NISKAYUNA — Dick Messmer is marking his 50th anniversary at General Electric Global Research this year.
And the scientist, a Pittsburgh native, still isn’t finished with a career that has seen the award of 22 patents, publication of over 180 technical papers and dozens of leaps forward in the technology he uses to do his work.
1969 was hardly the Stone Age — humankind first walked on the moon that year — but when Messmer arrived on the hilltop campus in Niskayuna, then called GE R&D, he carried a slide rule and did many of his calculations with a pencil and paper. The one mainframe computer on site took up several rooms and was programmed through punch cards.
In 2019, he’s using artificial intelligence and almost infinitely faster computers to help GE prepare a proposal to help the federal government modernize the electrical grid.
Even the smartphone he carries in his pocket is over 10 billion times more powerful than the old mainframe — and it’s just an ordinary iPhone, not a deluxe physicist model.
“I frequently say all the work I did in the first two years I was at GE I could probably do now in two weeks,” said Messmer.
Not only was that mainframe cumbersome to program and slow to compute, the manager let him use it only once a day, because he was consuming so much of its computing power.
He doesn’t miss the days of carrying a stack of nearly identical punch cards around in a box.
“If you ever dropped the box it would take you quite a while to get them back in the right order,” Messmer said.
While he has relied from the start on computers to do his work, they've never been the focus of his work.
“The computer is an aid to getting the work done,” he said. “I was always interested in solving physics problems.”
General Electric sold its mainframe computer business to Honeywell in 1970, so even if Messmer had been interested in developing computer hardware, the opportunity wouldn’t have lasted at GE.
Instead, his varied career has included materials science, risk analysis, quantitative finance and artificial intelligence, as well as computer science.
“The management at GE Research has always been very good at allowing those changes to take place and so I’ve always been very appreciative of that as a model,” Messmer said.
His experience before coming to General Electric was almost entirely in academia, having earned a doctoral degree in theoretical science at age 26, worked at MIT at 27, then came to GE at age 28. His sense was that he never would have ranged so far and wide in a college setting, as doing so would have reduced rather than boosted his credibility.
Messmer’s earliest work at GE included using that old mainframe to show that General Electric’s ongoing efforts to develop its synthetic diamond as a semiconductor would be in vain, that the concept just wasn’t workable.
He later led a team that convinced management to hold off on using parallel computing when that was still a new field. The technology was very expensive, the standards were fragmented, and as it turned out, most of the companies pitching the technology early on would go out of business. “There weren’t the standards there that make people comfortable,” Messmer recalled.
GE Global Research would later make extensive use of parallel computing, and still does today, but it may have avoided an expensive miscue by not jumping in immediately.
Messmer also led the team that developed an analytical approach to insurance underwriting that saved GE Capital tens of millions of dollars.
Along the way, he argued successfully for Global Research to upgrade to mini computers. “Each group then had the ability to do a lot more,” rather than wait their turn on the mainframe, he said.
Upgrades to work stations and then PCs followed, and the computing power at the disposal of the Niskayuna scientists just kept growing.
General Electric was willing to invest in these continual computer upgrades for its researchers, Messmer recalled, but not willing to just hand over a blank check.
“Budgets are always important,” he said. “One always has to make a strong argument. If you’re too far into the future, the cost of the equipment hasn’t taken the ride down.”
Messmer and his wife of 52 years, Beverley, have lived in Rexford the last 30 years and lived in Niskayuna for the preceding 20 years.
After a half-century at Global Research, he clearly still enjoys coming in to work each day, and has mixed thoughts about the retirement he knows will come, at some point.
“It’s going to be difficult because it’s so much fun to be doing stuff,” Messmer said. “On the other hand, I would like to do a little traveling and I’m not getting any younger.”
His latest project is a return to risk assessment modeling, which continues to get more accurate with improvements in computer technology, as better data and better analysis allow better predictions of real world results.
“Everything is an approximation to reality,” Messmer said. “The question is, what level of approximation is necessary in order to do something that is game-changing?
“In general the approximations are much better than they were” even a decade ago, he said. “It allows for a lot of new things to be done.”
GE Global Research has multiple projects underway connected to the federal government’s initiative to upgrade and modernize the nation’s aging electrical grid. It’s now seeking a contract with the Advanced Research Projects Agency-Energy, which is looking years ahead at technology so new it doesn’t attract sufficient private-sector investment to advance.
Messmer is part of a team that would analyze the many unknowns surrounding the new technology used for electrical transmission, and the consequences for the stakeholders should that technology fail.
“That’s a very interesting and very intricate problem,” he said.
“We’re looking forward to putting together a real nice proposal for that.”