In the extremely early days of a startup one doesn't have much optionality. With only three people at the company it really did not make any sense to hire a full time SRE for a pre PMF startup. I'm glad we made those trade offs back in 2011-2012 so that Firebase could be here today to be backed by a dedicated Google SRE team.
Don't mince words. In the extremely early days of YOUR startup, YOU didn't FEEL like you had much optionality. That is in no way universal. You very much do have optionality. You can quantify how much it costs to hire someone part time or full-time. You can quantify the risks of burning yourself out, or your manager can. Your firm had the optionality, you just chose to go in a very specific direction.
What I'm saying is that your firm survived in spite of, not because of this frankly amateurish decision.
Obviously nobody can take away your contributions towards making Firebase successful. I think many people may not realize the kind of fanaticism that it takes to get a startup off the ground, considering the odds are stacked against you.
I am curious though, how well did you sleep that year?
Thank you! Are there other examples of new improvements to the development process you've come across? Currently we're positioning Gold Fig to live alongside source control and IaaS. However, we feel there's a huge opportunity to bring the same level of rigor that "version control" encodes to all the (traditionally) non-technical SaaS, enrich existing process (e.g. Gold Fig's log can add more color to your PagerDuty alerts), etc.
Thank you! We are definitely looking at what else is happening in the space with companies like Pachyderm (https://pachyderm.io/) and DVC (https://dvc.org/features). We're also starting to see the development model starting to make its way in to more and more consumer apps as well (Think "document history" in things like Google Docs). Our view is that as more and more of these services specialize the analogous "Version Control for X" is inevitable.
Agreed that the surface area around the out-of-band changes is a concern. We find that educating our users around getting the extension installed team wide is the primary approach by way of policy (in GSuite orgs, you can install extensions team wide) or training (on-boarding docs/new hire orientation). Just as with source control we can commit straight to the main branch, it's through policy/training that we do things like mandatory pull-requests/reviews to not "allow" that. Long term we envision Gold Fig plugging in to your projects as a read-only member to pick up any other out-of-band changes.
I have many fond memories of Print Shop! I remember that program would also print "greeting cards" that you'd fold in quarters. The banners were a pre-birthday task that'd never get overlooked :)
It's interesting that an entire category of features fell by the wayside as a byprodct of the elimination of form-feed/perforated paper.
I fondly remember Print Shop printing cards and banners on the Epson LX-80 dot matrix on the Apple IIe monochrome (green and black screen). Piecing a banner together these days from a series of 8x10 sheets is just not the same. :)
