
He was strapped into a large, complicated chair in a smaller room a few doors down; then the technician
attached electrodes and placed a headpiece on his head.
"There's nothing to worry about,” the technician said, clearly reciting a set speech. “The monitors are just
to keep tabs on your bodily functions. Once we start the procedure, a sleep inducer will put you under
for the duration. When you wake up, it'll be over.” She smiled mechanically.
Casper smiled back shakily, and closed his eyes. The technician flipped the switch to start the sleep
inducer, and Casper quickly slipped under.
The technician checked him over swiftly and efficiently; then she waved the go-ahead signal to the
monitor camera and slipped out of the room. In the central control room another technician saw the
signal, hit a button, and turned away.
The procedure was fully automated, with technicians present only to troubleshoot when something did
not go according to schedule. Under most circumstances, unless an alarm went off or the machines told
them something was wrong, their attention was directed elsewhere. After all, watching someone sleep is
impossibly dull, even if the subject's brain is doing various interesting things.
Casper's chosen skill file consisted of a few gigabytes of data on a microptical disk, tagged and ready to
be fed into his brain; first, however, the scanners had to examine Casper's neural pathways and
brainwave patterns. The file would be imposed on these pathways, but the machines had to be sure that
the file was not so radically opposed to the recipient's mental structure that some harm could occur. Dr.
Jalali's preliminary survey had shown that Casper's brain could accept imprinting, but not that he could
accept any particular program; since the individual programs were all proprietary information owned by
NeuroTalents or their independent vendors, not to be distributed freely to other companies’ doctors even
within the Consortium, the doctor had not had the information to verify that Casper could handle this
specific skill-set.
The central computer began matching program details against neural pathways, checking for conflicts.
While the mapping was taking place, however, a badly-worn sector of old disk storage finally gave out,
dropping approximately sixty bytes from the system's primary command programming, from a total of
some two and a half million lines of code.
When the time came to check the scan against the waiting skill file, an uninitialized variable came up
garbage—the code that should have set it was missing. The error-handling software, never tested in this
particular situation, attempted unsuccessfully to compensate.
The waiting skill file was ignored. The mapping continued, into secondary and then tertiary areas of
detail, levels that were totally unnecessary for an ordinary skill imprint. A set of restricted-access files,
quite separate from the scheduled one, was accessed and readied.
A technician looked up casually from his magazine at the monitoring panel, then stopped and looked
again. He had thought the subject in Suite B was in for a regular skill imprint, but his instruments showed
that he was in the middle of optimization programming.
He didn't remember anyone scheduling any optimizations. Weren't there supposed to be extra
precautions for optimizations? A skill imprint just added a few new patterns to the subject's brain,
plugging in a little new information and some artificial habits, but an optimization more or less rebooted
the entire brain, streamlining the entire personality and redirecting it toward a predetermined goal, adding