key themes include the rapid shift from traditional SaaS to generic AI agent platforms powered by skills and memory, with dozens of new agent tools like OpenClaw, ClawX, and MaxClaw emerging. Major model releases include GPT-5.4, Gemini 3.1, Mercury 2, and Qwen 3.5. The talk highlights harness engineering as the critical factor for real-world AI performance. Important - Anthropic versus Pentagon controversy over mass surveillance and autonomous weapons, and Anthropic abandoning its safety-first pledge under competitive pressure. Andrew Ng shares his view that AGI is likely decades away Vertical SaaS will collapse into personalized AI agents. ----