GPT-6 is now running your regular ChatGPT 🤖

Good morning. OpenAI rolled GPT-6 into regular ChatGPT, and answers can now show up as interactive charts and tools instead of walls of text. Tencent dropped an image model that lets you edit a picture through a back-and-forth conversation. And absurd AI parkour videos, the last person you’d expect pulling off flawless flips, are blowing up. Would you rather chat with an AI or have it build you a tool on the spot? Hit reply and let me know! Let’s get into it.

OpenAI started rolling GPT-6 into the main ChatGPT chat tab, so the model behind your everyday conversations just got an upgrade. Paid plans are running GPT-6 Sol, free and Go users get GPT-6 Luna, and it’s rolling out to everyone over the next day or so. If you use ChatGPT at all, this is already changing under the chat box.

You might notice the new Intelligent UI. Instead of answering everything with a wall of text, ChatGPT can now build the answer as interactive stuff, a chart you can change, a little calculator, buttons, a form, a map, whenever it decides that fits your question better. It’s ChatGPT shifting from “thing that types paragraphs” to “thing that hands you a mini tool,” and you can dial the visuals down if it’s too much.

For creators, the upgrade is quietly useful. The model you lean on for scripts, hooks, titles, and research just got smarter, and the interactive answers are handy when you’re comparing tools, planning a content calendar, or sketching out numbers. No set up needed, it’s just there, and it’s worth playing with the new format to see where it actually saves you time.

Tencent released HY Image 3.5 Preview, a new AI image model, and the standout feature is how you edit. Generating an image and editing one now run through the same model, and you can refine it through a back-and-forth, “make the jacket red,” then “now add a sunset,” then “zoom out”, and it remembers each previous step instead of starting from scratch every time.

The conversational editing fixes the most annoying part of AI images. Normally every tweak risks blowing up the parts you already liked, so you re-roll endlessly. A model that holds the context of your edits means you can actually sculpt one image toward exactly what you want, the way you’d direct a designer. It also handles text in images and 4K output, though the heaviest 4K is upscaled rather than made natively.

It’s currently more of a power user tool right now. The main way in is through ComfyUI, a node-based editor that’s a step up in complexity from the apps most people start on, and generations are paid, not free. So it’s less “open it tonight” and more “worth knowing where image editing is heading,” toward tools you steer with a conversation instead of a fresh prompt every time.

The latest AI trend making people stop scrolling is unlikely parkour. Picture the last person you’d ever expect to be an athlete, a heavyset guy with long red hair, neon shorts, and yellow Crocs, pulling off flawless professional parkour, diving over ledges and nailing rolls through an industrial lot. It’s all AI, and one clip is sitting at over 326,000 likes.

Your brain expects one thing and gets the complete opposite, and that mismatch is a big reason it goes viral. The comments do half the work too, people quoting physics or tagging a friend, which is the kind of engagement that shoves a clip further up the feed. It’s dumb in the best way, and dumb-in-the-best-way travels.

Here’s the playbook. Take someone or something nobody associates with athleticism, then have them pull off genuinely impressive, physics-defying moves with a totally serious tone, the more casual the delivery, the funnier it lands. Generate it with a strong image-to-video tool that can handle fast motion, keep one recurring character so people come back for them specifically, and lean into a caption that matches the account. The gap between who they are and what they’re doing is the formula.

Stuck on a tool or have an AI video question? Reply, I’ll do my best to help.

Created something cool? Send it over and I might feature you. 📺

📫 Want to get your tool featured or advertise to our readers? Let's chat → [email protected]

Thanks for tuning in!

-Kevin