How to Train Your Inbox

Not using Open Source products more often, we are creating unknown damage to our collective knowledge

 · 3 min read

My dependence on Google is scary. I cannot image how much I will lose if I lose my GMail account. Make me wonder how much control of your life are you willing to give to large companies such as Google? When will security and ownership outshine convenience?

And why this dichotomy? As technology gets more and more advanced and widespread, more people should start providing services and it should become easy for anyone to start their own service, like say GMail or provide a directory of the web that is searchable. Why are there such few good alternatives?

I started thinking doing some small experiments.

First, I changed my default search from Google to DuckDuckGo (DDG), the alternate search engine run by Gabriel Weinberg. He started this as a personal project and is now serving upwards of 2 million queries a day. After using the service for a few days, I realized that Google serves better results in more than 75% of my queries. For example, Google is far more likely to return an “official website” of a product or a forum that is more widely read.

So this brought me to the next conclusion, Google Search is so good because we have trained it for so many years. It was not obvious to me earlier, but every time you punch in a few characters in Google Search box, you are telling Google something. Every time you select a result, Google knows you voted for it. The more number of people search for something, the better it usually is. So Google is actually better because more people use it.

Same with GMail and other services. Google knows what is spam because we diligently mark spam. Its machine learning (ML) algorithms are learning with every action I make. So in a way I am making Google better by using it. Google is learning to manage its server in the big bad world of the internet. It knows how to defend against attacks, it has sophisticated system to monitor breaches. It knows all of this because it has so many users. And this is making it harder for anyone to provide half a decent service, increasing our dependence on Google many fold.

Please note that I have nothing against Google. We are caught in a vicious cycle, where we are giving huge amount of data to a private company to and making their products better. Now what would happen if we gave this data to an open source company? The knowledge embedded in that product would become community property and anyone with would be able to setup their own email servers or web indexes.

The size of opportunity we are losing by not doing this is nothing short of a grand tragedy. Taking this analogy to ERPNext, we feel a bit better. All the users, whether they buy our hosting or they use on their own are “training” ERPNext to be a better product. Every feedback we get (and we get tons) is making the product better. Hence we love our free users (almost) as much as our paid ones.

Can we extend the idea to other products? How about EMail? How about web indexes? Personally I would love if there was an easy mailbox I could install on my computer and along with a cloud based directory service (DNS), so that all my mails would directly drop in on my machine. I could couple this with a simple backup service, or a spare hard disk so that I have copies in case my main hard disk crashes.

The reality is that it is extremely hard to do this at the moment. Logically it makes complete sense, but there is no such alternative. Opportunity anyone?

Rushabh Mehta

Rushabh is a software developer and founder of ERPNext. He usually writes about the startup experience, open source and the technologies he is working on.

Add a comment
Ctrl+Enter to add comment

Rushabh Mehta 10 years ago

Fred, thanks for the comment. The problem is that they are not easy to setup and maintain. But we

Fred Blauer 10 years ago

Thanks for your insightful blog posts. But I have a question on this one. There are several open