What Silicon Valley doesn’t say in its Trump attack

14.07.2016
Donald Trump will be "a disaster for innovation," according to some of Silicon Valley's technology leaders. But the innovation disaster they're warning of is already ongoing in America.

U.S. support for research is declining, and just last month China surpassed the U.S. in number of supercomputers on the Top500 list. Both countries are now in a race to build exascale systems (1,000 petaflops), a competition the U.S. is almost certain to lose based on published roadmaps.

The U.S. has set 2023 as its goal for exascale; China is aiming for 2020.

The Silicon Valley letter doesn't mention supercomputing, or any of the problems affecting science in the U.S.  

Vince Cerf, the vice president and Chief Internet Evangelist for Google, Steve Wozniak, Apples' co-founder, Padmasree Warrior, a former top technologist at Cisco who is now the CEO for NextEV, an electric vehicle company, and the approximately 150 others who signed the letter, fix a laser-like focus on Trump's shortcomings -- and nothing else.

It may well be, in the minds of Silicon Valley leaders, that Trump, the presumptive Republican nominee, will double down on the U.S. retreat from science investment.

China now has 167 on the Top500 list, a global accounting of the world's most powerful supercomputers -- at least those publicly known. The U.S. has 165. China also built the world's fastest supercomputer without U.S. chips.

"That's the first time this has happened," said Alex Larzelere, a senior fellow at the U.S. Council of Competitiveness, in an interview prior to the release of this letter.  Ten years ago China only had 28 computers on the list, while the number of computers the U.S. has is declining.

China's rise in supercomputing also means it is developing an ecosystem for high performance computing: The middleware, operating system software, mathematical libraries and other systems, said Larzelere.

"The Chinese are continuing up, and the U.S. seems to be level, at best" in supercomputing, said Larzelere, who has developed advanced modeling and simulation capabilities at U.S. Department of Energy's national labs. "It's not really a problem today, but it can potentially be a huge problem down the road.

"China has said you can't be a superpower without supercomputing," said Larzelere.

The Silicon Valley letter, signed by tech inventors, entrepreneurs, engineers, researchers and others, doesn't ask lawmakers to take a hard look at the direction of the U.S. That's because the real focus of the Silicon Valley letter is on Trump's immigration policies.

Trump would put some curbs on employment-based green cards, and particularly the use of H-1B visas.

Trump, "stands against the open exchange of ideas, free movement of people, and productive engagement with the outside world that is critical to our economy -- and that provide the foundation for innovation and growth," the letter says.  

The document isn't an endorsement of Hillary Clinton, the presumptive Democratic nominee, and doesn't even mention her. But Clinton's proposal to "staple" a green card on the diplomas of foreign students who graduate with advance degrees from U.S. schools is exactly the kind of policy the tech industry has lobbied for.

Clinton has released a tech agenda that calls for improved research funding, while Trump has yet to produce his plan. And even if Trump pens a tech policy to appease Silicon Valley, its condemnation of his immigration plan is total -- and immigration is at the heart of his campaign.

The Silicon Valley letter accuses Trump of holding "a fundamental belief that America is weak and in decline." This declinist belief, as it's called, has real roots apart from the U.S. retreat in science research funding.

In the 1990s, the U.S. abandoned its efforts to develop a super collider. Europe built the Large Hadron Collider near Geneva, making that region a center for physics research. Funding for NASA, an engine of innovation, as a percentage of the federal budget has been declining.

The U.S. had published plans about six years ago targeting 2018 as the rough time to build an exascale system, but the funding and support lagged. Meanwhile, China's push into this area is relentless and a high national priority.

(www.computerworld.com)

Patrick Thibodeau

Zur Startseite