ChatGPT Images 2.0 Business Guide (Free): https://aiwithkyle.com/mini/chatgpt-image-2

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

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

Summary: GPT-5.5 dropped yesterday, which means I spent this morning doing what I always do on a Friday: working out what it actually means beyond the benchmark chest-beating. The short version is that the model itself matters less than you think. What OpenAI has done with 5.5 is build something that works exceptionally well inside a harness, and if you don't know what a harness is yet, that's exactly what this video covers. The concept is simple once you see it, and it changes how you think about every AI release from here on.

We also get into Codex, why it's the most accessible entry point for anyone not spending serious money on Claude Code, and the quiet but very deliberate war OpenAI is picking with Anthropic over who actually gets to use these tools. There's also the small matter of Claude Code spending the entire month of March running at reduced intelligence, which Anthropic has now admitted was intentional. A lot happened this week.

—— Time Stamps ——

0:00 Intro: GPT 5.5 Just Dropped
0:34 Why Non-Coders Should Care About Coding Models
2:23 GPT 5.5: Intelligence For Real Work
4:32 3 Ways to Access GPT 5.5 (ChatGPT, Codex & API)
7:54 How I Run Agents Overnight (64-Page PDF Example)
9:32 Why You Should Start With Codex (It's Free)
10:43 How AI Agents Actually Work
13:50 Why You Should Ignore Every AI Benchmark
15:28 Testing GPT 5.5 Live: Building a Slide Deck
17:54 Sam Altman's Tweet: An Attack on Anthropic?
23:14 The Harness Matters More Than the Model
26:52 Claude Code Got Dumber in March (Confirmed)
29:13 OpenAI's Push Into OpenClaw
32:08 Choosing Your Harness: A Practical Guide
34:24 Outro & Newsletter

— Useful Resources ——

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