TensorFlow.js was from before and Daniel Shiffman created ML5.js from there?
Nikita Williams: Adjunct professor, from Trinidad. They worked in ML5. Created datasets.
AI generation in real time. The model in runway will let you do certain actions generating more content in real time.
Another 3D world modeling tool is Marble.
We’ll talk about about diffusion models and so many others.
There’s also realt-time Avatars. She’s working in that. Is based in Chat GPT’s talking in real time.
3 concepts:
This is the whole big field. It started in 1950 and then ML started in the 1980s. It didn’t go anywhere because there was not enough data and compute. Now Machine Learning, Deep Learning and Neueral Networks are driving the thing.
General AI: fabulous machines that have all our senses or even more. Think just like we do.
Narrow AI: They can only do one or a few of the senses. They do it better than human.s Machine doesn’t mean robot. Could be a software. This is the powerful one, usually we’ll be talking about that one.
Machine to perform human tasks
Gives computers the ability to learn. It is basically a program. Programming with examples not instructions. Difference:
Traditional Programming: we define the code and define the specific instructions. Computer will do it for us. Following the instructions and giving an output.
Machine Learning: In this case we give the data and the output but the computer does the program that eventually will do it. Learn patterns from what we feed it.