over 10 years on 2008-05-01


quantum computing & energy crisis:  the best thing we could do as a society regarding the energy crisis is invest in quantum computing - solving the traveling salesman problem would definitely save more fuel than almost any other optimization i can think of.  on the topic, quantum computing will never get declassified (even if it happens in the next 50 years) because it would instantly break basically all encryption and destroy historical 'security' (making privacy the only valid concept :) - via drop.io's paintball outing.

monocultures are extremely risky:  in biology, in food supply, in politics, in information, etc.  diversity is key, but there is a tendency towards centralization.  our civilization is more and more at risk as information is centrally pooled -- especially because security is a war of attrition, and the more value is centralized the more likely it is to be attacked (the more people are willing to spend to get it) - we are tending towards a monoculture on a dizzying number of levels.

the famous chinese 'firewall' probably most significant as cyber war defense:   as has been demonstrated in the worlds two cyber wars to date (georgia and estonia) turning off the outside world may be the only effective defense (essentially raising the drawbridge under attack) - is the chinese firewall actually more about being able to turn off the international internet when under attack than preventing their citizens to tunnel out?  that would be brilliant on their part, the us doesn't have a drawbridge.

what are the largest stores of social graph information in the world?  game is this:  pretend no regulation and infinite free compute power (aka you can get anything and you can blow through any 'dirty data' problems) - what are, in rank order, the largest real world existing stores of social graph information? or, put differently, what are the richest real consolidated human focused relational databases in the world?  credit cards, cell phone bill, internet browsing activities (isp data), social network data, im buddy lists and messaging activity, email address books and messaging activity, evite... whatever it is - social networking data currently ranks low, it is just the best organized data right now (cheapest to use) - and this will change.  if you get wider what are the theoretically most valuable information streams in the world (don't worry about the scale/centralization of the data)?  with the same rules real time purchasing data is probably the most valuable followed by real time location data.  what would real-time thought data look like?

both oil and data benefit from technology costs - but oil gets scarcer and data gets more abundant:  the enormous stores of social data that exist locked up in silos at credit card companies, email systems, etc have some of the same characteristics as the canadian oils sands, with one fundamental difference.  just like the oil sands, many companies are sitting on enormous stores of valuable social data which they could harvest for insight, trading, etc but given the form of their data makes it extremely expensive to extract the information (after all, these data structures were built in quaint times).  moore's law helps out here, because over time credit card companies, email providers, anyone who has useful data can expect it to become very cheap to harvest.... just as oil technology helped out with the oil sands.  unlike the scarce oil commodity information will get more and more abundant over time.  so, even as the costs of data extraction fall so too will the value of data extraction...  it means that companies should be going full tilt to harvest the data they have while it is still relatively scarce.  almost all of the normal social, purchase, location, etc data is going to get extremely abundant and therefore extremely worthless very very fast.

twitter is amazing:  as a format of pure signal - the 140 character limit is irrelevant, using urls, etc you can pack an amazing amount of signal into 140 characters...  it is going to become the backbone of the  next iteration of the web - i am not being a fan boy, i can back this up (but will save it for a much longer post).

original swl blogposts and letters 2007-2010