Slashdot Mirror
Trump Signs Legislation To Boost Quantum Computing Research With $1.2 billion (geekwire.com)
President Donald Trump last week signed legislation ramping up quantum computing research and development. From a report: The National Quantum Initiative Act (H.R. 6227) authorizes $1.2 billion over five years for federal activities aimed at boosting investment in quantum information science, or QIS, and supporting a quantum-smart workforce. The law also establishes a National Quantum Coordination Office, calls for the development of a five-year strategic plan and establishes an advisory committee to advise the White House on issues relating to quantum computing. "This next great technological revolution has far-reaching implications for job creation, economic growth and national security," Michael Kratsios, deputy assistant to the president for technology policy, said in a White House statement. "We look forward to building upon efforts to support the quantum-smart workforce of the future and engage with government, academic and private-sector leaders to advance QIS."
[...] Earlier this month, a report from the National Academies of Science said there is an urgent need to develop "post-quantum" encryption protocols in order to protect commerce and national security. QIS research could also produce new types of quantum processors, sensors, navigation tools and security systems. The challenges could bring about "new approaches to understanding materials, chemistry and even gravity through quantum information theory," according to a White House strategy paper issued in September.
[...] Earlier this month, a report from the National Academies of Science said there is an urgent need to develop "post-quantum" encryption protocols in order to protect commerce and national security. QIS research could also produce new types of quantum processors, sensors, navigation tools and security systems. The challenges could bring about "new approaches to understanding materials, chemistry and even gravity through quantum information theory," according to a White House strategy paper issued in September.
"Americans deserve the best quantums ever! We will out quantum Jiiina, believe me! Schrodinger didn't send us his best cats. No more cats from shit-hole dumpsters; we will hire THE best cats, and I know cats better than the Generals. I grab really terrific quantum bits; even those Q-bits, and all letters of the alphabet even. Know 'em well. T-bits are my favorite, of course [applause]. #MakeAmericanCatsHalfDeadAgain!"
Table-ized A.I.
WTF, I hate quantum computing now!
Although Quantum processors are unproven so far,
Wrong, I think what you meant to say is universal quantum computers are unproven.
Quantum Math co-processors, essentially what is done now, are proven to be better than classical computing.
Adiabatic QC works very well. It's already used by several companies now. Big companies. Big, smart companies at that. Not some random cunt with a Wordpress moaning. (or was it blogspot?)
We were just wrong at the qubit-counts required to outpace classical computing, which we thought were considerably smaller.
Whether general purpose QC will ever arrive is... eh, I dunno if it will work.
Some issues are inherently impracticable to do that way.
Same issue with serial vs parallel, some tasks just cannot reliably be done over multiple distinct steps because one step absolutely relies on previous steps data to continue. But sometimes sub-steps in each step can be accelerated. Depends how complex they are. But really, that's playing with terminology if anything. Fact is, some tasks 100% rely on previous steps to function. You can't do step N if step B isn't done no matter how hard you try.
At best, you can prepare some resources in advance, like reading a file in to RAM, opening a device, etc.
Whether this will boost performance overall or if it is a detriment is probably dependent on workload in question.
QC won't be able to accelerate that much unless preparing those resources can be accelerated too. Using the results of one step to sort of narrow down where data is likely to be in a large dataset before you finally get the full result of a further step, useful in fractal-related datasets or large science-project datasets, even machine-leaning.
Even humans suffer with serial order problems. We can do parallel tasks brilliantly, but when you come to tasks that rely on previous information to continue, we slow to a crawl.
It's just a problem in general that can't be overcome unless you start breaking the laws of physics and use time machines. Time-machine based computation throws a fuckhuge spanner in to the works. (if you ignore causality, that is!)
I guess someone in Intelligence got through to Trump.
I think it's more reasonable to assume he just likes signing things than assume someone changed his mind.
Anons need not reply. Questions end with a question mark.
As he stated himself more than once, he firmly believes that he has an intuitive grasp of science that eclipses any career expert in any field.
Since he doesn't read much I gather that his intuitions are pretty much based on whatever presentation he recently saw that impressed him. Maybe online or maybe someone who thinks they are talking to the president. Add to that add a healthy mix of preconceptions.
What a recipe. It results in things like this sometimes.
Trump has no idea what he is doing.