I had a pleasant afternoon coffee yesterday with Guy Lalonde, who has worked in biotech for many years, musing about what makes innovation work, and when does it get stifled -- not at the "Who is CEO?" level, but at the designer level. And it hit us both that often it is TOOLS that provide the insight. A new tool shows up -- say Mathematica -- and all of a sudden, designers have new capability, new insight, and new things happen. When gene splicing tools arrived, the biotech innovation scene exploded, for example. Not at the tools level, but at the applications level. The tools are enablers.
When I was graduating from Caltech, my major professor was Carver Mead, who has become reasonably well known in some circles (he did the work and calculations for 'Moore's Law' for example). I was trying to figure out how to go to grad school, but at 21 with two children, I needed a job. Characteristically blunt, Mead said, "you're not much good at classroom stuff, but great in the lab. And you like new things, being on the edge. You ought to consider Hewlett-Packard. They build instruments, and if you think about it, measuring tools have to be BETTER THAN the state-of-the-art in order to measure accurately things happening AT the state-of-the-art. So you're always on the forefront. And besides, HP will pay for your graduate studies."
How he knew me that well, I don't know. But he did induce me to go to HP, as their 98th engineer, before they passed $100 million in revenue (yes, they're a thousand times bigger today than when I first came to Palo Alto). And indeed, I found that instruments have to be ahead of the state-of-the-art to add value for designers. It's a great challenge.
What does that have to do with today? Our coffee klatch concluded that there aren't enough folk focused on tools in lots of areas -- integrated CAE design, IT shop tools for digital media production, system modeling tools, etc. -- anywhere in fact that the market isn't humongous, the tools wind up being esoteric, one-off designs, with little training or support or ongoing integration with other tools into a useful suite. This is somewhat overdrawn, but it is up to the user in many cases. Think about Autodesk's Maya in use with NVidia's awesome graphics chips, or Blender's 'bone-locking' capability for rigging in 3d avatar animation. Fun stuff, hobbyist stuff, or great animation sound and movies. Depends on your training, skill, and desire. And the specialists go nuts with this stuff.
But what made America wasn't just specialists, it was generalists. The greatest thing about Ford and John Deere might have been their modest (at best) reliability. Every farm boy learned how to fix darn near anything mechanical; 75% of HP's first four hundred engineers came off farms for example.
But tools today are taught mostly to specialists. So animation tools aren't found at the Moore Foundation or the Packard Foundation, and consequently, no one there can design animation movies to attract donors or to illustrate the fabulous work being done. Traditional companies, and most foundations, have IT shops focused on Microsoft Office or Google Docs, and Cisco routers. Geospatial decision support systems, large dataset visualization packages, and even Matlab are next to impossible to find; the business analytics (BA) software packages are rare enough (and obscure enough) that their utilization is modest.
The point, we concluded, is that the tools enable creativity, but the tools are taught and made available by domain and specialty, not so much in a generic or hierarchical manner. Too bad.
And that brought me back to HP. HP spun its measurement tools off into Agilent, a company which has spun off more than half of its own legacy. And HP builds (or vends) mostly high-volume goop, like printers and ink, or plain vanilla PCs, and maybe someday smart phones. At least that is what the world perceives. And HP's Enterprise Group builds backbone computing systems and infrastructure, with a touch of BA via the (asinine) purchase of Autonomy. Couldn't some of these apps be rolled up into some superset of true 'measurement' capability -- of system dynamics and processes, with special tools that are generalized for (somewhat) wider appeal?
I think of the HP-80 and eventually the HP 12C, as the epitome of taking a technology first done for scientists (the fabled HP 35) and making it truly useful for real estate agents and bankers, groups you'd have never thought of for giving them an easy capability for semi-logarithmic calculations and plots. And no business computer company ever made that connection. Could that be done again? Every art department, every marketing group, and every sales team needs attractive, compelling sales documents -- couldn't the Autodesk Maya and Nvidia chips lend huge value here if rejiggered to serve a substantially wider market? Dream on....
What might be done with these trenchant observations, made on a very sunny February afternoon at Starbucks in Menlo Park? Best to go grab a bottle of wine for dinner...