Product Management

People often ask “What is a product manager?”. Shreyas Doshi says the role is to “define the product & orchestrate actions across the org to enable its success”.

Another question we can ask is “Why do product managers exist?” Because of consumer choice. When the industrial age created consumer choice, companies needed to provide differentiated value through R&D and marketing - the brand and the brand manager were born. Consumer choice exploded even further with the internet age. User experience is therefore critical to software, and the customer voice must be embedded deeply into software development - that’s why software product managers exist.

The canonical book in the space is Marty Cagan’s Inspired. The canonical newsletter is Lenny’s.

Ken Norton’s classic essay on how to hire product managers doubles as instructions on how to be hired.

From the perspective of an investor-founder, Michael Seibel breaks down building products at start ups with great clarity.

A lot of product training sucks, Jeff Patton’s course does not, it’s great.

Strategic thinking is key to product management, Good Strategy / Bad Strategy is a defining modern text on the subject. Ben Thompson’s Statechery blog and podcast analyzes tech news with a strategic lens.

If you are an enterprise product manager it pays to understand how companies ensure productivity, there is no better book on this than High Output Management.

Writing

Good writing is indistinguishable from good thinking.

The Sense of Style defines style and teaches intuitively how to develop a classical style. If you have less time, watch the talk.

Writing Down the Bones by Natalie Goldberg, taught me about the habit of writing.

Paul Graham is sharp about many things including writing, and he thinks writing will only become more important with AI.

David Perell translates writing into the 21st century, follow his twitter.

Artificial Intelligence

AI is important and complex, but it is not impenetrable.

For the non-technical, here is a fun illustration of how Random Tree predictive models work. This is not the type of “modern AI” (i.e. deep learning) that powers Chat GPT or self-driving cars. But Random Tree Models are still one of the most accurate and important predictive techniques in use today, and this illustration is an approachable way to understand how computers find patterns in data - which is the essence of machine learning.

For those who want an intuition of how “modern AI” works, 3blue1brown explains really well how to think about the math behind neural nets without doing any actual math.

Karpathy is one of the luminaries of the field and he has a gift for teaching, his deep dive into LLMs like ChatGPT is the perfect mix of simplicity and detail.

Python is probably the coding language you want to learn if you want to get into machine learning. Learn python the hard way follows the philosophy that the hard way to learn code, literally typing stuff, is actually the easy way. They’re right. machinelearningmastery.com teaches machine learning with a similar philosophy. Karpathy’s Zero to Hero is that for neural nets.

A good one stop shop for learning AI is from Andriy Burkov, The Hundred-Page Language Model Book.

Andreessen Horowitz is an important venture capital fund that has kept a close eye on AI for years, they’ve done a good job of collecting what they consider the canonical documents on modern AI as of 2023. Make your way through this and you’ll have a strong foundation.

Lastly and perhaps most importantly, much thinking on AI ethics is shallow, Vervaeke is anything but.