You Are What You Read
Last week’s article, “On the Death of Amazon’s Writing Culture,” got me thinking about a lot of things.
I had mostly written it for Amazonians. I figured some current and former employees would read and react to it, and that would be that. Instead, it went everywhere. Posts about the article got nearly a million impressions across LinkedIn and X, and many of the people sharing it had no connection to Amazon.
I suspect it struck a nerve for many people because the same thing that has affected Amazon is happening where they work, too.
That article was about writing. But as I read more and more of the responses, I began to think that I may have told only half the story. That perhaps there was more to the Amazon writing culture than writing.
A while back, Business Insider ran a story about me with a headline that I had found funny at the time:”Former Amazon principal engineer says he spent ‘1-4 hours’ reading daily”. It came from something I had said on the Pragmatic Engineer podcast. But more than just being mildly amused, my reaction to the headline was mostly disbelief. How could they turn my offhand comment about reading into the subject of an article? It didn’t seem newsworthy.
Because deep reading was just what we did at Amazon.
I’ve come to realize that there was another equal part to that culture that complemented the writing, and the two parts working together is what truly made Amazon so special. The “secret sauce” had two ingredients: writing and reading.
This issue is sponsored by Linear, the planning tool all sane engineers pick for themselves. It’s what OpenAI, Cursor, Coinbase, Figma and Substack run on, along with Clay, whose AI writing policy I referred to last week. Linear is giving readers from this community six months of Linear Business for absolutely free, with no cap on seats, which works out to around $5,000 for a typical team. Claim it at linear.app/ale.
What you told me
The comment most relevant to where I’m going with this article came from The Wiz: “Nobody is writing their own ideas anymore,“and nobody is reading anything that gets written either.” He’d sat through several doc reviews where the document hadn’t been read. “It’s not just the death of the writing culture, the reading culture is dying too.”
Noble Udoh, who is still at Amazon, confirmed that AI summarization of AI-generated text is a growing trend. The moment he senses that a document he is reading has been written by AI, he tags it AI;DR.
Stop the Slop summarized the whole problem in two sentences. “A doc used to be proof somebody sat with the problem for a week. Now it’s proof somebody has a subscription.”
Then there was LinkedIn. A lot of people pushed back on the article. I have no problem with that; I welcome all forms of feedback. The most common objection to the article was that things were much more nuanced than was suggested by the title and its subtitle (“How AI has turned what was once a superpower into a liability”). Fair enough.
But, many of those replies had been clearly and unironically written by AI. Ugh. It was not clever in the slightest. All it did was make me a little sad.
If you’re a developer or scientist heading into year-end reviews, I’m running a live session next Tuesday, October 27: Visibility for High Performers: How to Make Meaningful Impact and Be Seen Before Year End. Recency bias means performance reviews weigh the work you’re doing now over the impactful work you did earlier this year. I’ll cover how to tell which of your work will count this cycle, how to make sure the people who decide your rating see it, and what to do about the new expectation that you produce more with AI. Grab a seat here.
Garbage in, garbage out
Every programmer learns the phrase “garbage in, garbage out” early in their career. If you feed a system bad data, it will always produce bad results. This also applies to the content we consume.
The opposite is also true. “Quality in, quality out” was the reason why I read so much at Amazon, and the benefits that accrued as a result of that approach formed the basis of that Business Insider article. At Amazon, I could reasonably assume that the documents I read were high-signal, and reading for so long each day improved my thinking.
That assumption was reasonable because writing was expensive back then. A six-page memo worthy of an Amazon presentation would take at least a week to write. Nobody spent a week on something they didn’t care about, so the fact that a document existed at all told you something about the effort that had gone into it.
That world has gone. Why? Because now that writing is essentially free, it has lost its underlying value.
Cal Newport has compared writing to food. Some food is nourishing and healthful, made with care by someone who knew what they were doing, and you’re better off for having eaten it. But then there’s fast food: cheap, convenient, but low in food value.
You are what you eat.
AI writing has created the equivalent of fast food. It’s cheap, and it’s engineered to be easy to consume. But ultimately, it’s empty.
The difference is that food comes with a label and is easily recognizable. You know what you’re getting when you pull into the drive-through.
In contrast, bad writing has no label. (Unless someone before you has tagged it with AI;DR.)
Newport recommends that we treat reading the same way we treat food when we are serious about our health. We need to slow down our consumption and be conscious of what we’re consuming and why. We need to find a few sources that reliably nourish us and go back to them regularly. And we need to stop snacking.
You are what you read
That was the real lesson of Amazon’s study hall, and I miss it dearly. Amazon’s writing and reading culture was the result of the company deciding, as a matter of policy, to sit down and eat a proper meal.
I’d like your help. Finding good sources of nourishing content takes time. Give us the name of a curator that you trust, so that we can share this information with others.
We are what we read, so let’s choose accordingly.
A Life Engineered is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
A Life Engineered will always be a free publication, but if you’d like to support it, we’d be honored if you upgraded your subscription. Paid subscribers get the full experience, including access to the archives.
© 2026 Steve Huynh
3100 Airport Way South #12, Seattle, WA 98134