72 Science & technology The EconomistSeptember 28th 2019
2 The trick is to maximise the chance of
choosing the right answer instead of one of
the zillions of wrong ones. This is where
the third counterintuitive idea comes in. In
classical physics, probabilities must be
positive—a 30% chance of rain, say. Quan-
tum mechanics uses a related concept,
called “amplitudes”. These can be negative
as well as positive. By ensuring that ampli-
tudes which represent wrong answers can-
cel each other out, while those that repre-
sent the right one reinforce, programmers
can home in with high confidence on the
correct solution.
That is the explanation which textbooks
present, anyway. In the laboratory, things
are rather more difficult. Superpositions
and entanglements are exceedingly deli-
cate phenomena. Even the jiggling of adja-
cent molecules can interrupt them and sul-
ly a calculation. Most designs for quantum
computers require the machines to be
stored at temperatures colder than that of
deep space, and to be tended by a basement
full of phds, to keep things on track.
No height of education or depth of cold,
though, can altogether preclude errors
creeping in. The biggest problem facing
quantum engineers is how to spot and cor-
rect these, because most of the useful ap-
plications of quantum computing will re-
quire many, many more qubits than
current devices sport—with a concomitant
increase in the risk of errors. That has
spurred a huge effort, both by well-known
firms such as ibm, Intel and Microsoft, and
by an eager band of newcomers, such as Ri-
getti, to build better, less error-prone kit.
There is also, in parallel with this race to
build better machines, a race to develop
useful quantum algorithms to run on
them. The most famous example so far is
probably Shor’s algorithm. This is the piece
of quantum-turbocharged maths that al-
lows rapid factorisation of large numbers
into their component primes, and thus
scares cryptographers, a group whose trade
depends on this being a hard thing to do.
But if quantum computers are really to
earn their keep, then other algorithms will
be needed. Developing them will be assist-
ed by the fact that a lot of the proposed ap-
plications (drug design, materials science
and so on) themselves depend on quantum
processes. This, indeed, is why those appli-
cations have been so intractable until now.Little acorns
Despite the promise of quantum comput-
ing, many in the field are uncomfortable
with the phrase “quantum supremacy”, for
it implies a threshold that, once crossed,
leaves decades of existing computer sci-
ence in the dust for something weird and
wonderful. And for all the “before” and
“after” that Google’s paper represents,
building practical, error-corrected ma-
chines will be far from easy.Itisthereforeamistake,mostpeople
think,tobelievethatquantumcomputing
willreplacetheclassicalsort.Thepracti-
calities of low-temperature operation
aloneare likely to see to this. Govern-
ments,bigfirmsandtherichersortsofuni-
versitywill,nodoubt,buytheirownma-
chines.Otherswillrent timeondevices
linkedtoquantumversionsofthecloud.
Butthetotalnumberofquantumcomput-
erswillbelimited.
Andthatwillbefine.Butitisworth
bearinginmind asimilarprediction of
limiteddemandmadeintheearlydaysof
classicalcomputing.In 1943 ThomasWat-
son,thenbossofibm, isallegedtohave
said,“Ithinkthereisa worldmarketfor
maybefivecomputers.”Hewasoutbya fac-
torofperhapsa billion. 7T
he mid-pleistocenetransition was a
significant event in the history of
Earth’s climate. It marks the point, be-
tween 1.2m and 900,000 years ago, when
the ice-age cycle of freezing glacial periods
alternating with warm interglacial ones
(which began about 2.6m years before the
present day) flipped from being 40,000
years long to 100,000 years. Climatologists
would like to know why.
The answer is important because, on
past performance, the cycle should be
about to turn cold again. Studies of post-
transition cycles, though, suggest that one
important regulator of what is happening
is carbon dioxide, a greenhouse gas that
people have been pumping into the atmo-
sphere in unnatural quantities for a cen-
tury or more. Understanding CO 2 ’s influ-ence on climates gone by may help predict
the details of its role in the future. Teams
from Australia, China and Europe are
therefore engaged in a friendly competi-
tion to gather samples of air that are as
much as 1.5m years old. These they hope to
find trapped in the lower layers of what will
be the deepest ice cores drilled from the
continent of Antarctica.
Mere depth, however, is not necessarily
enough to achieve the desired goal. The
horizontal flow of the topmost layers of an
ice sheet can mix up those lower down,
making them difficult to date. And older
ice, closer to the bedrock, may be melted by
heat rising from Earth’s interior. Research-
ers from all three teams have therefore
spent the past few years seeking the opti-
mum place to drill. They have dragged ice-
penetrating radars far and wide across Ant-
arctica’s surface to map the layers beneath,
and sunk exploratory boreholes to try to
gauge how warm it is likely to be in the
deepest sections of the ice.
The Europeans, led by Carlo Barbante, a
climate scientist at the Ca’ Foscari Univer-
sity of Venice, seem to be the first to have
struck metaphorical gold. In April Dr Bar-
bante and his colleagues announced that
they had identified a spot in an area called
Dome C (see map) that probably includes
ice undisturbed by melting or folding. This
site is some 40km north-east of Concordia
station, a base run jointly by France and Ita-
ly. The process of extracting a core nearly
3km long from this site is scheduled to start
in 2021. The actual drilling will take six
months, but because those months are re-
stricted to two per year during the Antarc-
tic summer, the whole project will last sev-
eral years. Dr Barbante expects preliminary
data to be available by 2025.
Tas van Ommen of the Australian Ant-
arctic Division, a government agency, is
also planning to drill near Concordia. He
and his colleagues expect to start in 2022 at
a location 5-10km from Dr Barbante’s site.
On September 23rd they unveiled the new
drilling equipment with which they hope
to extract their core.
The third project, organised by the Polar
Research Institute of China, is in Dome A,
closer to Antarctica’s centre than Dome C.
Dome A has low snowfall and thick, sta-
tionary ice. These are propitious for the
preservation of ancient ice layers, but the
dome is located over buried mountains,
which are likely to complicate the pattern
of geothermal heating from below.
Local difficulties aside, these three pro-
jects should together push understanding
of the mechanisms of glacial and intergla-
cial periods back through the barrier of the
Mid-Pleistocene and closer to the point in
time when the ice ages began. With luck,
after that is done, the past will illuminate
the future and the nature of the climate to
come will be clearer. 7A quest to obtain the oldest ice core
from Antarctica is beginningPalaeoclimatologyData from the
freezer
DomeARoss DomeC
Ice ShelfWeddell
SeaRoss
Sea0°180°90°W 90°E
South PoleANTARCTICA1,000 kmSOUTHERN
OCEAN