• Former Amazon principal engineer Steve Huynh says every company should replicate Amazon's culture of reading.
• On the Pragmatic Engineer podcast, Huynh said that he spent one to four hours every day reading six-page memos.
• He said meetings with a "study hall" where attendees read the memos ensured everyone was up to speed.

When Steve Huynh was a principal engineer at Amazon, meetings began with a "study hall."

Amazon had a "reading culture" even among engineers, Huynh recently told the Pragmatic Engineer podcast, speaking of his time at the tech giant. Employees frequently drafted six-page memos, he said, which they shared with the company to update progress and demonstrate new projects.

"I spent on the order of like 1-4 hours every day reading while I was a principal engineer," Huynh said. "What an amazing culture that I think that almost every other company should replicate if they could."

Huynh, who said the company's embrace of writing and reading the 6-page memos was part of its "secret sauce," said Amazon employees' writing was often constrained to the format during his tenure at the company, whether it was a business strategy, system design, or press release.

Huynh started at Amazon in 2006, only a few years after the company turned its first profit and while Jeff Bezos was at the helm. Bezos famously instilled this culture of memo-writing from the top down.

Bezos insisted on dense, direct memos in 10-point font. In his 2017 letter to shareholders, Bezos wrote that "we don't do PowerPoint," instead opting for these six-pagers. "Not surprisingly, the quality of these memos varies widely," he wrote.

Before meetings, Amazon employees read these memos together. On the Lex Fridman Podcast in 2023, Bezos explained why he didn't ask employees to read the memos in advance.

"The problem is people don't have time to do that, and they end up coming to the meeting having only skimmed the memo, or maybe not read it at all," Bezos said. "They're also bluffing like they're in college, having pretended to do the reading."

Andy Jassy, Bezos' successor and Amazon's current CEO, has worked at the company since 1997. When first pitching what would become Amazon Web Services, Jassy described writing his own memo.

"I remember this six-page narrative, we called it a vision doc. We asked for 57 people, which felt so ballsy at the time. I was so nervous, I wrote 30 drafts of this paper, and Jeff didn't blink," Jassy said in a 2017 talk to the University of Washington.

Jassy has continued the culture of memo-writing under his own leadership. In his 2024 letter to shareholders, Jassy wrote that a mere six-page allotment made the memos "much easier for the audience to engage with and ask the right 'why' questions."

"I got really really good at just reading these documents to get up to speed," Huynh said on the podcast, explaining that reading enough six-page memos taught him to express himself in the same format.

Huynh no longer works at Amazon. He left to pursue YouTube content creation full-time, as he told BI in 2024. But Huynh still reveres the company's reading culture — even if he acknowledges it may not be easily reproducible.

"The difficulty would be, you actually have to be disciplined and principled," Huynh said. His interviewer, Gergely Orosz, argued it could only be done from the top down. Huynh agreed.

• Shahad Ishraq was a systems development engineer at Amazon in Germany for just over three years.
• He quit at the end of May after the company implemented a five-day RTO.
• Ishraq said Andy Jassy's memo wasn't surprising, and he isn't concerned about his future career.

This as-told-to essay is based on a transcribed conversation with 30-year-old Shahad Ishraq, from Germany. The following has been edited for length and clarity.

Reading Andy Jassy's new memo on generative AI, I'm not surprised by anything.

I worked at Amazon for nearly three and a half years and left in late May because of the 5-day RTO mandate. My commute took an hour and a half each way, so I wanted to move to another job where I could still see my career progressing and do interesting work.

The memo feels consistent with what I'd been hearing from management and Jassy while working at Amazon. I think Jassy's comments are to show shareholders he's invested in the technology.

I initially worried about AI when I wasn't as familiar with the products. However, using AI will give you a better understanding of what it's capable of and what skills you can develop to differentiate yourself.

At Amazon, there were some eager adopters of AI and some skeptics

I joined Amazon in 2022 as a systems development engineer, working in Leipzig, Germany. My day-to-day work involved designing and implementing software and performing operational tasks.

When AI tools first came out a few years ago, we were told we could use them, but we should be very careful and follow the company's policies on their usage.

Amazon is a huge company. Within it, I spotted and heard about different approaches to AI adoption. There seemed to be a bunch of excited early adopters who shared their findings with everyone. There were people like me who followed the first bunch and saw what went well. There were also some skeptics and a small number of engineers who were outright against using AI.

For me, the dawn of AI was a bit scary at first. Everyone was saying it would put me out of a job. Unless you test the technology yourself and see what it can do, you'll fear the unknown. AI wasn't part of my job until sometime in 2024.

There were also some barriers to using the technology. When I first joined Amazon, ChatGPT wasn't even available, but when it did come out in 2022 we couldn't use it that extensively because of data security issues that come with copying our code into those models. When Anthropic's Claude became available within Amazon Bedrock — the company's internal service for developing generative AI applications — we were able to make more use of AI.

In my last few months at Amazon, I started experimenting a lot with approved AI tools, doing extensive tests with them. They don't do everything for me, but I've integrated these tools into my workflow, such as by asking it to create a plan for my tasks or spot differences between documents.

I noticed it often fails, and I have to make changes, but overall, it has improved my speed and increased my throughput significantly.

AI won't eliminate software engineers anytime soon

Andy Jassy's memo feels very consistent with what I've been told internally before I left Amazon and what the company has communicated publicly.

News articles talking about the memo focus on Jassy saying that a lot of jobs will be taken by AI. However, in the same sentence, he also says jobs will be created.

I've tried creating production-level applications using AI, and it takes a lot of effort to get these products ready. A company like Amazon can't roll out an application that breaks and causes havoc. They have to have firewalls, checks, and tests.

I don't see people going out of jobs in huge numbers soon. Amazon went on a hiring spree during COVID. If we see more layoffs, I think it will be associated with cutting back after that spree, rather than the impacts of AI.

AI agents are helping out software engineers a lot, and the amount of work agents do will probably increase gradually. I'm able to get agents working on three different things, while I look into other tasks.

But AI hallucinates quite a lot. It does things it's not asked to do. I often have to correct an AI agent producing code. Humans will be required to build guardrails and act as guardrails themselves. Implementing these guardrails will take time, and I think this will slow down the AI agent hype.

There's nothing new in Jassy's memo

I think Jassy's comments about AI are to show shareholders he's invested in the technology. Memos have to come out. Jassy has to place a lot of optimism around AI; otherwise, shareholders will think they're not doing anything with AI.

My advice to Amazon employees is to start using AI as much as possible to overcome their fear of the unknown. I now work as a software engineer at a utilities company. The more I've been using AI, the more comfortable I feel about myself. I can see what skills I have that I can use to stay relevant.

In tech, languages and developments come really fast. My guess is that people will need to use AI to write code and increase their throughput, and pure software engineers will gradually be replaced by people who have both software engineering and AI skills.

I'm personally trying to learn these skills because I think they'll become more important.

A spokesperson for Amazon told Business Insider, "Amazon employees use internal generative AI tools every day to innovate on behalf of our customers. We have safeguards in place for employee use of these technologies, including restrictions on sharing confidential information with third-party generative AI services."

Do you have a story to share about the AI job market? Contact this reporter at [email protected].