This morning I had the opportunity to read Steven Levy’s Wired article profiling “How Users took over Twitter“. It was the fire set under me to write down some thoughts about a possible future for Twitter and how it will fit into a future landscape. I should preface this post with the warning that it may steer further into ’speculative fiction’ than business strategy but I haven’t met anyone who’s created a 10-year plan without a little speculation.

The thesis here is that Google is missing out on a major long-term opportunity by not aggressively pursuing Twitter in its quest for more acquisitions. We’ve all seen rumours about talks but the trend seems to be taking Twitter away from Google rather than towards it. I’d like to put forward a hypothesis of what could happen if Google acquired Twitter tomorrow. Although I agree that their current valuation is astronomically high for a service with less users than YouTube gets in a day, I hope this post might explain how that value could rise over time if leveraged properly.

Day 1
Google acquires Twitter for one kajillion dollars. I revisit that moment when I watched Chad and Steve announce that they were now billionaires. I am violently sick with envy but recover with a steady diet of YouTube clips and tweets – all free of charge.

Day 60
Google adds real-time search to their results. Nothing really changes but Bing pouts a little. We now see fads rise and fall within minutes.

Day 150

Twitter leverages a viable business model for business users to issue real-time promotions and acquire premium ‘validated’ accounts. This is a success only due to the modern world’s obsession with SEO for Google. It quickly becomes integrated with Google Adwords platform.

Day 300

Twitter and Google bring shortened URLs in-house, collapsing the tiny URL industry. The additional data collected by linking a tweet and its destination URL serves to enhance Google’s evolving search algorithm by telling them what users are clicking on RIGHT NOW. Timeliness is one thing spiders have trouble with, until now. It also integrates well with Google Analytics to show ROI on specific tweets for motivating clicks.

Day 1500

Twitter realizes that they have amassed trillions of tweets over the years and most of them seem to be about nothing. Most companies would then collapse under the data-storage demands and evaporating venture capital, but since Twitter is now supported by Google’s ad sales it notices something interesting instead.

In 1984, the Douglas Lenat began a project called CYC with a fascinating premise.  The divide between artificial intelligence and human intelligence was not in developing ever more complex rule sets but in learning ‘common sense’ – discrete facts that most humans dismiss as obvious but a machine needs to make logical leaps that people do every day. Typical pieces of knowledge represented in the CYC database are “Every tree is a plant” and “Plants die eventually”. The concept is that real intelligence requires this massive database of connective ideas that lead to emergent pattern recognition. The trouble at the moment is that CYC requires an arcane language to input discrete thoughts and collects data like (#$capitalCity #$France #$Paris). The project has been active for 25 years and spawned applications like OpenCYC for creating expert systems already, and the link to Wikipedia’s library will no doubt be a massive leap for the project. Wikipedia tells me that Lenat’s early estimates expected 350 man-years would be needed to create a useful version of CYC, and more recent estimates have put the number higher.

Hang on a minute! Google’s been working for years on parsing natural language via their search engine and GOOG-411! And Twitter is filled with all sorts of previously-considered inane crap like “Monday again. Ugh. Need coffee to face another work week.”! On this day 1500 of the new Google-Twitter hybrid, the project begins which I am code-naming “The Mechanical Twit”, or TMT for short (may I suggest it be pronounced TOMATO?).

TMT takes that tweet and parses it into the following facts:

1. Monday is a repeating thing
2. People dislike Monday
3. Monday is the start of a work week
4. Coffee is something people ‘need’, especially on Monday and especially in the morning (when the tweet was posted)

TMT is then able to extrapolate in many directions such as ‘people dislike work’ or ‘morning arrives on the East Coast at noon GMT’.  Each individual tweet can be so insignificant, yet combine to create a completely crowdsourced hive mind.  And we all know this is Google’s master plan anyway now that they’ve got our email communications, calendars, interests and location.

The reason tweets are so good for this is their 140 character length. You really can’t give more than one discrete complex thought in a tweet which is virtually identical to CYC’s required input. The users are already semantically tagging their ideas for crying out loud!

Day 3500
TMT launches, in beta of course. Ray Kurzweil, author of ‘The Age of Spiritual Machines‘, does an exuberant pirouette in his beautiful new robotic body. Verizon has launched their new ‘Borg‘ video-phone as an invisible implant in your jawbone and retina. The killer app in the new Android 4.0 is good ol’ TMT. Here’s how it works:

TMT knows everything that’s happening right now and everything that’s ever happened in the world. It doesn’t store it as data – it KNOWS it through kajillions of neural network links that are the exact reverse of Keanu Reeves ’seeing’ the Matrix. TMT knows that pizza is still good when it’s cold but that hamburgers are not. TMT knows that it will rain in New York today and that means people will want umbrellas and that the best place to buy an umbrella is in the drugstore.

But TMT will also know everything about you, starting today. It will parse your tweets as you tell your friends that Terminator:Extinction jumped the shark. It will know that you are in Stitch Media’s Halifax office today and it will know why you are there, since it’s right there in your Google Calendar.

And before you know it, TMT will become your best friend. Good ol’ Tomato will pipe up with ideas like ‘Hey I just heard about this great concert that I’m sure you’re going to like in Boston – I found a cheap flight there and I see that OUR buddy Chris wants to go too. Want me to set it up?’ or it’ll answer you when you say aloud ‘What’s that girl’s phone number who sat next to me in ‘Avoiding Robot Slavery 101′ class six years ago? She gave it to me and I’d really like to call her’. (of course, there won’t be phone numbers in the future though, stupid)

Many people might think this would be unbelievably annoying. But that’s not the way it works – TMT understands what annoys you. It offers a good sushi restaurant and you say ‘Shut up Tomato I hate when you do that!’. It will learn to stop offering. When you ask it for that girl’s phone number and TMT knows that she’s married now and really wasn’t that into to you anyway – he’ll tell you, kindly and firmly.

Can anyone put a price on a best friend? Would a kajillion dollars come close?

---

This entry was posted on Thursday, October 29th, 2009 at 9:08 am and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.