Memory warning
Sandra Pe帽a de Ortiz and Yuri Arshavsky suggest the idea that memory
works in a similar fashion to the immune system
(15 September, p 24).
Could memory not in fact be part of the immune system?
While the “standard” immune system deals with internal dangers, a memory
“immune system” would deal with external threats. An example would be touching
something very hot and getting burnt鈥攜our memory immune system would then
be able to inform you the next time you are near a very hot surface that it may
burn you鈥攁 form of inoculation, if you will.
This memory immune system could also lead to language, as a means of
transmitting a meme about a danger to others. Consciousness could be thought of
as the result of the mind evaluating immunity memes for relevance and extending
them to new situations.
This raises the possibility of immunising our children against various
dangers such as traffic before they encounter them, daft as it sounds.
Patently absurd
I really enjoy the patents column鈥攊t’s often the funniest thing I read
all week. The 15 September issue included two classics
“Silencing radioheads” (p 21)
and “Booted out” (p 21).
“Silencing radioheads” suggests replacing all car radios with new models that
automatically mute in response to signals from emergency vehicles. Exactly who
is going to pay for this? The obvious, cheaper and far more effective solution
is to fit fire engines with low-power, wide-band FM radio-jamming
transmitters.
“Booted out” proposes a CO2 detector in car boots to prevent kids or
kidnap victims being locked in. Placing this so close to the exhaust doesn’t
strike me as terribly clever. Wouldn’t it be a lot simpler to fit a
press-to-open latch inside the boot? This would be cheap and effective, and
would work even if the car battery was flat or removed.
If this can go wrong. . .
The funny thing about Murphy’s Law is that it may (or may not) be exempt from
itself
(15 September, p 36).
If it is exempt from itself, it’s invalid as a
general law. If it isn’t, it’s invalid because it, too, can go wrong. So it’s
not a very good law and can’t be taken very seriously. (Nor should this
letter.)
More practically, I offer Sugerman’s Quantitation of Murphy’s Law, which
states that “the less likely it is that something will go wrong, the more
probable it is that it will”. Consider for example situations when you feel as
though nothing is at all likely to go wrong. You may unwittingly drop your
guard, thereby increasing the probability that something will.
Stand in the corner
Feedback really ought to have applied some school physics before ridiculing
the concept of a product being labelled both “Withstands a shock of
1500g” and “Do not drop”
(15 September).
The deceleration following a drop is
the ratio of the fall height to the stopping distance, in units of the
acceleration due to gravity, g. (Assuming uniform deceleration and no
aerodynamic drag.)
Suppose that the innards of the Microdrive are fairly rigidly fixed, but the
casework and mountings can flex up to 1 millimetre on impact. A 1.5-metre drop
onto concrete will then generate about 1500g. Of course the
1-millimetre flex is purely an assumption, and the g forces on impact
for particular parts may vary greatly.
It just so happens that the standard drop test for most small consumer
appliances (phones, watches etc) is a fall of 1.5 metres or 2 metres onto steel
or concrete, depending on the industry. The most fragile electronic components,
such as quartz crystals, are usually specified by the manufacturers to withstand
1500 or 2500g. Many electronic components such as solid resistors are
capable of withstanding considerably more than this.
No smoke without ire
Richard Peto states that half of Britain’s 10 million smokers will die from
the habit, almost as if they wouldn’t die if they didn’t smoke
(1 September, p 44).
But in developed countries, apart from emphysema and lung cancer (around
0.2 per cent of smokers die of these diseases), the vast majority of smokers
tend to die of the same diseases as non-smokers鈥� they just die sooner.
But Peto tells us this is not the case in developing countries where,
depending on country, smoking causes death by emphysema, liver and stomach
cancer or even TB. The variety of cancers and disease caused by smoking calls
into question the central tenet of smoking-related carcinogenesis, namely that
smoke carcinogens cause mutations that lead to cancer. Either that’s not the
case, or we are confronted with some pretty dastardly carcinogens that can
distinguish which country they are in before they decide what disease they will
cause.
Most smokers indulge because they enjoy it, and as such their actions are not
irrational. Many of us engage in activities that may shorten our lives. Isn’t it
about time we left smokers alone? If the coercive utopians win this battle they
will only turn their attentions to some other “undesirable” behaviour that the
rest of us enjoy.
I think it thinks
Brian Aldiss goes out on a limb when he claims “intelligence cannot exist
without consciousness”
(15 September, p 40).
One problem in the debate over
human and machine intelligence has been defining exactly what intelligence and
consciousness are鈥攐r even whether they are, for any given definition,
real.
If you define intelligence as “the ability to solve problems”, it becomes
easier to draw at least some conclusions. Then you can start discussing the
ability to apply problem solving in one domain to problem solving in
another.
In the world of existing technology, you can argue that computers are
intelligent at adding numbers, but not so intelligent at, say, assessing human
emotions or language translation. This does not require positing computer
consciousness.
The more difficult question is: what is consciousness? Without an answer to
that, any debate about its relationship with intelligence is difficult. Is
consciousness simply an inevitable artefact of systems beyond a certain level of
a certain sort of complexity? If so, it might be possible to argue that certain
classes of problem solving (and hence certain types of intelligence) require
consciousness, but the original assumption is a big one.
Shhhh . . .
According to the article “Snorers are silenced with a harmless shot”, a
technique developed at the Walter Reed Army Medical Center in Bethesda,
Maryland, “can reduce snoring loudness from 11 decibels to 2 decibels”
(22 September, p 16).
Compare this with the ambient night-time noise levels in a
typical British bedroom, with all the windows closed: between about 20 and45dB(A). This depends mainly on how close it is to major roads, and the type of
heating and ventilation devices installed.
Against these background levels, snoring noise peaking at 11dB would not
interrupt sleep, in spite of its transient nature. And 2dB is virtually
inaudible鈥攐nly specially constructed acoustic laboratories are that quiet.
Loud snoring is roughly equivalent in peak sound power to conversational speech,
which would produce sound-pressure levels at the ear of a sleeping partner in
excess of 65dB(A).
Maybe the article should have said “reduce loudness by between 2 and 11 dB”.
A 10dB reduction is, by definition, a halving of sound energy.
Ditch the air
“Out of thin air”, your article about aligning particles in response to
standing-wave configurations, was interesting
(1 September, p 32).
What puzzles me is why one would resort to such heroics as flying parabolic
arcs in aeroplanes to demonstrate the concept.
Surely the effect would also apply to particles suspended in a liquid of the
same density, giving them neutral buoyancy? It would seem that the greater
energy densities applicable in the liquid would make the process more efficient,
while the use of tuneable ultrasonic transducers to set up the acoustic field
would permit high shape resolution as well as saving perfectly good loudspeakers
from the scrap heap.
Not as much fun, to be sure, but easier on the ears.