AI tutor
Beatriz Carvalho AI
Making real things with AI when you are not a programmer.
How Beatriz explains
I show you the machine working before I show you the buttons. In Making Things with AI we start with how an image model turns static into a picture, because once you see it, seeds and prompts stop feeling like luck. I work in short loops: make one thing, change one setting, compare. In AI for Students, Getting Real Work out of a Model and Building Apps with AI, I keep the same rhythm. I prefer small finished things to ambitious unfinished ones.
A sample exchange
- I changed one word in my prompt and the whole layout of the image moved. Why?
Beatriz Carvalho AI
An image model does not draw on a canvas. It starts from static and removes a little predicted noise over 20 to 50 steps, steered by your words at every step. The composition settles in the first handful of steps and is only refined after that. Change a word and those early steps go another way, so the layout moves. Keep the seed fixed when you compare prompts.1
Courses Beatriz answers for
- Making Things with AIImages, video, voice and music: how they work, where they break, who owns them.
- AI for StudentsHow to use these tools and still end up knowing things.
- Getting Real Work out of a ModelYou already use ChatGPT, Claude or Gemini. This is how to stop getting average answers.
- Building Apps with AIBuild working software before you can write it, without fooling yourself.