Winding Down
An idiosyncratic look at, and comment on, the week's net, technology and science news
by Alan Lenton
23 February 2020
A bit of an odd issue this week, because there are a couple of larger items taking up the bulk of the issue. The first is a look at so called artificial intelligence (a friend of mine always refers to it as 'artificial stupidity'), then there is a piece about modelling the spread of the coronavirus epidemic, and problems with the generally accepted theory of how the early universe developed, the latter with an extra bit on how scientific revolutions work.
There's a couple of nice astronomy pictures (well one is actually a bit creepy), and two quotes, one on self-driving cars, and one on banks - both of them with a cynical, but undeniable, bent.
Finally, scanner contains URLs pointing to material on robots and jobs, the rationality of voting, a ransomware attack on a natural gas production facility, a 3-D reconstruction of ancient Athens, the FCC ordered to repeat its net neutrality public consultation, the Total company in court over greenhouse gas emissions, and a round up of recent space news.
That should keep you going for the rest of the week!
Credits: Thanks to readers Barb and Fi for drawing my attention to material for Winding Down.
Updates:
Amazing! No updates this week. I think the usual suspects are keeping their heads down...
Artificial Intelligence:
The first thing you should know about 'artificial intelligence' is that it's a much abused term bandied around by companies in an attempt to appear 'hip' and on the cutting edge.
What they are really offering is an attempt to model the effects of some sort of behaviour, either human or otherwise, for the purpose of prediction, often using (or misusing) statistics. Often the claims are backed up by studies apparently showing that the software performs better than humans.
Be cautious about this. Such studies are full of pitfalls. Often they use college students as the humans. Students are cheap, plentiful and available en mass. But they are a distinct sub-set of humans with different characteristics from other sections of humanity. Similar problems occur when using people selected from Amazon's Mechanical Turk souk. They rarely have specific training in the topic they are making decisions on.
Take the case of facial recognition. There are plenty of studies out there claiming that some piece of software is better at recognising faces than untrained humans are. I've never seen a single study that checked to eliminate test subjects with face blindness (officially called prosopagnosia). It wouldn't require very many people with this problem to drag the human score down. Actually, it works the other way as well. There are what are termed 'super-recognisers'. They are very rare and much in demand by police forces! (OK - I admit I'm partially face blind, which is one of the reasons why facial recognition interests me...).
At the end of the day it's very difficult and error prone trying to compare human and computer performance, especially where the work involves making predictions about individual humans is concerned. Statistics is a science which deals in large numbers - the larger the better. It can tell you, for instance, that in general a large number of people will get it right 65% of the time, but it can't predict -which- of those people will always get it 65% of the time!
https://www.scientificamerican.com/article/will-past-criminals-reoffend-humans-are-terrible-at-guessing-and-computers-arent-much-better/
Of course, there are cases where computer models can be very helpful. For instance, modelling the spread of the coronavirus infection can be very helpful in efforts to contain it. Of course, assumptions have to be made, in this case about the reproduction rate of the virus (i.e. about how many people each person with the virus spreads it to) but researchers are refining that as more info becomes available.
The results of such modelling so far indicate that the Chinese government is being less than truthful about the scale of the epidemic in China. Somehow, I'm not surprised... Take a look at the URL for an interesting and informative article and video explaining how this all works.
https://www.scientificamerican.com/article/heres-how-computer-models-simulate-the-future-spread-of-new-coronavirus/
Cosmology:
Astronomers have a generally agreed version of how the universe started (the so-called Big Bang) and behaved early in its evolution. Of course, all this happened a long time ago - 13.8 billion years ago, in fact. Recently, a new ultra massive massive galaxy has been discovered which some time in the first 1.8 billion years stopped forming new stars.
According to accepted theories of the early history of the universe, that shouldn't have been able to happen! Our models of star formation predict that star formation shouldn't have been 'turned off' that early.
Still, that's the way science works. Everything trundles along happily for a while, and then suddenly something occurs which the old theory not just didn't predict, but often says couldn't happen, and everyone has to go back to the drawing board and amendments, exceptions and fudges have to be made. Eventually someone comes along with an upstart new model that removes all the fudges and is an integrated whole.
This new interpretation is resisted by the powers that be, who have built their professorial reputations on tweaking the old model, but eventually they die off and the usurper becomes the accepted model, and the whole thing starts over again!
OK, OK. I admit that there is an element of my cynical sociologist coming to the fore in that last bit, but it's not without a strong element of truth. If you want to know more about this topic I'd suggest you read 'The Structure of Scientific Revolutions' by Thomas S Khun. I first read it in the early 1970s and it was a revelation to me. Later, as I studied more sociology I came to realise that things were much more complex than it suggested, but I still consider it to be a brilliant analysis.
https://www.sciencealert.com/astronomers-have-found-an-ancient-galaxy-that-grew-fast-then-died-suddenly
https://en.wikipedia.org/wiki/The_Structure_of_Scientific_Revolutions
Pictures:
This week two pictures from NASA's Astronomy picture of the day
The Fastest Rotating Galaxy Known
https://apod.nasa.gov/apod/ap200219.html
The Dark Nebula in Orion - rather creepy looking, I think!
https://apod.nasa.gov/apod/ap200221.html
Quotes:
Two thought provoking quotes this week.
The first is from a Bruce Sterling article "Artificial Morality"
"In the hermetic world of AI ethics, it’s a given that self-driven cars will kill fewer people than we humans do. Why believe that? There’s no evidence for it. It’s merely a cranky aspiration. Life is cheap on traffic-choked American roads – that social bargain is already a hundred years old. If self-driven vehicles doubled the road-fatality rate, and yet cut shipping costs by 90 percent, of course those cars would be deployed."
http://blog.lareviewofbooks.org/provocations/artificial-morality/
The second is about traditional banks and their attitude to change
“Big banks are set up to prevent change, basically. A lot of the systems inside the bank – particularly the risk and compliance functions – are there to stop things changing. Because if things change it creates risk. And so they’re like these antibodies that go around hunting change and trying to kill it.”
Tom Blomfield, CEO of British digital bank Monzo. (CNBC (https://www.cnbc.com/2020/02/10/monzo-ceo-tom-blomfield-big-banks-are-set-up-to-kill-change.html) )
Scanner:
Crazy idea but hear us out... With robots taking people's jobs, can we rethink this whole working to survive thing?
https://www.theregister.co.uk/2020/02/11/french_robot_study/
Given how little effect you can have, is it rational to vote?
https://aeon.co/ideas/given-how-little-effect-you-can-have-is-it-rational-to-vote
When the air gap is the space between the ears: A natural gas plant let ransomware spread from office IT to ops
https://www.theregister.co.uk/2020/02/19/dhs_confirms_ransomware_attack/
Explore Ancient Athens 3D, a digital reconstruction of the Greek city-state at the height of its influence
http://www.openculture.com/2020/02/explore-ancient-athens-3d-a-digital-reconstruction.html
FCC forced by court to ask the public (again) if they think tearing up net neutrality was a really good idea or not
https://www.theregister.co.uk/2020/02/21/fcc_net_neutrality/
French oil company Total in court in landmark case over greenhouse emissions
https://www.euronews.com/2020/01/28/french-oil-company-total-in-court-in-landmark-case-over-greenhouse-emissions
SpaceX's next Starlink volley remains stuck on Earth to glee of astronomers everywhere
https://www.theregister.co.uk/2020/02/17/space_roundup/
Footnote
Please send suggestions for stories to alan@ibgames.com and include the words Winding Down in the subject line, unless you want your deathless prose gobbled up by my voracious Thunderbird spam filter...
Alan Lenton
alan@ibgames.com
23 February 2020
Alan Lenton is an on-line games designer, programmer and sociologist, the order of which depends on what he is currently working on! His web site is at http://www.ibgames.net/alan/index.html.
Past issues of Winding Down can be found at http://www.ibgames.net/alan/winding/index.html.
If you have any questions or comments about the articles on my web site, click here to send me email.