AI Tools

🎧 Prefer to listen? Your browser does not support the audio element. Dario Amodei, CEO of Anthropic, testified to Congress in May 2025 that Chinese labs are running “industrial-scale distillation” on American AI models right now. He claims this process—using a top-tier model’s outputs to train a smaller, cheaper one—is being used by military-linked researchers to build defense systems. His push for controls on computing power, not just model usage, marks a new kind of choke point for the AI industry. ...

🎧 Prefer to listen? Your browser does not support the audio element. I build AI products for a living, and when OpenAI announced ChatGPT for Teens, my first thought wasn’t “good for them.” It was “how much did this cost them in legal fees before they got here?” They didn’t add teen safety features out of the goodness of their hearts. They did it after Florida sued them over chatbot interactions linked to teen mental health crises. That’s a liability gap, and OpenAI’s IPO filing makes it even more awkward. ...

OpenAI’s Codex Micro is just the beginning — Apple’s lawsuit reveals how seriously OpenAI is betting on physical AI hardware.

🎧 Prefer to listen? Your browser does not support the audio element. Microsoft just announced its AI and Agent Platform as a full enterprise stack — build, ground, govern, and operate agents at scale with security, compliance, and Responsible AI tooling. That’s great if you have a team of 50 engineers and a seven-figure budget. But what if you’re one person trying to build something that actually works? ...

🎧 Prefer to listen? Your browser does not support the audio element. Two weeks ago, I wrote about OpenAI’s Codex Micro and the pattern every major technology follows: a breakthrough becomes usable when someone builds a physical interface for it. The mouse made computers accessible. The touchscreen made smartphones intuitive. The Codex Micro is OpenAI doing the same thing for AI coding. ...

Anthropic shipped four models in two months, and the naming has scrambled half the developers I talk to. Claude Opus 5 — launched July 24, 2026 — is the one worth your attention. It matches frontier performance at half the frontier price, and introduces something no previous Claude model had: an effort dial that controls how much thinking you pay for per request. I’ve spent the last week running Opus 5 through actual work. Coding tasks, research synthesis, long-document analysis. If you read my breakdown of what Fable 5 and Mythos mean for AI users, you know Anthropic’s lineup has been a moving target. Fable 5 got banned, came back, and sits at $10 per million input tokens. Sonnet 5 is cheaper but weaker. Opus 5 lands in the middle — and on paper, it’s the smartest buy in the entire Claude family. ...

OpenAI is launching a physical macro pad for Codex, and it follows the same pattern that made calculators, mice, and touchscreens mainstream.

🎧 Prefer to listen? Your browser does not support the audio element. I’ve been testing AI image and video tools for over a year now, and the hardest part has never been the generation itself. It’s picking the right tool. Every week there’s a new model claiming to be the best, and unless you’re deep in the weeds, you’re basically guessing. Runway just launched something that could change that — a model router that automatically picks the best creative AI model for your specific request. ...

🎧 Prefer to listen? Your browser does not support the audio element. Glow raised $180M at a $1.2B valuation from Sequoia, Cyberstarts, Greenoaks, and Redpoint to build AI-powered endpoint security that prevents threats before they execute. The company has already stopped malicious npm packages from being installed in customer environments and detected AI agents attempting to pull in compromised software components — attacks happening in production right now. ...

Update August 2025: Google’s Gemini has surged past 1 billion users, making it the company’s fastest-growing product ever, while Google has also announced it will retire the classic Google Assistant on phones starting September 4. More details below. 🎧 Prefer to listen? Your browser does not support the audio element. If you’re a solo developer shipping code to production, you probably have a nagging feeling you should be running security scans. You also know that the tools capable of finding real vulnerabilities — not just linting errors, but actual memory corruption and logic flaws — cost hundreds of dollars per scan. Google just made that problem a lot smaller. ...