AutoML - or "Deep Learning for Dummies!"

I'm always intrigued by all the myriad "{Subject} for Dummies!" books, because they allude to the idea that mere mortals can accomplish most any complicated task given some basic training. Now, Google is leveraging that approach with AutoML, a new framework for putting Deep Learning expertise into the hands of everyday engineers and programmers. Stand back and Pay no Attention To That Man Behind the Curtain! 😮

Cory Doctorow explains why FLOSS matters

A worthwhile video presentation by Cory Doctorow about why Free/Libre and Open Source Software (FLOSS) matters, "How you got here". Thanks to O'Reilly for making this OSCON presentation available online.

Shades of HAL-9000 (maybe Clarke got it right!)

Here's an interesting article written by Will Knight, the Senior Editor for AI, MIT Technology Review. Regardless of your point of view on the philosophical underpinnings concerning Deep Learning and Cognitive AI, it's a worthwhile read (Caveat: It's a long article, but you can pick up the salient points pretty quickly).

The Big Three - And Self-aware AI isn't one of them!

An interesting read from Bill Gates - Where he states, "I view the threat of deadly pandemics right up there with nuclear war and climate change. Innovation, cooperation, and careful planning can dramatically mitigate the risks presented by each of these threats." Interesting that he doesn't include Self-aware AI in the top three of his list of worries for the future.

Don't get upset about CPython concurrency - just get Grumpy!

Here's a short recommended read that was brought to my attention by a teammate. If you've grumbled about poor concurrency support in conventional Python, Grumpy looks like it's coming to the rescue.

I suspect this break-through reflects the benefits of new concurrency approaches, as reflected in recent packages (like Akka) and languages (Go).

It's a fundamental shift, since Go extensions will be supported over C extensions. I assume that means Grumpy and CPython will co-exist for some time, unless most existing C extensions get re-coded in Go? It will be interesting to see how this unfolds!

Installing Robot Framework on a MacBook

I recently faced challenges getting the open source Robot Framework to work on my MacBook, which is running OS X El Capitan. In general, the installation process should be trivial, but here's what finally worked:

1. If you have the brew package manager installed, update it:
   
       brew update
   
2. Use brew to install (or upgrade) the chromedriver program:
   
     brew upgrade chromedriver
   
3. Make and use a Python virtualenv for the framework installation:
   
     pip install --upgrade virtualenv
     source virtualenvwrapper.sh  # Note: Depends on where virtualenvwrapper.sh was installed to.
     mkvirtualenv robotframework
     workon robotframework
     cdvirtualenv
4. Install the necessary framework packages:
   
     pip install --upgrade robotframework
     pip install --upgrade robotframework-selenium2library
     pip install --upgrade requests

Now, you'll be able to use the robot command to run test scripts OK :-)