Boris Cherny, the dude who invented Claude Code, says it's all about the loops now. Some people are having a difficult time with this reality because they built their kingdoms on that principle. Never a good idea with tech that moves this fast. If you're going to build your future in this space you better be prepared to pivot from day to day. Being nostalgic will make you poor.

Don't be afraid to give away your stuff. Give them your stuff. Sell them the implementation. I know how to change the oil in my car but I still have someone else do it.

Google dropped an entire ecosystem of new releases yesterday centered on one strategy: turn Gemini into the agent layer running everything you already use. New tools and models released: Gemini 3.5 Flash: Google's new speed model. Benchmarks near Claude Opus 4.7 and GPT-5.5 at 4x the speed and half the cost. This is the one to watch for builders running AI at scale. Gemini Spark: A 24/7 personal agent that runs on Google Cloud virtual machines. It takes autonomous actions across Gmail, Chrome, Workspace, and chat without you having to prompt it. This is Google's answer to the agentic AI wave. Gemini Omni Flash: Converts text, images, audio, or video inputs into video outputs. Google is calling it their multimodal video model that edits via chat. Early days but the direction is clear. Google Intelligent Eyewear (announced, shipping fall 2026): Gemini-powered smart glasses in partnership with Warby Parker and Gentle Monster. Voice first, camera, translation, navigation. Google's direct shot at Meta Ray-Bans. Google Search overhaul: Framed as its biggest redesign in a generation. Cross-modal inputs, 24/7 info agents, generative UI. The traditional search experience is being replaced with an agent doing the searching for you.

Anthropic just removed the last excuse non-technical entrepreneurs had for not building AI agents. Most of you scrolled past the Claude Managed Agents announcement. That was a mistake. Let me break down what this actually means for you. Before this, building an AI agent that could do real business work: reading emails, updating your CRM, generating reports, following up with leads... required servers. Security infrastructure. Engineering teams. Months of setup before the thing does a single useful task. That was the wall that kept most entrepreneurs out. Anthropic demolished it. With Managed Agents, you describe the job. They run the infrastructure. Your agent gets a secure workspace, access to tools, and the ability to actually execute work. No server setup. No dev team. No months of pre-work. The barrier just dropped from "hire engineers" to "describe the problem clearly. Here is what nobody is saying loud enough: The businesses that need this most, law firms, real estate agencies, accounting practices, medical offices, will never set this up themselves. They do not want to learn it. They do not have time to learn it. They want someone to do it for them. That someone is you. The playbook is simple: Identify their biggest time waster. Build an agent that solves it. Deploy it on Anthropic's infrastructure. Charge a monthly retainer to maintain and scale it. Four clients at $500 a month is $2,000 in recurring revenue. You are not trading time for money. You are building systems that run while you sleep. The people who win in the AI economy are not the ones who waited until they felt ready. They are the ones who understood the opportunity while everyone else was still confused by the terminology. You do not need to write code. You need to understand business problems and know how to describe them to an agent that can solve them.