Insight. Applied.
Trust seems to be a rare commodity these days.
Game theory helps us understand how distrust grows within a society and—encouragingly—suggest ways in which we might stop the rot.
Nicky Case has brought this dry subject matter to life via a wonderful interactive description of how (dis)trust evolves.
It’s about 30 minutes long. Just play it.
Using Excel in business critical operations is risky to the point of recklessness.
Phocas have just published an article highlighting seven problems with spreadsheets. The problems are
See the original article for details.
Microsoft has said that the recent wave of cyber attacks that has hit 150 countries should be treated by governments as a “wake-up” call.
Security experts have constantly warned intelligence agencies about the risks of weakening security to make it easier to undertake surveillance. As UK hospitals are plunged into chaos we can see the ramifications of ignoring these warnings.
This has been a textbook demonstration of why weakening infrastructure that we all rely on is a terrible idea. Let’s hope that companies who’ve resisted pressure to insert “back doors” in their systems feel emboldened to continue the fight.
I have a lot of sympathy for the view expressed in the following tweet
Good CS expert says: Most firms that thinks they want advanced AI/ML really just need linear regression on cleaned-up data.
— robin hanson (@robinhanson) November 28, 2016
Many organizations who dive into machine learning haven’t even started to extract value from their data.
I understand why they want to get on the train. If you’ve not managed to draw value from your data yet why not just shovel it all into these amazing algorithms and let insights flow out the other end. Jump to the latest technology. Makes sense.
Unfortunately that’s not how data science works. The success of any data science project depends on you understanding your business and your data.
Some things you might do before adopting machine learning
If you’re ready for it, machine learning is fantastic. However, if you deploy it as a magic bullet, don’t be surprised when you shoot yourself in the head.
People, in my experience, tend to find it hard to get their head around many “big data” concepts. It’s only when they attempt to implement initiatives, and are frustrated by the basics, that they start to “get it”.
One of the most basic things that people seem to misunderstand is the challenge of moving data around. Most big data tutorials assume that you already have terabytes (or petabytes) of data on your cluster. However, how does it get there in the first place? If you have hundreds of terabytes of data in traditional storage how do you get in onto your Spark (for instance) cluster?
There’s no magic answer. Moving that amount of data is painful—plain and simple.
Microsoft has recognized this and introduced an Azure Import/Export service. Basically you snail-mail them your hard disk(s) and they upload the data to distributed storage via their high-speed secure internal network.
It’s a start, but is trailing Amazon’s solution to this problem by, oh, around 100 petabytes.