Loss, Compassion, and leaving AWS

At the start of 2020, I had big plans. I was going to create. A lot. I had multiple startups I wanted to launch throughout the year. I had a lot of goals.

Then, like everyone, I was forced to adjust. Family died. Work situations changed. Everything changed.

—

I threw out most of my plans this year to deal with myself, my family, and improve my own mental state. COVID really affected me. At the end of April, I devoted a lot of time to creating, maintaining, and moderating ThisWebsiteWillSelfDestruct.com.

I haven’t written about this much, but ThisWebsiteWillSelfDestruct was really about what I needed in my life at the time. I needed a place to write about what was going on in my life. I’d recently been to the funeral of a woman who had meant a lot to me, and I was having a very hard time internalizing her death. I was seeing the death toll rise. I didn’t know how to handle it.

I theoretically needed to write about it. To journal. To just get out my feelings to someone, instead of focusing my anxieties on doom scrolling. And I thought some other people might feel the same way.

It’s been incredible reading well over 100,000 kind and honest messages from people in the same situation. And I think it’s been useful to others, as they’ve been read over 25 million times.

It was a distraction, and it was needed.

—

Let’s jump to December, and all those big plans from January are mostly a distant memory. But I still remember that I expected to spend a lot on AWS this year. At least $5,000. Instead, I’ve spent $130.55, all in the past two weeks.

I did a real launch of my first planned project for 2020, ListenLater.fm, last week. It had launched to a very early alpha in February, but then I had to stop putting any time into it, and it had $0 in AWS spend in January thru November of this year.

Well, nothing gained, nothing lost, you might say. Next year I’ll have another chance to make something new. And that’s mostly true.

But there is this one cost that hurts. At the start of this year, I paid $1,000 to be a part of an AWS Activate partner program to gain $5,000 in AWS credits.

I built ListenLater.fm on AWS rather than Google Cloud Platform exclusively because of this offer. They both offer the same $4/million character Text-to-Speech service, and Google offers a slightly more name brand Text-to-Speech implementation in WaveNet for its higher quality service. But as a bootstrapping founder, I went with the service that would give me a bit more runway.

Except those credits expire at the end of this month. By the end of the month, I’ll have essentially paid $1,000 for $200 in credits. My mom was right, I truly am a great business person.

—

I’ve reached out to AWS Activate’s support staff. Even if I could have the credits extended the six months where I put absolutely zero energy into AWS and development because of COVID, I’d be thrilled. But no. It didn’t matter that I’d decided to build on AWS because of the AWS Activate program. It didn’t matter that I was requesting a one time COVID hardship extension to the credits. It didn’t matter that I’d incurred $0 in charges from signing up thru December. It didn’t matter that I’d actively promoted AWS Polly to developers as an alternative to Google’s WaveNet, creating samples walking people through documentation.

I was left with a form letter that AWS Activate doesn’t do extensions, and that my account will never be eligible for a $5,000 AWS Activate credit again. I’d wasted my free credits on family, depression, anxiety, and COVID. Thanks 2020.

I sent a followup, trying as politely as I could to make sure it was clear what I was asking for, and that it was a little frustrating to receive a form letter back. I received a few sentences:

Thank you for reaching out to us. We understand the tough situation and we apologize for the inconvenience. However as previously mentioned the final review is that Activate Credits cannot be extended for any reason as this has remained the policy of the Activate program. 

Thank you for your understanding and we hope you have a wonderful day.

Well, at least they understand the tough situation.

—

This kind of response hurts.

Support matters to me. I really try to be on top of support. I try and answer emails as soon as I can, and even in two cases this week have pulled off the proactive emailing of a customer about an issue before they realize it or contact me. I try and do what I can to accommodate reasonable requests. It almost never takes a lot of time, and it usually helps someone.

Maybe my request is less reasonable than I think. Maybe it’s ridiculous to think that in 2020, we should rethink hard line rules in light tough situations.

I don’t know. But I just know that I fundamentally don’t understand.

—

I’ve been in touch with Google Cloud Platform’s support. It’s been nice that they’ll actually talk to me on the phone, and seem to have at least some level of freedom in working with individuals. So in early 2021, I’ll be moving new users of ListenLater.fm to Google Cloud Platform’s WaveNet text to speech.

I don’t think GCP is any panacea of customer support. But I know it can’t get worse than this.