25 July 2020 | New Scientist | 55
The back pages Feedback
Casso-wariness
It has been a while since Feedback
last covered units of measurement
used in guidelines for social
distancing. That’s partly because we
thought we had spotted all the good
ones, but also, we confess, because
we took our eye off the ball. So we
are grateful to Simon Kravis for his
example from the signage at Cairns
Airport in Queensland, Australia.
According to a news story from
ABC, the airport is planning to use
the length of the cassowary – a
flightless Australian bird – to keep
travellers the necessary distance
apart. Adult cassowaries are
apparently around 1.5 metres
from nose to tail, so imagining
that one exists between you and
your nearest neighbour seems like
sound advice. But with claws that
can exceed 10 centimetres in length
and a predisposition for kicking
nosy passers-by, you are probably
going to want to stay even further
away from that imaginary
cassowary. Those travelling
through Cairns are likely to be
very socially distanced indeed.
Amen to that
Feedback would be nothing
without our colleagues. They
selflessly share with us their
anecdotes and expertise, keep us
informed of the goings-on in the
world and are always willing to
trawl the depths of their spam
folder for entertaining nuggets.
This week’s nugget is
particularly amusing, for, after
a bit of throat-clearing, it runs:
“churches like NEW SCIENTIST
are struggling more than ever
during the new normal. Giving,
attendance, events, and more
are down or non-existent.”
We hear you, unnamed email
sender, we hear you.
That’s why, it continues,
“you will have 6 months to take
advantage of a complete risk-free
church solution”. If risk-free
solutions are available, then
rebranding as a religion is
certainly an attractive prospect.
But we are curious to know what
I spy a joke
Scientists shouldn’t be allowed
to name things. All too often,
they crowbar in their own names
or come up with an unconvincing
pseudo-acronym that elicits
groans at every conference.
Or, if they are astronomers, they
just go with a dull and pedestrian
option. This mindset has gifted
us the (existing) Very Large
Telescope, the (soon to be)
Extremely Large Telescope and
the (perhaps mercifully cancelled)
Overwhelmingly Large Telescope.
Unabashed, Anna T. P. Schauer,
Niv Drory and Volker Bromm have
uploaded a paper to the arXiv
preprint server in which they
speculate about a hypothetical
Ultimately Large Telescope. It is so
powerful that it can see back to the
cosmic dawn when astronomers
finally lost the spark of creativity.
Fake News-scientist
Nine times out of 10, when
Feedback comes across a humorous
fragment of text on the internet
purportedly written by an AI, the
trace of a human hand is clearly
apparent. Items in the exceptional
10 per cent have – so far at
least – originated from GPT-2,
a text-generation programme
developed by OpenAI. Its fearsome
ability to predict what should
follow a given sample of text led
to its creators not releasing the
full version, for fear it be used to
create fake news indistinguishable
from the real thing.
It turns out that a comparatively
well-kept corner of the internet
forum Reddit is devoted to the
output of GPT-2, wherein every
post, comment and response is
artificially generated. And lo and
behold! Among the recent threads
is one responding to a completely
made up New Scientist headline:
“Human body is the oldest DNA
ever discovered and it came from
a fossil of a human who was
2000 years old.”
Indistinguishable from
the real thing, eh? But while our
subeditors may not be quaking
in their boots just yet, it is the
comments that really confound
us. “That’s a new record for us
humans,” says one. “You got to
admit, it’s pretty cool to look at
old genetic sequences being
compared to other old genetic
sequences,” says another.
“I’m not sure how much
weight this paper holds,” begins
a third, with laudable scepticism,
adding that “the fossil is from the
Hoxne-Dahydran culture, which is
the earliest known culture to have
agriculture and a complex society.
This is the only known evidence
for a pre-agricultural society
anywhere in Europe.”
Astonishingly coherent,
distinctively erudite and, alas,
involving entirely made up facts.
As a simulated version of an
online comment, that takes
some beating. ❚
Written by Gilead Amit
our distinct appeal might be. What
sort of rituals would we engage in?
What holy texts – if any – would
we adhere to? Your offerings
(via email, post or bank transfer)
are welcomed.
Have you finnished?
Nominative determinism in English?
Old news. Hackneyed. Boring.
Nominative determinism in Finnish?
Exciting. Frisson-inducing. Moreish.
This is Feedback’s new policy,
spurred on by an email from Ilpo
Salonen who informs us that
a leading figure at the Finnish
mining and metal-refining firm
Metso Outotec is called Markku
Teräsvasara. The Finnish speakers
among you will scarcely need to
be reminded that this surname
translates as “Steelhammer”. “What
a waste of a useful heavy metal
band name,” says Ilpo. Ba-dum-tsh.
Got a story for Feedback?
Send it to [email protected] or
New Scientist, 25 Bedford Street, London WC2E 9ES
Consideration of items sent in the post will be delayed
JOS
IE^ F
OR
D