Pebbling up

“If we knew what it was we were doing, it would not be called research, would it?” ― Albert Einstein

[mathiru-net]

[mathiru-net]

pywit - Python library for Wit.ai

[mathiru-net]

Early on, all of our movies suck. That's a blunt assessment, I know, but I… choose that phrasing because saying it in a softer way fails to convey how bad the first versions of our films really are. I'm not trying to be modest or self-effacing by saying this. Pixar films are not good at first, and our job is to make them go… from suck to non-suck.

Ed Catmull, Pixar CEO (quoted on BBC)

[mathiru-net]

If something that seems like work to other people doesn't seem like work to you, that's something you're well suited for.

Paul Graham

[mathiru-net]

Defining roles: CTO & VP Engineering

When technical titles get overloaded

CTO's are often the worst manager, and the best coder. But the title of CTO is a fallacy here — it means that they are the producer of the solution, not the team leader. They are the person who is referred to (or incredibly hated) when new coders come to the table and ask about what part of the code does what.

[mathiru-net]

[mathiru-net]

Innovation: La démocratisation des outils est le vrai facteur de rupture

La vraie rupture, c’est le développement d’outils toujours plus puissants, toujours plus simples et toujours moins chers, qui démultiplient les possibilités et, surtout, la productivité.

[mathiru-net]

What is a metaclass in Python?

Fantastic introduction to metaclasses in Python.

[mathiru-net]

Beginners' Guide to Concurrency and Parallelism in Python

[mathiru-net]

Seedcom is a newly founded venture fund in Vietnam. It’s one of the silent outfits building startups at the heart of Vietnam’s startup scene. If you’re a foreigner, it’s likely you’ve never heard of them, because the founders are homegrown Vietnamese successes.

[mathiru-net]

[mathiru-net]

[mathiru-net]

This tutorial offered by Standford will teach you the main ideas of Unsupervised Feature Learning and Deep Learning. By working through it, you will also get to implement several feature learning/deep learning algorithms, get to see them work for yourself, and learn how to apply/adapt these ideas to new problems.

[mathiru-net]

We do a lot of interviewing at Palantir, and let me tell you: it’s hard. I don’t mean that we ask tough questions (although we do). I mean that the task of evaluating a candidate is hard. The problem? Given a whiteboard and one hour, determine whether the person across from you is someone you’d like to work with, in the trenches, for the next n years. A candidate’s performance during an interview is only weakly correlated with his or her true potential, but we’re stuck with the problem of turning the chickenscratch on the whiteboard into an ‘aye’ or ‘nay’.

[mathiru-net]

What can a Deep Learning system do that other machine learning systems can’t do?

Previous systems, which I guess we could call “shallow learning systems,” were limited in the complexity of the functions they could compute. So if you want a shallow learning algorithm like a “linear classifier” to recognize images, you will need to feed it with a suitable “vector of features” extracted from the image. But designing a feature extractor “by hand” is very difficult and time consuming.

An alternative is to use a more flexible classifier, such as a “support vector machine” or a two-layer neural network fed directly with the pixels of the image. The problem is that it’s not going to be able to recognize objects to any degree of accuracy, unless you make it so gigantically big that it becomes impractical.

[mathiru-net]

CSS flexbox layout can be applied to SVG in order to simplify the task of constructing charts with D3. This approach has been made possible by the JavaScript flexbox implementation that Facebook recently open sourced to support ReactJS Native.

[mathiru-net]

Making emotive games from open data

Microsoft researcher Kati London's aim is "to try to get people to think of data in terms of personalities, relationships and emotions" (...) In one experience she invited children to play against the real-time flow of London traffic through an online game called the Code of Everand. The aim was to test the road safety knowledge of children between the ages of 9-11 and "make alertness something that kids valued." The core mechanic of the game was that of a normal world populated by little people, containing spirit channels that only kids could see and go through. Within these spirit channels, everything from lorries and cars from the streets became monsters. The children had to assess what kind of dangers the monsters posed and use their tools to dispel them.