Foundation · Boot Camp
What is AI, how it learns, where it lives. Short videos — watch any, in any order.
Two words: Artificial. Intelligence. Break them apart and the whole idea clicks.
Up next
Talks, sees, draws, drives, helps doctors. The wild stuff AI is already doing today.
Some gadgets look smart but aren't. The test: does it learn from experience?
From Turing's wild question to ChatGPT — the whole story in one scrubbable timeline.
Supervised, unsupervised, reinforcement — three recipes for the same big idea.
Every program is a box with knobs. Until now, WE set them. What if the computer could?
The 5-minute rule: if you can write the rule in 5 minutes, don't use AI.
Every AI has one teacher: data. If the teacher is biased, the AI is too. Garbage in, garbage out.
Guess. Check. Adjust. Repeat. That's it — that's learning.
Watch a computer adjust its knobs until its "wrongness" number shrinks.
Learning answers vs learning the pattern. Hide some examples (train/test split) and see if the AI truly gets it.
The box with knobs, revealed: inputs, weights, a squish — that's a neuron.
Stack neurons. Feed outputs into inputs. Now you have a brain of sorts.
The weights ARE the AI. Training fills them in (weeks). Inference runs new input through them (milliseconds).
CNN sees. RNN reads in order. Transformer understands context. Three shapes, three superpowers.
AI is never "done". Collect → Train → Use → See new data → Improve → Repeat.
AI always guesses. It shows confidence, not truth. And sometimes confidently wrong — hallucinations and all.
Bias starts in the data. Fairness is everyone's job — including yours.
Your data is yours. AI hallucinates. Voices and photos can be faked. 5 rules to stay safe.
Hands, eyes, text glitches, deepfake tells. Your digital-forensics field guide + the SIFT method.
Kids, adults, animals, birds, vehicles, props, sea creatures — all generated from 7 rigs. The agentic character pipeline at full breadth.