Today marks four years since I joined AWS. My last day will be Friday.
I must say that being fired from AWS is really a relief. There have been a lot of changes in the company since I joined in 2022, and the company I wanted to work for is no longer the same company.
Last year, while I was trying my best to make AWS good in the open source communities, there were two main drivers that were making me unhappy in my job: organizational change and the accelerating focus on generative AI.
Organizational change came in the form of the guy who hired me, David Nally. I was skeptical about joining AWS, especially since I work in open source, but David reassured me that his team, OSSM (Open Source Strategy and Marketing), is dedicated to making AWS a better citizen in open source communities.
Amazon has a really weird approach when it comes to the people who work there. They view almost all employees as “replaceable.”
Now I first heard the term “fungible” in the context of non-fungible tokens (NFTs), but it basically means “fungible.” Amazon built a huge retail business on processes that could take someone who was relatively healthy and relatively intelligent, and turn them into a productive fulfillment center employee in a matter of weeks. Although this may work for the shipping business, it does not translate as well to information technology, as much of being successful in that business depends on institutional knowledge that must be acquired over time.
It also assumes that there is an unlimited supply of people with the necessary skills and willingness to work for Amazon.
In any case, David called me “non-fungible” during the interview process (which still sounds dirty in my mind but makes me feel proud) and I got the job.
While my official role was to act as a liaison between AWS and those customers who were commercial open source companies, I simplified it to bring a human face to a huge, faceless corporation.
David was a very good manager. In fact, he’s in the running to be the best manager I’ve ever had, although that title is still held by a guy named Jay Clapsaddle (who’s long since retired). He has an intuitive understanding of how AWS works, and he was always pushing me into situations where my unique but limited talents could be put to good use.
Well, last year David was promoted to run the entire AWS Developer Experience organization due to being so good at his job. OSSM is a part of it, but I no longer interact with him in a meaningful way. My “David Time” dwindled to almost zero.
Additionally, last year AWS’s focus shifted completely and almost entirely toward GenAI.
This post is already very long so I won’t throw out all the examples I was going to bring up at this point in the narrative, but we are beginning to be pushed to use AI as much as possible. People were writing things like “I use AI to summarize my emails!”. I mentally responded by saying, “Why don’t we write better emails?”. And the thing that really bothered me was “I used a prompt to create my conference presentation!”
The most valuable commodity in the modern economy is attention. I really appreciate the attention my three readers give to my posts, even when I lose them midway. I love giving conference talks and I spend a lot of time preparing them, and when someone still wants to speak but doesn’t want to do the work, it annoys me. Really, why do that?
It’s gotten better, but I used to see AI generated images with a lot of illegible handwriting or misspelled words in the slides, but the speaker left them out anyway. “Good enough” is not a customer obsession.
In all this pivot to GenAI, AWS has lost its focus on the customer. Instead of working backwards from the actual customer need, the goal is to create as many things as quickly as possible, throw them out into the world, and see which ones get traction, whether they meet the actual need or not.
There is an emphasis on using AI to create content, which will ultimately be consumed by AI, and we have lost the human in the process.
When AWS first introduced a viable cloud to the world, it was amazing. In the 1990s when you wanted to implement an enterprise software solution, the first thing you had to do was estimate what computing power you would need. Next, you have to order hardware from companies like Sun Microsystems or Dell, and delivery can take weeks, if not months. Then it would need to be racked up, operated and provisioned, and then you would be penalized if you undersized it or criticized if you spent too much and made it larger.
The cloud solved those problems, and AWS set the standard with services like S3, EC2, RDS, etc.
Go to Re:Invent these days and try to find a session on those devices. While you can, AI will still dominate presentations.
This whole matter made me question my role. My personal goal is to make AWS the default choice for running open source workloads, but what’s the point when you can “vibe code” the same functionality bypassing the license?
The customer focus on AWS has also changed. Instead of appealing to people focused on the infrastructure needed to build stable and feature-rich applications, focusing on the level above that has become abstract, because the whole promise of GenAI is to make those no longer necessary; To make those people “replaceable”.
The accomplishment I’m most proud of in the last year involves restoring a suspended AWS account. The financial impact on the company was negligible because this customer was not a big spender, but they are one of the people who made AWS successful in the first place.
A man in North Africa posted that his decade-old AWS environment had been shut down without notice and without any recourse. In fact, they were told that their data had been deleted.
I reached out to her to see if I could help, but I wasn’t optimistic. If his data was gone, it was gone, but I really wanted to capture as much as I could about the experience to prevent others from going through this.
In the process of converting this person from an account number to a human being, I learned more about his situation, although I won’t share the details, but losing his AWS account was just one of a long list of issues he was dealing with at the time.
To make a long story short, I was able to restore his resources. I just tried to catch the right bear and the support team did the rest (and they were amazing). He wrote a nice post that mentioned me, but the main point of it was that this issue shouldn’t have happened in the first place.
No one in senior management seemed to care after the case was closed, but this attitude was not ideal, especially among the rank and file. When that post hit, several random Amazonians pinged me on Slack to thank me, some even saying I renewed their faith in the company. It was hard because no one in leadership cared that I did it.
This past year has been difficult in other ways. There were massive layoffs last October but it did not affect many of the people with whom I worked closely. The January mass layoff was bad, and I had made several friends at AWS who were now looking for work. Stress affected my health. I’ve gained another ten pounds (bringing my four-year total to almost thirty), I constantly set new high scores on the blood pressure machine, and my sleep has become so disrupted that I haven’t had a good night’s sleep in weeks (I wrote most of it in a hotel room). [checks watch] 1 o’clock).
I can’t stress enough that AWS employs some amazing people, but between reduction in force and people moving on to better companies, I’m not sure how much longer that can be maintained. Many good people have left on their own and others like me have also been asked to leave.
Then there are the many things that made me embarrassed to work at Amazon. Cory Doctorow did a long post on how Amazon creates “reverse centaurs.” Not a single Amazonian I worked with could read it and feel the slightest bit ashamed.
One thing AWS does right is that it allows calling Slack channels #actual-aws-memes to exist. Although it’s highly moderated, it’s a place for people to vent by posting memes about life on AWS. I posted my first (and apparently last) last week.
Note that I don’t believe I got fired because of that meme, and I want to emphasize that in my four years at AWS I have never been asked to do anything I found unethical, much less illegal. But there seems to be a level in this country, and in the world in general, where following the law becomes optional.
I didn’t know what my future held at AWS, so being forced to leave is really a relief. After attending GrafanaCon this year, I really wanted to get back to my open source roots.
Open source has always, at least for me, been about putting technical power and control into the hands of the user, not the vendor. How will this play out in GenAI, when every state-of-the-art model can only be accessed through APIs? Even if you want to try running the model locally, who can afford the hardware?
And what do you do when your job is to be a human in a world of AI?
<a href