317 karma · joined April 15, 2011
I might make open-source some of the "Git toolkit" I built for this app, but it's too early to say.
The reason for expiring builds is that 1) it's pre-release, 2) it's a software intended for professional engineers and 3) the app is rapidly evolving, so I don't want to have people using old versions which may have bugs fixed since then: http://forums.gitup.co/t/gitup-release-notes/16.
Then on top of it is an extensive toolkit I built in a few months for repo manipulation and rendering.
Mostly because I didn't have time for one and also because I wanted not to require people to listen to some speech to understand how GitUp works (which you can't always do at work anyway).
In such cases, GitUp just prompts you if you want to force push.
If you decide to force push, then your updated commit message is on the remote repo as well. This is completely safe if this is your private repo, or you are pushing to a private branch. In any other situation e.g. a shared repo with a team or a public repo on GitHub, you should pretty much never force-push as this will mess up other people clones (as there are now 2 histories of the repo in a way).
The Map view documentation does have a large warning on rewriting history (which obviously is required for a number of operations), but I agree that an in-app warning might be good too: http://forums.gitup.co/t/using-gitup-map-view/34.
I think some people like to be all high and mighty about security issues and derive pleasure from pointing out (often adamantly) what they see as flaws while often these are expected and well understood side-effects, and in no way "real" security issues.
Of course, it could be used poorly, like a number of other libraries (not validating HTTPS certs anyone?), and introduce a security issue which would only have an effect in apps that deal with sensitive data and transmit said data using this Telnet library.
IMO it's no different from running, say, an HTTP web server from an iOS app which a number do.
The primary use of this library is for development and field debugging and for that it can be very valuable: you can connect to your app (while it's running of course) from pretty much any computer. There are also some use cases in production that expose no security risks (apps are sandboxed after all and far from every app deals with password or sensitive data).
I provided some extra context and counter-points here: https://news.ycombinator.com/item?id=7053473.
1) IMO you can't apply such "idealized" reasoning to an early stage consumer startup: the vast majority run at a loss. It's expected and as a matter of fact not a single investor had problems with our financials. For better or worse, Everpix was on the VC track, not the bootstrap track.
2) You're essentially assuming that infrastructure costs are set in stone. I addressed this point in another comment: https://news.ycombinator.com/item?id=7043555.
3) You're assuming photo collection sizes are also constant, which is not the case: early adopters will have more photos than mass consumers. For instance IIRC average subscriber had 15K photos, while we would expect the average consumer to have 5-10K. This gets compounded with infrastructure cost savings.
4) You're assuming the only revenue stream are subscriptions. We had concrete plans post Series A of printing based monetization leveraging the uniqueness of our photo platform.
So what you demonstrated is that, assuming some hypotheses, it would be really difficult to run profitably such a business on AWS. Again, I agree with you: none of the large players do it on AWS (I know for a fact and that savings would be 3X-8X), and our #1 engineering goal post Series A would been to get off AWS (since we would have had the capital and resources to do so).
I would however counter that using a different set of hypotheses, more appropriate to such a business, combined with our "track record" of driving infrastructure costs down, we would have fixed that marginal loss short-term. I am actually 99% sure of that but of course I cannot prove it :)
On a side note, some questions which IMO are more critical to address are for instance what's the total size of the market or how much does it cost to acquire customers at scale? Now these are externalities you cannot "change" contrary to infrastructure costs. So if these don't work out, then your business can absolutely not be viable. I would love to read blog posts as detailed as yours on such topics.
Thanks again for writing this insightful post!
PS 1: About infrastructure costs, I shared some extra context here: https://news.ycombinator.com/item?id=7041640
PS 2: About the team size, Android & Windows were actually outsourced to contractors. Regarding the rest, we're going to have to disagree that a quality and complex product like Everpix can be built with a smaller or less experience team in such a time frame (for instance, you're ignoring the science part and I assure you no junior iOS dev can built highly responsive and dynamic apps displaying 20,000 photos) ;)
The reason we were getting closer and closer to being positive on variable costs (looking at revenues on a sales recognition basis, since from the sales volume perspective, things were already positive in the last few months) is, yes, improved monetization, but more importantly AWS optimizations. We had squeezed a lot out of S3, then EC2 and our last step, the one we were working on before shutting down, was RDS where there was a ton of room. Looking at our trajectory, I'm pretty confident we would have been positive even on AWS (but unlikely much). You can even compound that with AWS discounts which apparently even startups can get looking at other comments in this thread.
To re-iterate, you would not build a freemium business like Everpix, with intense computing and storage requirements, at scale on AWS. Large photo platforms have their own storage and servers. Everpix was never intended to grow out of AWS either and you know you can cut your infrastructure costs to at least half. That was the plan post Series A. Most CTO / CEOs of large established photo companies I talked to said they were doing much better than this 1/2X, so that's conservative (I've been told ranges of 3-8X in savings vs S3 for instance).
Anyway, Everpix was bringing about $6 revenue / user / year IIRC, which is the part that really matters considering the denominator, infrastructure costs, is something you know you can really bring down. One relevant reference point is 500px for instance which is doing about $1 revenue / user / year [1]. I would be surprised if they were growing "sustainably" too ;)
[1] http://techcrunch.com/2013/08/07/500px-scores-8-8m-series-a-....
However, on the database side of things, there was a lot of room for optimization (and we only got to some of it).
However it's important to keep in mind in a situation like this, and at this scale of burn rate, you're not looking for a solution to shave off a few thousands $ left and right, which is kicking the can down the road (if even), but for a long term solution.
The key problem here is that I don't think you can realistically switch a business built on the VC funding track to the bootstrap track without the collateral damage of killing your company's essence i.e. the product.
For instance: when suggesting to have fewer employees, that is implying you have room for that or that your employees are interchangeable, which is far from the truth. It's not like we had 5 tech support people, and we could have downscaled to 3. We had hired quite slowly and were already as lean as it gets i.e. 1 person per project: iOS, Web, Infrastructure, Science, Design, Imaging / QA. All these folks were working at full capacity managing a multi-platform service managing hundreds of millions of full-res photos with ~100 servers, displaying photo collections of 10,000+ photos in very responsive apps, and running cutting edge computer vision science at a scale no other consumer photo startup was doing (the closest was the semantic search in Google+ Photos). You can certainly cut salaries (which we did and only works up to a certain point in the very competitive SV hiring market), but you can't remove positions which would be akin to extending a bridge to nowhere.
Regarding the infrastructure expense, I shared some thoughts here: https://news.ycombinator.com/item?id=7041640
I can assure you that without the facilities of AWS, which saved us a ton of time and development overhead early on, there would have been no Everpix. Period. Actually, the very first Everpix was on Google App Engine, but that's a different story.
As an early stage startup, whose core business is not building infrastructure, do you really want to be in the business of managing for instance your own S3 type storage with the same performance and reliability at 200+ TB scale? We never lost a single user photo or account. Never lost a database server either. When your users trust you with their life photo collection, that does matter.
Then of course, if your company starts gaining traction, your bills increase so things change. As a matter of fact, post Series-A, our plan was to switch out of AWS, concurrently to our infrastructure's redesign to go from managing 100s of million of photos to billions.
Doing the switch earlier would have been premature IMO: there was no one the team with such experience, even less bandwidth. The opportunity cost would have been significant on all the other aspect of the company. Let's not forget the required upfront capital which takes some time to recoup so savings would not kick in from day 1. Finally there's also the fact our infrastructure was still rapidly evolving (type of EC2 machines used, database architecture, etc...), so even buying reserved instances was very difficult as there was a good chance of buying machines we would not need anymore 6 months down the road ultimately wasting money (it did happen).
So what we did instead what focusing on driving continuously driving down AWS costs through various optimizations - if you were to look at our dashboard graphs, you would notice AWS costs were growing slower than users / photos, which reflected that work.
Long story short, the infrastructure was paying for itself through subscription revenues. The real "killer" was payroll. I use quotes because in this type of business, it's expected to have payroll your largest expense - there was no surprise. BTW note that the team was already 1 person = 1 entire product component like iOS or infrastructure so quite lean and highly productive. For a VC backed consumer business, you pretty much need the VC money to cover these fixed costs until you reach profitability.
It's far from simple though: you need to find the people who can build such systems and also be able to afford the capex. And then there's the risks of running your own system and opportunity cost.
It's really unclear we would have been able to go this far in the first place without AWS.