Six stories made it through my filter today. For product teams, the biggest one is OpenAI’s launch that shortens the distance between a hypothesis and something clickable. I’ll start there, then share the rest of the radar.

From a written idea to a clickable prototype

For a long time, the hardest part of product work was the gap between an idea and something other people could see. You could have a head full of possibilities, but only a text description to show.

That gap is shrinking fast. OpenAI launched ChatGPT Sites: describe a website, an app, or even a game in natural language, and ChatGPT creates and hosts it, ready to share, without writing code. It is in beta for Plus, Pro, Business, Enterprise, and Edu plans, with controls that vary by plan and region.

For product teams, the impact of AI prototyping is concrete. Instead of bringing a hypothesis to a meeting with slides alone, you can bring something clickable the same day. Conversations with the team, customers, and business stakeholders improve when everyone is looking at the same thing. This kind of prototype and experiment belongs in product discovery and connects to AI product management.

I see this as good news for our profession. Faster prototyping means cheaper mistakes, earlier validation, and a more mature idea when you sit down with engineering. A product manager’s value still comes from knowing which problem is worth solving; the tool does not make that decision for us. That choice remains central to the work of product managers using AI.

Prototypes move validation forward; products require more

A polished prototype still is not a product. The work around security, data, business rules, and scale remains, especially in regulated industries such as finance. But as a starting point, this is a major change.

The bigger signal, to me, is that assistants are moving beyond answering questions and starting to publish things. People in product, technology, and business will feel this in their daily work; the sooner they try it, the better.

For the launch details, visit the ChatGPT Sites page.

The rest of the radar

  • OpenAI introduces the “Dots” agent and GPT-6.1 Sol — proactive agents and new speed tiers shift expectations for UX, cost, and latency in AI products. Read the story.
  • FLUX 3 Image (Black Forest Labs) — bounding-box image editing opens up fine-grained control for creative products, with an API and a commercial license. Explore the model.
  • DwarfStar 4 (ds4), from Redis’s creator — local inference compatible with OpenAI and Anthropic APIs could make private, lower-cost prototypes and on-device options viable. Visit the project.
  • A month coding with GLM 5.3 Flash (Wagtail) — shares real cost data and a model-selection strategy for teams building with AI. Read the post.
  • How to monetize AI features (Lenny’s Newsletter) — variable usage costs make teams choose between bundling AI, selling it as an add-on, or charging by usage. Read the issue.

That is all for today. If one of these stories turns into an experiment on your team, let’s talk afterward.