Your work on developing polysilicon gate electrodes to replace aluminium, while you were at Fairchild, resulted in a giant leap in the manufacturing of integrated circuits. It must have meant you were really in demand at this time. What particularly made you choose to join Intel over everyone else, and likewise what were the reasons that made you leave them years later? Well, back in '68 I created the first commercial self-aligned gate process technology. We knew this would greatly reduce parasitic capacitance and result in faster and smaller devices. Less known is that it also drastically improved reliability since the high temperatures required to remove impurities that could not be tolerated by the aluminium were not a problem for polysilicon. By then Bob Noyce and the rest of them had gone off to start Intel, and I was beginning to feel that Fairchild wasn't going anywhere. Intel had a far more aggressive mindset. So actually it was me who asked them for a job, not the other way around. I left Intel because they were primarily interested in memories. They saw microprocessors as second class citizens. I believed in microprocessors and I wanted to start a company dedicated to the emerging microprocessor market. Gordon Moore, your illustrious contemporary from the 'Intel days' predicted the exponential rise in processing power. As we all know this law is still as accurate today as it ever was. How long will it be before physical constraints dictate that this cannot continue. Or do you think that bio-genetic computing can sustain this still further?Obviously nothing can increase at an exponential rate indefinitely; the big question is, for how long? Currently the state-of-the-art lithography is 90nm of course. Eventually we will be able to reach critical dimensions somewhere between 1 and 10nm (in lab conditions, MOS transistors with 10nm gates have already been demonstrated). So that gives us a factotor of 100 to play with. Then we can put circuits on different layers (this stacking technique is just starting to be used). So let's say we can stack 1000 layers. If the chip area is increased by another factor of ten, then all this would give us a factor of at least 1,000,000 and that could take 30-40 years to reach Beyond that we will need brand new technologies, like nanotubes. But do you think that the demand that has driven this progression can continue, surely at some point the increase in processing power will not justify the R&D needed to get there. For an analogy, the engines in today's cars may be far more advanced than that of a Ford Model T, but they aren't so much different to those from ten years ago. Will there be a stage when things will have to level out?It will take more man hours and require larger sums of money to create each new generation of process technology, and this will mean that smaller companies may have to share their resources. But as long as the volumes needed to drive a new technology exist, there will always be an evolution towards smaller, thus less expensive, and more powerful chips. Some people talk about a major dislocation as we go from 200mm to 300mm wafers, and from 130nm to 90nm nodes, but this is only because we have two big changes at once. We will stay on 300mm wafer size for years now, and so the move toward the next generations of chips will be less traumatic.

...

Let's talk briefly about one of your other offspring. Synaptics' human interface products have recently had some real boosts, following contract wins with Symbian, Samsung and IBM. Can things like fingerprint security for notebooks and laptops become standard in the future or do you think it will be just high end devices that need this sort of functionality?Of all the potential biometric solutions available, fingerprint seems to be the best; after all, you carry your finger around all the time. The system is easy to use, small in size, and as the unit volume grows it will become relatively inexpensive. People want to have security, but they don't like the inconvenience that usually comes with it. Fingerprint recognition is very comfortable and the validation time is the lowest possible. As the frequency of secure transactions increases, these factors will become ever more important. Nothing is certain but you have to make a bet to have a chance of winning in a new market, so we're betting on this.

...

With regard to IDT. The networking companies are still not in a position to carry out major upgrades to legacy infrastructure, and this makes the packet processing technology that companies like it supply less of a necessity. Could the roll-out of wireless access networks be IDT's saviour, or will an increase in SOHO gateway demand (due to telecommuting becoming more common) be the impetus to drive this market, until things pick up in other parts of telecom sector?IDT is involved in both wireline & wireless networks, and have had to re-adjust to the changes in the telecom industry. Three-four years ago there was so much money flying around in telecoms and a lot of Internet-based companies were created that had no right to exist. The backlash from these excesses has hit the sector hard. At the moment few new major installations are being made, but demand will come back, and IDT will then be well positioned to take advantage of it. Hopes of a swift economic recovery seem to have disappeared altogether. Do you think that VCs reluctance to risk money on electronics start-ups, and investor's fears about backing tech firms will push this back still further. If so when do you think things will begin to get better?The VC community is scared at the moment. Start-ups are finding it hard to survive in such a grim investment environment. The fix for this is to hold out until the market picks up again, sometime next year. Hopefully a lot of lessons have been learnt, I don't think that we will see the sort of excesses that were taking place in the 90's for a long time. After achieving so much in the industry, what makes you still want to be part of this crazy circus?Now that I am older (for sure) and wiser (I hope), I have a chance to pass on some of the knowledge I built up over the years. If I can help Jim Thorburn (CEO of ZiLOG), Greg Lang (CEO of IDT) and Francis Leeee (CEO of Synaptics) by showing them my battle scars and sharing my experience, then that will give me much satisfaction.