I'm in Computational Biology, and I'd say that the most valuable skills you should learn (and the ones most often seen in this field) are more mathematical and/or statistical than "big data." Understanding how to properly normalize your data or calculate a p-value will take you much further than being able to spin up a 100-node Hadoop instance in most labs.

I think you should spend the first year on your home PC. Download RStudio and work through a few R Tutorials, then find some data/questions that interest you and poke around. Post your results to a blog so that you'll have something to show for the time you spend and release the code on GitHub so that it's open to future employers.

I'd say get comfortable with a data analysis language (R will probably serve you best currently), and a data manipulation language (Python, Perl, etc.) and start asking questions of data that's around you (your email archives, a log of Internet sites you visit, your spending records, etc.). Once you've found that well-designed algorithms can't handle some dataset you're looking at, then look at Hadoop and other "big-data" projects.

When you're ready, I'd steer your towards Next-Generation Sequencing data. Most of the bioinformatics questions being asked (and funded) now have at least some interaction with NGS, and analysts capable of working with that data are highly valuable. Check out the 1,000 Genomes project when you're ready to start playing with free Sequencing data.

I think most people pick these skills up at work -- seeing these practices employed by others. I guess the short answer would be: go work for a company which is already employing people who understand these topics and learn from them for a couple of years.

Of course, that doesn't help you if you're not looking to change careers...

Interesting. I view this from a completely different perspective: if DNA sequencing really is outpacing Moore's Law, that just means that the results become disposable. You use them for your initial analysis and store whatever summarized results you want from this sequence, then delete the original data.

If you need the raw data again, you can just resequence the sample.

The only problem with this approach, of course, is that samples are consumable; eventually there wouldn't be any more material left to sequence. So this wouldn't be appropriate in every situation.

The incoming data is image-based, so yes, it will be huge. Regarding the sequence data: yes; in its most condensed format it could be stored in 750MB. There are a couple of issues that you're overlooking, however:
1. The reads aren't uniform quality -- and methods of analysis that don't consider the quality score of a read are quickly being viewed as antiquated. So each two bit "call" also has a few more bits representing the confidence in that call.
2. This technology is based on redundant reads. In order to get to an acceptable level of quality, you want at least ~20 (+/- 10) reads at each exonic loci.
So that 750MB you mention for a human genome grows by a factor of 20, then by another factor of 2 or 3, depending on how you store the quality scores.

Your suggestion of deduplicating the experiments could work, but definitely not as well as you think because of all the "noise" that's inherent in the above two steps.

If you really just wanted to unique portions of a sample, you could use a SNP array which just reads the samples at specific locations which are known to differ between individuals. Even with the advances in the technology, the cost of sequencing a genome still isn't negligible. For most labs, it's still cheaper to store the original data for reanalysis later.

I often stumble across some product on Wikipedia that I'm interested in buying (album, book, etc.). I actually would find it very convenient if such pages had a "Purchase this Item" link. I'm sure Amazon would kick in a few million for that privilege, or you could use their pre-existing referral program. I think most users would view those links as added value to Wikipedia.

And in light of this, why are we assuming GPS? I can't get find GPS satellites through the metal in my car roof, let alone through my entire car. Is it more likely they they're just tracking the cellular connections?

This thing looks like "futuristic" technologies from the a 1980s movie: picture.

And the FCC ID is the same as the one in a mobile credit card terminal)...

I guess it's comforting to see that, in this instance, the government isn't decades ahead of the rest of us...

Of course! I'll just fix their laptops! Why hadn't I thought of that?

Passengers in cars (, boats, and trains) may object to that one...

Only on /. would this be rated as "Insightful" instead of "Funny"...

I wonder what it would mean to the RIAA (or any IP-based litigation) to have multiple ISP customers consistently NAT'ted to the same IP.

... Maybe this won't be so bad after all!