[mtlynch]

Thanks for reading!

It sounds like we had similar experiences. I agree that the biggest benefit has been the freedom to explore my interests. At Google, I always wanted to get deeper into machine learning, but I felt pressure to lean into my strengths so that I could deliver something and get promoted rather than slow down and learn something new.

Now that I'm on my own, I do still feel pressure to release something, but it's also a lot easier for me to control the pace and devote more time to learning because I'm on my own roadmap.

[dwoot]

Hey Michael - if you don't mind sharing, which team did you work on and why did you feel the pressure to lean into strengths instead of learning something new? Could it be that this was just your perspective on it at the moment? The alternate perspective that I see is that learning something new pays off into the future even though it may not feel like you're moving somewhere at the moment. Did existing responsibilities require too much bandwidth that it made attempts at fitting in any learning a daunting endeavor?

Additionally, did you feel that you had the freedom to learn and apply ML?

[mtlynch]

Most of my time was on Google Maps.

The reason it was hard to slow down and learn was that promotions were so difficult to get. You can only apply every six months, and so many stars have to align in your favor to get promoted and so many things can derail it. Your project can get cancelled. Your project can flop even if you personally did everything right. One of your senior teammates (whose word carries a lot of weight with the promotion committee) can leave the company.

It was so hard to get promoted when I was trying to get promoted that I felt like if I wasn't focused on promotion, I'd just never get a promotion.

I wrote a longer post about this a few months ago:

https://mtlynch.io/why-i-quit-google/

>Additionally, did you feel that you had the freedom to learn and apply ML?

I theoretically had freedom to learn and apply ML, but it would have been at the expense of my career. For example, my last major project was to launch a new ML pipeline in six months (it sounds like a long time, but it's really hard to do this with all of Google's bleeding edge infrastructure and bureaucracy). I was leading a team with two other developers, both of whom had graduate degrees in ML-focused subjects. I could theoretically have assigned myself more of the ML work, but the most likely way for us to meet our deadline was to let my teammates handle the deep ML work while I did more infrastructure work.