Everyone tells you to learn Python. Nobody tells you when you are done. This is the whole of step one: which language, the six things you actually need, the seventh nobody mentions, and a finish line you can actually check yourself against.
Both are official. Anthropic ships maintained kits for Python, TypeScript, Java, Go, Ruby, C# and PHP. OpenAI ships Python and Node. Neither choice locks you out of AI work, so pick on what you already have, not on what the internet argues about.
It reads closest to plain English, and almost every AI example you will find online is written in it, so the tutorials you land on will match what you are typing.
Stay where you are. You will have something people can actually open sooner, and the AI kits are just as official on this side.
These are the same in both languages, which is why the argument above matters less than it looks. Learn these six and you are through the door.
calories = 250meals = ["eggs", "rice", "steak"]if calories > 2000:for meal in meals:def total(meals):One line to install it, one line to bring it in. You are never writing the model. You are borrowing it, and the whole job of an AI engineer is knowing what to borrow and how to wire it together. If you can do this, you can call a model from your own file today.
pip install anthropic
>>> import anthropic
That is it. Not a certificate, not finishing a 48 hour course. Here is a real one. Read the highlighted parts and say what happened.
It tells you the file, the line, and the reason. Line 6, and I treated a piece of text like it was a list. If you can do that out loud, you are done here. Move on to step two.
Skipping the right things is most of the speed.
You only need one of these. Pick the one that matches your language and start today, not Monday.
Harvard's Python course, and it covers almost exactly the six things above. Free to watch and to do. If you picked Python and you want one structured thing, make it this.
cs50.harvard.edu/python →Written by the people who make the language. Drier than a course, but it is short, correct, and you can read the parts you need instead of sitting through hours.
docs.python.org/3/tutorial →freeCodeCamp's Python track. Project based, so you are typing from early on, which is the only way any of this sticks.
freecodecamp.org →If you picked JavaScript. The clearest free JavaScript book on the internet, and it covers the same six things in the same order.
javascript.info →Mozilla's reference. The thing working developers actually keep open in a tab. Use it to look things up rather than to read cover to cover.
developer.mozilla.org →Step one's real finish line is reading errors, and almost nobody studies them on purpose. This is the official page on what each one means.
docs.python.org/3/tutorial/errors →Get a free key with no card, send a model a message from a script, print the answer. Plus the mistake that costs people real money, and why the model forgets you the moment it answers.
read part 3 →Free tools, honest feedback on your work, and people who answer. Job hunters, builders and learners.