Yep, I started another company. Couldn’t help it.

I have a passion for media, and I was getting way too anxious watching the TV/broadband world change without being a part of it. Coincident with that, the commoditization of media creation tools is incredibly exciting, and of particular interest to me is the impact that is having on teens.

The amount of content being created by teens is astonishing; a significant amount of it is actually good, and some is even great. This isn’t dad’s home movies anymore. Today’s youth have grown up consuming media at a rate beyond what we could even have imagined when we were kids. The availability of multiple forms of media on multiple devices and honed multitasking skills give modern American teens the ability to expose themselves to over six hours a day of media during their leisure time. (And they accomplish it in around four hours.)

From these hours of eyepounding exposure, teens are developing styles and techniques for storytelling that are uniquely their own. The content that they create is imaginative, funny, disturbing, insightful, and deeply personal.

At Uth TV, we’re finding the best of the young media creators and giving them a highly visible outlet for their work.

Rendering complex computer animation can easily demand thousands of hours of compute time, a task generally relegated to racks and racks of submissive CPUs fondly referred to as Render Farms. In days of yore, the cost conscious production company would triple shift every computer in the facility, creating an overnight render farm out of people’s desktop computers. Interns, aka Render Wranglers (seriously – go check those old film credits, you’ll see…), would seek their big break in the film industry babysitting the machines each night. Generally the job involved walking around looking at screens to make sure everything was going well, and fixing things when they weren’t. A control-G beep would generally signal a frame had finished rendering and a new one was queuing up. It was satisfying to hear the occasional beep to know that progress was being made, but there were no modern dashboards showing all the machines at once, and no proactive notification of problems.

As technology evolved and better sound became an option, one of our competitors (and I’m ashamed I can’t remember who) made a brilliant move. Each time a frame rendered successfully, the machine would let out a contented “Mooooo.” If ever the program ran into an error, an authentic cowbell sound would clatter. With dozens of computers in a room, the wranglers were treated to the pastoral sound of happy cows until attention was requested.

I’ve always loved this idea. The use of two unobtrusive and complementary sounds to signal “good” and “bad” allows people in the area to be aware of positive progress as well as to be alerted of problems when necessary. The mooing exists as a comforting background noise, which can be consciously ignored. However, at any point in time one intuitively knows how much work is being done by the general “feeling” of the room.

Anheuser-Busch has decided not to run a “controversial” ad during the Super Bowl, but instead to launch it just on their web site. Personally, I think they’re really wimping out, as the commercial, called “Wardrobe Malfunction,” is as tame as they get. But I love their quote in USA Today:

“Our beer drinkers tell us we shouldn’t overreact to the media scrutiny, but we have to live in this world.”

I think every corporate PR missive should be required to start with, “Our beer drinkers tell us…”

Regarding our simplified version of Moore’s Law (see prior post), in the mid-90′s I built a spreadsheet to see what had actually happened. At that point we had about 15 years of technology investing behind us which I charted out. The result was that there was about a 1.8x price performance improvement every year. This was not a measurement of raw compute speed per dollar, but rather the average cost of what we were spending to support an animator – so it was really an amalgam of compute power, memory costs and storage costs, and a very realistic indicator of what we could expect to be spending.

I’m a big fan of Andy Kessler. His articles are always worth reading, providing insight to parts of the financial world I don’t play in, and giving his unique spin to it all. His most recent emailing included an excerpt from his new book, Running Money. In this section, he recalls meeting an entrepreneur who was pitching a new keyboard product, I cracked up at the following part:

“Here is photo of our chip, pretty small, doesn’t cost us more than $5 in volume. TSMC in Taiwan makes it for us. Over here is the harmonizer, the pitch control, tempo stabilizer. Here is a DSP to do Fourier transforms. There’s not much else to it.”

“What is that section at the bottom corner, labeled 780?”

“Oh that. Well I grew up programming DEC machines, so when I needed something to control the whole process, I just sat down and threw in a VAX.”

“An entire VAX?”

“Sure, it’s no big deal, just a little 1 MIPS VAX 11/780.”

“You’re kidding. I bought one of those 15 plus years ago for close to a million bucks.”

“Yeah, it’s pretty amazing. It fit right there in the corner, cost me about an extra 25 cents in chip cost, well worth it. Every chip has some 1 MIPS controller on it these days, it was just easier for me to stick in a VAX than to buy someone elses.”

We had a VAX 11/750 at PDI around 1984, and indeed it was worth about a million bucks. It was the machine of choice in the computer graphics world at the time, a time when computers were compared in size to home appliances rather than books. Ours was the size of three washing machines; replacing the far inferior PDP 11/44 refrigerator (don’t stick a magnet on it or you’ll erase the tape!).

Around that time is when we decided our goal was create our own fully animated films. The problem, though, was very succinctly put by someone at NYIT (who’s own VAX was working on a film called “The Works”) when he said, “If we start now, we’ll have the film rendered in 15 years.”

That wasn’t going to work, nor would spending $15 million to get it done in one year. We were smart enough to know Moore’s Law, though we used a greatly simplified version: “The price performance of computing doubles every year.” A simple extrapolation said that in a decade we’d have 1000x the computing power for the same cost. That might start to work.

Because of that we knew we would be technologically limited from doing what we really wanted to do for a decade. Nice to be a little ahead of the curve, huh? But this was incredibly powerful knowledge; we could instead focus our efforts on building up everything else we would need so that when the technology was there we would be ready. That approach set the tone for our steps from broadcast graphics to commercials to film effects to finally getting to do our own films.

It took us 12 years to get our deal with DreamWorks, so we weren’t too far off.