A fireside chat with avuncular Gordon Moore - some things you didn't know about Intel's co-founder - Industry - News

A chat with industry luminary Gordon Moore

HEXUS had the opportunity to sit in and listen to an engaging interview with genial Gordon Moore, Intel's co-founder.

How did Intel come about?

Gordon was working under William Shockley at the Shockley Semiconductor Labs but left to join Fairchild Semiconductor, along with 7 other 'traitors', in 1957. Another of the supposed traitors was Robert Noyce, who was consistently overlooked for promotion. Bob decided to leave and Gordon, being a sympathetic chap, left in sympathy.

Bob and Gordon saw that a profitable opportunity existed in engineering and production of semiconductor memory. Intel was formed by the two in 1968. Gordon admitted that Intel got pretty lucky when they pioneered research into silicon-gate MOS and SDRAM memory, surmounting early problems, and made hay whilst other established companies looked elsewhere for profits. Intel had a monopoly on silicon-gate MOS for 7 years, Gordon beamed.

The name?

Gordon and Bob tried a number of options but settled on Intel. The name had to be bought from a midwest hotel chain (imagine how much they could charge now!). If you don't know, Intel denotes Integrated Electronics.

Moore's Law

Gordon's probably more famous for his eponymously titled Moore's Law than for his co-founding of an industry giant. The term was coined by physicist Carver Mead and predicts that the number of transistors on an IC will double every two years.

The legend of the Intel cubicle

Segueing on to some personal nostalgia, Intel's HQ in Santa Clara is somewhat is akin to a prison block. Gordon reckoned that a open-plan cubicle design would make for a better working environment, hence the design was one of hundreds of cubicles (it's quite sobering, let me tell you), and practically everyone, irrespective of position, had to work in them. Gordon still maintains the largest cubicle at Intel, required to house his huge table.

Moore's Law and imminent danger before hafnium

Moore's Law will come to an end reasonably soon, according to the man himself. Before Hafnium (45nm breakthrough) Intel was just 5 molecular levels from disaster. Moore reckons that the next decade will see an inexorable march towards the end of Moore's Law, as we know it. We come up against fundamental barriers that cannot be overcome by tradition methods, so the race is on to develop newer tech.

How has the semiconductor industry helped other sciences?

Gordon studied chemistry at graduate level and chemistry and physics for his doctorate. Gordon reckons that the advances made in the semiconductor industry are key enablers of progress in other sciences. The economies of scale from large-scale production has brought the cost of transistors to a microscopic level. Thinking about it, a 65nm quad-core desktop processor carries around 580M transistors and costs from £160.

The bumpy road of the 80s

Intel had a somewhat bumpy road in the 1980s, Gordon mused, as roadmaps were somewhat predictable, especially with respect to DRAM engineering. Japanese manufacturers effectively copied Intel's lead and became the pacesetters the field for a while.. Gordon mentioned that there was a time in the 80s when the US semiconductor industry was in real, tangible danger.

Surprising innovation

The most surprising innovation? Taking a moment to answer, Gordon mentioned that an Intel innovation that was particularly striking was the size of current silicon wafers, measuring 300mm. Thinking about it, ensuring that you garner effective yields on a 300mm, 45nm wafer, encompassing trillions of transistors, is still a mindblowing achievement.

Interests

Fishing and spending his hard-earned money. Gordon famously donated around $600M to Caltech some six years ago. He's of the opinion that he knows how better to spend money than the government (tax!).

Where will it go

Gordon reckons that the next evolution that will bring a greater number of users 'closer' to their computer is advances in speech- and language-recognition, where computers will be able to decipher and translate in real-time. We've heard this kind of thing before, but we're still some way off personal computers efficiently processing voice commands and becoming truly interactive 'partners'.