How Not to Network a Nation. The Uneasy History of the Soviet Internet

(Ben Green) #1

The Undoing of the OGAS, 1970 to 1989 177

car and driver. Buoyed by such a reputation, he wrote a strong-willed let-
ter to Pravda in 1954, the year after Stalin’s death, detailing a long-term
strategy for world domination without having to go to nuclear war. His
suggestions involved calculated moves and countermoves through which
the socialist leaders would grant the masses of petty capitalist owners their
material wealth in exchange for their acceptance of the socialist revolution
without atomic combat.^38
For such a brazen public stunt, the political secretariat rebuffed him and
threatened to throw him out of the Communist Party. In the 1960s, he
publicly repented and avowed his Communist credentials by publishing
Computers, Chess, and Long-Range Planning, which describes how the domi-
nation of the Soviet school of chess over the Americans was an expression
of superior long-range socialist planning. In 1968, having been influenced
by Claude Shannon’s less well-known 1950 work on computer chess, Bot-
vinnik published An Algorithm for Chess, which successfully demonstrated
how to algorithmically organize attacks against an opponent’s position
from challenging tactical positions. Even though his algorithm excelled in
solving technically stressful positions, it also had the frustrating tendency
to overlook the simplest tactical moves.^39
Backed by major cyberneticists and computer engineers including Vik-
tor Glushkov and his colleagues—such as Bashir Rameev (who developed
the Ural computer series), Viacheslav Myasninkov, and Nikolai Krinitskiy
in the 1970s and 1980s—Botvinnik poured his energies into what he called
(drawing on the Stalinist vocabulary of his youth) the Pioneer Project, a
computer chess program that was designed to imitate how the brain of a
grandmaster works.^40 The OGAS Project algorithms were designed to ignore
the bulk of all computationally possible moves and instead to concentrate
on the most probable moves. Attempts were made to develop an algorithm
with a long-term intuitive “feel” for the board. The brute force approaches
(which calculate all possible moves in branching decision trees) eventually
won out with the arrival of faster computers in the 1970s, outpacing Botvin-
nik’s selective but theoretically more sophisticated approach. Nonetheless,
his prodigies in training—Kasparov among them—remember their surprise
at hearing that Botvinnik was confident that his selective program would
one day consistently beat them all. Nevertheless, it—much like Glushkov’s
attempt to develop intuitive macroprocessing and natural language pro-
gramming that mimicked the neural processing and speech patterns—bore
fruit in other spheres of application. For example, Botvinnik, in his career
as an electrical engineer, reconfigured his Pioneer algorithm into planning
maintenance repair schedules for power stations across the Soviet Union.^41

Free download pdf