Archive for November, 2008
People in the Computer
I just finished reading The Turk: The Life and Times of the Famous Eighteenth-Century Chess-Playing Machine, which Dean lent to me to read on the plane. Like lots of history novels, it’s chock-full of interesting facts and tales, and a large amount of it is written with the benefit of years of hindsight (which always makes the writer look much smarter).
For those of you who aren’t familiar with the story, it revolves around a mechanical “automaton” that apparently could play chess. In its 200 year lifespan, it played against (and beat) world chess masters, and chess aficionados, including Napoleon Bonaparte and Benjamin Franklin.
Now there was obviously no computing power around in the late 1700′s to effectively program a machine, so The Turk (named for its oriental robe and turban) had to use some other kind of mechanism. Hidden inside the cabinet was a human chess player, with a second chessboard, who was ‘watching’ the game, through an elaborate mechanism involving magnets and levers. The Turk was, in effect, a conjuring trick.
The thing that made the Turk remarkable, and that led it to be the talk of the courts of Europe in the 1700′s was the notion that it could somehow ‘think’, and react to the moves made by it’s human opponent. The prospect of thinking machines held amazing promise for the future.
Nowadays, computers can play chess, and often extremely well — using a brute-force, compute all the possible moves approach to the chessboard. Now, this is impressive, but it’s not thinking. It’s much more mechanical and rigid than the way a human thinks about the problem. Add to that, the notion that all of these chess-playing rules have to be programmed by an army of human programmers to start with.
So, the Artificial Intelligence of the future is still there — in the future. Tasks requiring repetitive manual labor may have been replaced by robotic machinery, but the knowledge worker isn’t losing their job to thinking computers any time soon.
Even with the amazing advances in software, in search analysis and indexing, computers are really only good for two things — doing math, and remembering stuff. If you need information, you may well be able to find it through your enterprise software. But if you need the analysis and counsel that adds value to that information — maybe we could dare call it ‘knowledge’ — you need to connect with the most sophisticated thinking machine you will find in your organization — another person.
And that’s why social software is so important. It’s a lot like the Turk. Sure, It looks like the computer is helping you out — but really, there’s a guy hidden inside the cabinet…
The Twin Perils of Social Software
The first peril of social software is that nobody uses the application, and you’ve wasted time and money deploying it. The second peril is that everyone uses the application, and you’ve got to spend all your time scaling the system and filtering out all the noise.
Successfully navigating the dangerous waters of social software takes boldness and skill. Expect that you will need to help row the boat. And be prepared to lose a few sailors on the way.
Epic heroes don’t come through these tests unscathed. But they don’t let the wrecks of unsuccessful voyages scare them away, either.
There’s Something About Aardvarks
Looks like there’s a new entrant in the social search space with an aardvark mascot. VentureBeat wrote a glowing article about aardvark.im. They’re taking an instant messaging approach to the problem of getting expert advice. In that regard, they’re more like Qunu than like us.
Mechanical Zoo, the company behind aardvark.im, recently closed $6 million in VC money. Not bad for a company that started at the same time as we did, late 2007. They’re currently in private beta.
Other than the potential for confusion with our own lovable Infovark pal, it’s good news for us. It shows there’s still interest — and funding — available to companies trying to reinvent collaboration tools. We’ve got some product announcements lined up for the coming weeks, and we’re just about to enter our second Alpha test. Watch this space.
Watch “Everything I Know About Startups”
Dharmesh Shah gave a presentation at the Business of Software conference called Everything I Know About Startups. He posted a video of it to his blog, OnStartups. If you’re in the software industry or work with startup companies, you should watch it. The video lasts about an hour, but the insights Dharmesh gives you will last much longer.
Dharmesh lists the ten key points from his presentation — I’ve copied them below — but you need to hear his explanation to understand the rationale behind each.
1. Your idea can suck. Just get started.
2. You can be in the middle of nowhere and still build a great business.
3. Not having cash breeds good behavior. It’s helpful to have constratints.
4. In defense of the modest outcome: You don’t HAVE to build the next Facebook. Modest liquidity events are highly under-rated.
5. “I’m a complete introvert. It’s not that I don’t like people, I just don’t like beind around them a whole lot.”
6. Something’s changed here. You don’t have to spend a lot of money to get your message out there.
7. The real issue with VC is not the cost of capital (which is high), but how hard it is to actually raise it.
8. You have to go through the 12 flaming hoops of venture capital.
9. All the time you should’ve been spending solving your customer’s problem, you use to start to solve the VC’s problem.
10. Write a blog, not a business plan.
We’ll be giving all of these points deep thought over the coming months.
