Free as in “a free beer” (as Stallman’d say). People doesn’t use tools
like joomla or wordpress because they’re free as in freedom, but because they’re
gratis. Welcome to the world of the cheap web.
The cheap web is a market, and it’s a huge one. But it’s a place that is dominated
by asymmetric-information between buyers and sellers as in
the market of lemons
(George Akerlof, Michael Spence, and Joseph Stiglitz won the Nobel Price for this research).
What happens in the market of lemons is that the quality of goods fall.
In the previous posts we talked about how to
predict a continuous value
using a linear function and a way to
measure the error
given a matrix of test data and an hypothesis set of values theta.
In this post we’re describing a function that will converge the vector theta
to its optimal values (or local miminum) which is called gradient descent.
In the previous post we covered the hypothesis function,
which is the function that will predict a value given a set of features for a new unknown case. In this post we’re
going to build the cost function, a way to mearsure the error of the prediction function with a specific set of weights .
Python is becoming the defacto standard for Big Data and Machine Learning,
in particular this is because of some amazing tools like IPython Notebook
that helps visualizing your data or scikit learn
some of the most popular machine learning algorithms.
So implementing a ML algorithm in go is a pure excercise.
This is the situation: the team is very small, just 3 people in the engineering department.
It’s a young startup in a early stage with a very tight budget and the engineering
team is building the product as a side project, aka in the spare time.
If it wasn’t enough the team is spread in different cities, so being in the same
room frequently is impossible. It’s pretty obvious that the team cannot follow
a regular scrum approach to development.
These days I’m working for a stealth startup in a very early stage. One of the best team I’ve ever worked with. A couple of days ago we had an interesting discussion about the UX and the priority of an additional feature. I’m talking in the specific about a map view of the items additional to the default list view, but it doesn’t really matter. In this case my point is very easy: since we cannot measure the impact of this UX feature (because we basically have no user base yet for an A/B test) it’s not important by default.
Bitcoin sign has been accepted into unicode. Here’s the symbol: ₿ (it will work someday)
People who know me is aware of my aversion against social networks and blogs. So why a blog?
I’m thinking of this space as a personal set of notes and thoughts about technical things that I can share with friends and colleagues.