Reading “Blown to Bits” by Harry Lewis, Ken Ledeen, Hal Abelson, and Wendy Seltzer.

Everything we are discussing now consists of bits. Bits are uniform, and for machines they make no sense, as machines do not have “sense”. But people assign meaning to different forms of bits, and represent images, audio, sculpture, texts, and many other things.

People are used to apply to those bits the reasoning originally aimed at various objects of the physical world, and have historically justified pre-suppositions fit for the objects, but not necessarily for bits.

When the US government imposed restrictions on exporting of cryptographic algorithms from the USA, Phil Zimmerman printed his program as a book and published it internationally.

It looks very weird not just for people who discovered computers in the 1980s, but even for people used to the 56k modem links of the 1990s to see how cheap bits are nowadays. Digitising the world is fun and profit, everything which can be digitised, will be. Hard drives are big enough to fit everything one wants to fit, and will be kept for as long as someone wants to keep them. Moreover, you do not necessarily know who collects the bits about you. And even you do not want to have data collected about you, you probably have no option to say “no”.

Bits live a long time, if at least someone cares about them. This is both good and bad. Good in that making archives is very easy, bad because the amount of accumulated noise is also very big.

Bits can be transmitted very fast and very far. Expect people to work from home, and your daughters falling in love with guys from Yemen.

Since bits live long and can be accessible to everyone, what if someone publishes something bad about you on the internet? If it is false, you can sue. But what if it is true?

Suppose you cheated on your wife, years ago, and want to forget that case, but somebody remembered, possibly copied your messages to took a photo.

Do you really deserve the public hatred and bad fame after all those years?

When I was young, people were much more privacy-loving than now. Even today, if you ask people if they care about privacy, they usually say “yes”.

But in reality, even if people care, they do not take action to assure that privacy, especially since it is very laborious and expensive. People do not read “privacy policy”, and in many countries even consider them laughable.

People are very ready to lose their personal data for a 1% discount at a shop or the convenience of ordering food over the phone.

Copyright was invented to stimulate the production of knowledge when knowledge was scarce. It gave an artificial monopoly over physical book production to reward authors and protect them from predatory competing publishers. Since that time most “copyright violations” are by end users, and printing books is not really an investment.

DRM refers to tools which artificially limit copying. It is largely a failed effort, but the books discusses it thoroughly.

Conducting fraud is much easier in the age of the Internet.

For a few years I could not log into my bank app, because I am a Caucasian man living in a predominantly Mongoloid country, and the app just crashed on seeing my face. This is not a bug, this is a feature.

Algorithms have bias in them, because biases make sense. If group A has crime rate 16%, and group B has crime rate 10%, by excluding group A the average crime rate will drop. But this is not a solution people are happy with, we say that any particular member of group A is not responsible for all the group.

In the case of neural-networking AI the problem is even worse, because AI methods cannot explain why they make a decision. You fail a job interview, because “AI considers you untrustworthy”. (Real case.)

American law enforcment agencies once wanted to outlaw any uncontrolled encryption, and wanted to set up a “key escrow” system in order to wiretap all Internet communication. They failed because the (1) public key algorithms, (2) Internet commerce industry.

It is always better when the forces of Good are supported by large money.

Bits are only uniform if you can get access to them. In the USA the ISP market is deregulated, but the “invisible hand of the market” did not lead to an increased competition, so internet is slow and expensive.

If they are publishers, then they should be liable if some crappy content appears on their website.

If they are newspapers, they cannot refuse to publish true content, because “freedom of speech”.

How do you solve this issue?

A lot of attention is paid to the fact that legislation of the new technologies uses metaphors and use-case from the past technologies. The laws about Radio from 1920s used books, newspapers, and telegraph as analogies.

New Internet laws referred to Radio and Television.

Search is much more productive than hierarchies in finding good information, but it also makes Google a monopolist and a gate-keeper. (I use Microsoft Bing, by the way.)