AI Playbook & Primer (Andreessen Horowitz, A16Z)
Frank Chen puts together explanations and resources for ‘a general audience’ — like you!
I’m encouraged by new resources surfacing that seek to distill this topic’s major concepts into reader-friendly summaries. I wish Frank had built this 5–6 months ago when I was looking for ‘the primer’ on AI, but lucky for you it exists now!
The microsite is a compact walk-through of the fundamentals for history, terminology, APIs, popular applications, architectures and mountains of external resources for deeper reading (no pun intended).
It’s loosely written for people new to the field that are thinking of applying it in their existing businesses, but anyone can get benefit from reading it.
Calling APIs
My favorite sections in the playbook were the guides that focused on bringing in interactive examples, where you could call NLP and computer vision APIs and check out results and JSON outputs. His preceding explanation on the number of existing APIs and how to call them was also helpful.
Concept Overviews
Frank also concisely explains network architectures and types of machine learning with myriad extra reading on both. You can find countless versions of these types of explanations online, but even if you’ve already read a couple, I recommend re-reading Frank’s. Unless you work with these concepts regularly, it can be hard to get them to stick. Repetition helps.
External Links!
Once you get through the guides, you’ll already have enough links to keep you reading for weeks. The bittersweet truth is that you will find no end to the eminently helpful material available online — it’s always growing too!
You may begin to feel overwhelmed once you hit the ‘Where to Next’ section of Frank’s playbook, particularly the page on ‘videos, tutorials and blogs’. It’s not the longest list I’ve seen, but tackling it all at once is not advisable — I recommend cherry-picking from each space and subscribing to handful of the newsletters. Newsletters will give you some space between ‘link dumps’ and give you more current context. The places they send you often end up referencing the best explainers for a given topic anyways (i.e. Chris Olah’s LSTM post).
From Frank’s list of newsletters, highly recommend Jack Clark’s Import AI and Rob May’s Inside AI (formerly called ‘Technically Sentient’, now easily confused with Jack’s). Jack’s has more commentary and humor per link, Rob’s is more extensive and structured — let’s say they’re complementary.
If You’re Experimenting with ML…
Then you must have datasets. Frank helpfully aggregates some of the most popular openly available sets for you to work with here:
If All of This Feels Way Too Complicated…
Best to start with the ‘assume I know nothing’ video primer:
For more from Frank Chen:
Frank Chen (frank@a16z.com / @withfries2) on behalf of the a16z team.
