It was ten-thirty at night, and Peter and Todd were staring at yet another screen with yet another presentation on it. Except this was the one that mattered. They had been given all the rope the VCs would allow, and now they would be slowly and gently pulled back in. They were utilizing a program called Director, used by just about every high-concept start-up in the industry. The reason for this was simple. With Director, one could emulate the software or service desired, without actually having to build the infernal thing. Peter and Todd were going to use all its powerful features to show their high-concept software without ever programming a single line of it. Although this was well and good, it also had to be close enough to the ultimate product so as to not hang them by their own petard. Thus a fine line was being drawn at the moment.
Peter had a scotch sitting next to him. In moments of extreme stress and concentration, he sometimes found it helpful to lube the gears. About half a bottle of lubing had occurred so far. They were trying to carefully craft a story that had substance at its heart, but style all around it. The core of the company was a very valuable piece of technology, but selling it by itself would be practically impossible. And the market was unproven enough to make raising money on that alone laughable. They needed a Trojan Horse. They needed something to smuggle their plumbing into the offices of venture capitalists up and down Sand Hill, and make them get so excited that they opened up the gates to the city and invited them in. This was where Snow Crash came in. Although Neal Stephenson's book had been published way back in 1992 (practically a century in Internet years), it held sway over the Internet crowd like no other piece of literature. This was due not only to the technological view of cyberspace, avatars and 3D worlds, but also to its egalitarian philosophy that everyone would be able to join in this society of equals. Valley types liked this utopian view of the world, and the VCs liked it because it meant that the potential audience was, well, everyone.
But everyone knew that Snow Crash was for all intents and purposes impossible to implement. So Peter and Todd had a bit of convincing to do. This is where things got sticky. They knew they were pushing the envelope. They just couldn't burst the damn thing. So they had to build the vision slowly. Couch it in something easily digestible and then work their way up. Their packets were going to be smuggled under Snow Crash, and housed in an on-line service. A next generation on-line service that would rival and then surpass anything else currently running. This was something that would generate customers, revenue, advertisers, and, oh yes, scads of hype. Just what the doctor ordered.
Now the problem. There was already a big 800-pound gorilla in that particular forest. The gorilla was America Online. The world's biggest on-line service. Ten million users and growing every day. But Todd and Peter knew where the chink in the armor was, and they planned to drive themselves and their little company right through it. The Achilles Heel was the fact that AOL was a kludge. Through years of updates, modifications and changes, AOL had become in essence the greatest legacy system of all time. Spaghetti Program Maximus. A tired old dog that was kept alive by the technology wizards back in Virginia. And this was always assumed to be okay, because who else could spend the money to build a nationwide network to catch up with them?
Could you believe the government? That’s right, good old Uncle Sam had gone and built one. Originally called the DARPANET for the initial funding from the Department of Defense, it had been built biologically by a strange ragtag collection of government agencies and universities. A national network, just waiting for a couple of lunatics to start sending Snowcrash over it. And in a final piece of irony, this network, now known by its more current moniker of the Internet, was a hard piece of technology for AOL to get along with. In fact, there were all sorts of compatibility problems between AOL and the Internet, which meant that Peter and Todd had an instant network and the added benefit of one that didn’t play nice with AOL. And had more people using it than AOL. And was growing faster than AOL.
Todd and Peter were going to use the Internet as the backbone for their virtual on-line service, and in doing so throw AOL a curveball that it had no chance of hitting. And this was going to get potential investors very excited. “You know,” said Todd, staring at the screen. “This actually has a chance to fly. I mean, if we can show that the computer market is moving into the broad consumer space, I mean like VCRs and TVs, then the interface for services like on-line companies will have to be as easily accessed as any old appliance. And if we can prove that a purely graphical interface is the method to get into the sixty percent of this country that doesn’t yet have a PC, then we have something.”
Peter was playing with the slide in front of him. “That’s the angle I’ve been hunting for. This is the on-line service for the rest of us. By using a 3D graphical interface, we are presenting people with a view of the world that they already understand. No windows, icon interface, drop down menus or any of that silliness. To us in the valley, those all make sense, but it’s complete gibberish to the rest of the country. We are going to fix this. This is about putting a human face on technology.” A grin slowly came to life on his face. “ I think we have our hook.”
“We have the excitement. Now any idea how much money it is going to take to build this funhouse?”
“That is the question,” said Peter. “And like everything else in life, it is all about timing. If we think we can build this thing fast enough to generate some serious heat, then we raise only enough money to survive, and then hope for a major pop in the valuation when we got out for the serious money. All told, I think we are probably going to have to raise in the twenty to thirty million range before this is all over. So if we don’t get our valuation up fast, we are going to get diluted back to the stone age.”
“Peter, it seems maybe we should get this bird off the ground before we start worrying about whether we have 19 percent or 18.5 percent of a publicly traded company. I mean, building a company instead of a bank account is the point of all this, right?”
Peter glared at Todd. Maybe it was the time, maybe it was the scotch. “Todd, I am not on some ‘making the world better for democracy’ jag here. We did that last time and all we got for our troubles was some kiss off money and a raft of aggravation from Lazarus. This time we drive this car all the way to the finish line, and we hold onto as much as possible. I want this to be the swan song Todd. Next stop retirement and a beach house in the Bahamas.”
Todd didn’t like the direction this was heading, so he decided to change tracks. "OK, so let’s say we are going for the gold. How much do we need to raise in this round to get to the bonus round and still have most of the company available?”
“This all depends on the minimum we need to start getting partners. We have to show that other companies will come onboard with the technologies we are introducing. If it is just us versus AOL, no VC on the planet will touch us. So it’s all about the partnerships that we can forge. Ray Noorda may have taken Novell headfirst into a brick wall, but he was right about one thing. Coopetition is the name of the game. This isn’t a zero sum game. We are going to have to give a little on the content side to get our products leveraged into the market.”
“So what is the minimum functionality we can show to close some big time partners?”
Peter looked at the slides in front of him. “Well, obviously all the back-end necessities like authentication, billing, security and the like are not necessary. And we can probably get away with just running the demo over our own internal network. The rendering engine has to be up to snuff, and we have to have enough distributed firepower to show at least five avatars on at least three machines.”
Todd leaned forward. “That raises an interesting question on the system architecture. Are all avatars human-controlled, or are we going to have machine-controlled avatars wandering around also. Bots, if you will.”
Peter shook his head. “I am against computer-controlled characters, because it confuses the user unnecessarily. Is the avatar they’re approaching human-controlled or computer? It breaks down the experience.”
“True, but we also then could run into an empty world syndrome at the beginning, where there aren’t enough users to populate the worlds. It will be like being the only person in a town. I think we really need some bots just to make the worlds seem inhabited. Don’t want the user playing an endless game of find-the-avatar, do we?”
“I still don’t like it. It sets up expectations that are not realistic for the user. Am I talking to a bot or is this person ignoring me because they find me the most boring person in the world?”
Todd sighed. He had seen this approach before, where Peter slowly dug in his heels until he was buried up to his waist and no longer in a mood to compromise. This was not a make or break decision in the grand scheme of things, so he decided it was time to blow the canopy and bail out. "Okay, let's table this one for now. This is will become obvious one way or the other in user tests anyway."
Peter grimly nodded. "So we have two machines, five avatars, rendering engine working, network protocols working, network based for now, on standard consumer PCs. That should be enough for any investor. However, this raises the next question. What exactly is a standard machine?"
This was yet another decision that could instantly turn them into a nuclear waste dump in front of any investor. Their software was obviously pushing the edge of the performance envelope. Assuming a more powerful machine would mitigate several problems, but at the cost of reducing their addressable installed base. This would make their universe of happily paying customers grow smaller, which would inversely make the grimaces on the faces of their investors get larger. So a wonderful dance called the Minimum System Requirements Tango was about to commence. This involved a little crystal ball gazing, trying to figure out what the standard machine would be in the eight to twelve months it took to get the product out. Too high and they would eliminate all but the bleeding edge techno-adopters of the Bay Area; too low and they wouldn't have the power to bring their virtual world to life. This was compounded by the additional fact that their software also was sensitive to additional components such as modems, line connectivity, sound cards and graphics. All these components, with all their different configurations and manufacturers, had to work flawlessly with the software.
Todd pondered the question in front of him. "Well, the problem I see it is that although the majority of PCs that are going to ship this summers are Pentium IIs, the majority of the current installed base is Pentium I. That probably isn't going to change, and especially since this is a consumer product. I think we have to operate in a 32 megabytes of memory, because we can't count on 64 megs, we can't ask people to buy more, and asking them to close all other programs before they run us is also not very user friendly. I don't think we should count on anything above a 56K modem speed, and let's default on the sound to the standard Soundblaster-compatible sound card."
Peter thought about this for a moment. "Are we being too conservative? I mean, we don't really need to worry about this becoming a problem for another twelve months. I would hate to have us launch and discover we are behind the curve and not offering enough for success."
Todd decided to play safe here. “Well, we are going to get a little bit of a benefit from using our own proprietary rendering engine, but I really think that making a choice to go to Pentium II as this point is a little premature. There is also the perceptional problem of our current investors and partners thinking we are pushing the hardware platform too hard. It seems to be a silly fight to be getting into with people at this early junction. Besides, if we are off on this, we can always rev up in the first post release. We both know that the launch software is going to be a fraction of our total functionality package anyway."
"Actually, that's something else that has been bothering me. I agree that we are not going to be able to launch initially with all the functionality that we want. The question is what do we think is critical out of the gate that absolutely has to be in the first release. It also begs the question of what exactly constitutes a launch. Do we do the official alpha-beta-launch routine where we have official little steps to the rollout process, or do we start softly releasing code to the outside world during the beta program and try to roll quietly into a full launch. That would allow us to sneak in functionality and get a bigger bang when we actually launch."
"It would also help with the empty-world problem. If we can soft launch with all our beta customers, then the first paying customers, partners and reviewers will be able to experience a world that is in essence already up and running."
Peter turned to the forgotten presentation. "This is good. We are starting to get an outline of a schedule, which will lead is to what we launch when, which will lead us to what everything is going to cost. Then we can do the Director presentation that will exactly lay out our battle plan and what we need in capitalization."
Peter hung his head in contemplation for a moment, and then started to type as quickly as he could, numbers and words spilling out onto the screen in a sudden torrent.
Todd looked over his shoulders as slide after slide appeared. After twenty minutes Peter had the outline done. As he finished the last slide, he turned around and grinned back over his shoulder. Todd looked at the final slide, outlining the next investment round.
"Jesus, that is a big number."