“We couldn’t start inkjet again if we had it in our hands because we’re not meeting the rules.” That stark admission from former HP Fellow Will Allen reveals why even the most innovative companies struggle to recreate their own breakthroughs, and what it takes to actually scale disruptive technology.
In this episode, Will Allen, holder of 102 US patents and the first HP Fellow promoted within HP’s Global Print Business, takes us inside three decades of Silicon Valley innovation from logic analyzers to consumer inkjet printing to his current role as CTO at Kaspix, where he’s pioneering ultra-low-power AI inference using analog circuits. We explore why research should be treated as investment portfolio management (not an expense to cut), how “showing beats telling” when getting buy-in for radical ideas, and why getting AI to the “true edge” – directly at sensors and actuators – will fundamentally change computing economics.
What You’ll Discover:
- [00:00] Why Research and Development Are Two Different Things → The fatal mistake of treating R&D as a single expense line when research is actually an investment with portfolio-level returns
- [06:03] Has Silicon Valley Run Out of Ideas? → Why scaling success creates the very constraints that prevent future innovation, and whether we’re less innovative than decades past
- [10:12] The Scaling Trap That Kills Success → Real HP story: how field-fixing problems scaled so badly that engineers couldn’t design problems out, and what this means for any growing business
- [16:23] Getting Past the “$100 Million Question” → How to navigate corporate demand for predictable returns when developing something the market has never seen before
- [18:03] “A Functioning Proto Is Worth a Thousand Pictures” → The clownfish story: how a weekend demo got low-drop-volume printing approved after months of rejection, and the art of communicating on stakeholders’ terms
- [21:16] Signal Spotting and Fundamental vs Killer Apps → Will’s framework for distinguishing noise from transformational trends—and why asking “what’s the killer app?” might be the wrong question
- [24:47] Kaspix and the True Edge → Why analog circuits for AI inference could be as transformative as the mouse, enabling intelligence directly at transducers without memory-compute bottlenecks
- [29:56] Where AI Is Actually Heading → Beyond the hype: specialized AIs, “AI middle management,” and why rapid societal change from deterministic technology creates uncomfortable transitions
- [36:04] The Advice Will Would Give His Younger Self → Why leaders who invested years in education suddenly think quarterly, and how to reclaim the long-term thinking that got you there
Key Takeaways:
- Treat research like an investment portfolio with measurable returns, not a cost center to cut during downturns
- Success at scale creates antibodies to future innovation—be intentional about preserving innovation capability
- “Show don’t tell” with functioning prototypes that communicate value on stakeholders’ terms, not yours
- Fundamental technologies solve problems across multiple domains; if it only has one killer app, question the scale potential
- The “true edge” isn’t your smartwatch—it’s at the sensor/actuator level where intelligence eliminates downstream processing
- Quarterly thinking kills long-term breakthroughs; reclaim the investment mindset that built your career
About Will Allen:
Will Allen is CTO at Kaspix, pioneering ultra-low-power AI inference through analog circuit design. Previously, he spent 30 years at HP, becoming the first HP Fellow promoted within HP’s Global Print Business. He designed the color imaging pipeline used in HP’s first 4 million color consumer inkjet printers, led IP production in HP Labs’ AI and Emerging Compute Lab, and holds 102 issued US patents across printing, displays, robotics, and digital imaging.
Related Links:
Spotify
Apple Podcast
Pocket Casts