My “to read” backlog has grown quite large, as measured by the huge number of open tabs in my browser. Here are a few that I can close now:
- Why There Was No New Hardware At Google I/O
RedMonk’s Stephen O’Grady posits that Google’s real play (pun mine is services, therefore it only needs hardware, and Android, to be “good enough” to provide a sizeable channel for these services.
- Black box software: a problem for science that extends to big data
Scientists place a high degree of trust in their analytical software without understanding how it works or validating the results against other methods. As big data grows in importance, this will be an issue that concerns business as well. An argument for open source.
I am reminded of another recent article, perhaps still open in my tabs, that talked about erroneous models playing a role in the financial trouble that befell of one of the big banks. The rather insidious scenario that played out at the bank was that models that produced errors requiring the bank to be conservative in its liquidity posture (and thus limit profits) were quickly identified, while those that allowed the bank to take a more aggressive position were not questioned.
List of Cloud & Vendor Events for 2013
A pretty thorough list.
Behind “Amazon Redshift is 10x faster and cheaper than Hadoop + Hive” slides
Interesting discussion of Hapyrus’ recent efforts to compare analytics performance on Redshift vs Hadoop + Hive. Redshift offers significant cost and performance advantages for some use cases, particularly where data is of average size (a few TBs) and queries must be performed frequently (a few per hour or more).