Curated Content October 2022

Articles

How to solve a real machine learning problem with Nx

An awesome post from Tiago Davi, a member of one of my teams at Savi Solutions. This one covers how to use Elixir's Nx in order to solve a linear regression problem.

This computationally heavy work would be something that Elixir would struggle with traditionally, but thanks to Nx is actually well within reach of Elixir from a performance perspective.

Know your carrying capacity

An interesting idea from Mac Chaffee to periodically think about the tasks and/or knowledge that you're carrying uniquely, especially the non-obvious ones, and ensuring that's something that is within your cognitive ability to keep up.

Books

Lead from the Future - Mark W. Johnson & Josh Suskewicz

A book on how to ensure your organization is comfortable cannibalizing itself, and creating continuous innovation.

The authors do this by introducing a future-back mindset, which is where you project yourself into the future based on extrapolating certain trends, and walk backwards into what that implies for your business through concrete tools and frameworks.

Would recommend this for all product focused executives who are already familiar with folks like Marty Cagan and Teresa Torres work and want to sharpen their thinking on larger time horizons.

The Choice -  Eliyahu Goldratt

Rather than the usual business fable approach to teaching the theory of constraints, Eliyahu and Efrat Goldratt engage in dialogues to teach the philosophies behind the theory of constraints and how they can be applied to life broadly.

I found this one really interesting, and would recommend it.

Conference Talks

You are a Program Synthesizer

An interesting technical talk from an under-appreciated tech author - Jimmy Koppel.

My favorite takeaway here is the analogy drawn between citrus advice and precise advice in the treatment of scurvy and software engineering processes.

He boils this down as:

The advice to eat citrus is good advice. It saved people’s lives. But it was not precise enough to capture the reality, and so it came with this long list of caveats that made it hard to use.

The precise advice is to get vitamin C. And while it took about 200 years to get from citrus to vitamin C, this is the ideal to be aiming for: the simple thing that explains all of the caveats.

We have many kinds of advice that are citrus advice in software engineering.

He then takes us through different kinds of logic, how they're used in program synthesizers, and how you can use them to clarify your thinking about programming.

Podcasts

No podcasts to recommend this month.

Tweets

I love empirical based software engineering. Unfortunately as Hillel notes in some of his other tweets, aside from the fact that code review moves the needle, there isn't much we actually know about software engineering, but this one seems like it may get there, or if not, actually merits further investigation outside self reporting, using actual observation.