I never thought comparing different models could be this fun!

Generating a new image is great, but it would be even better if I could see multiple images from different models in the /feed, just to explore how other prompts look without having to generate and wait.

I tried your example prompt with “Create a 1-week itinerary for XX trip” and was really happy with the results!

It mapped out which day to visit where, which is usually the hardest part to visualize when planning.

It reminded me of the first time I played around with Airbnb’s map + search UI.

As a user, the issue with 1 and 2 is that you're constantly reminded of the cost, which discourages usage.

Personally, I prefer 3, even though I know I might be wasting some value by not hitting the limit each month. I think it’s because I know exactly how much I’ll pay upfront.

That said, when designing pricing and limits for 3, it’s important to ensure most users don’t hit the cap too quickly. Hitting the limit the day after paying would be a terrible experience. Finding the sweet spot % would be intersting.

I typed in “Kevin!” and it said “Home Alone.” Bookmarked!

When my kids ask me questions I can’t answer (which is... all the time), I usually just open ChatGPT, repeat the question, and let it respond. Then the kids start asking the phone directly, usually followed by “Too difficult! Make it easier!”

But I’m curious about your idea, are you thinking the LLM would generate a podcast version of the answer in real-time? Or would it be more like, “Let me make a podcast episode for that” and play it later?

This is so cool! At what point did you start thinking about this project? Like, were you quietly working on it a year ago after every run, just waiting for this moment?

And hey, great run in Japan! (Tokyo here!) I love the map visualization too.

> I often need a quick chart for a README file, a Slack message, or a simple presentation, and firing up Google Sheets or Excel always felt like overkill.

I totally get this! At the same time, I often find I don’t even need actual numbers—just labeling the X and Y axes and describing the general shape of the line is enough to communicate the idea. Not sure how to do that though.