Recent Readings on AI

The Seven Deadly Sins of AI Predictions

by Rodney Brooks, MIT; founder of Rethink Robotics and iRobot

Mistaken extrapolations, limited imagination, and other common mistakes that distract us from thinking more productively about the future.

https://www.technologyreview.com/s/609048/the-seven-deadly-sins-of-ai-predictions/

Artificial General Intelligence: Why Aren’t We There Yet?

by Gary Marcus, NYU; Geometric Intelligence (acquired by Uber)

https://www.technologyreview.com/video/604429/artificial-general-intelligence-why-arent-we-there-yet/

Artificial Intelligence Is Stuck. Here’s How to Move It Forward.

by Gary Marcus, NYU; Geometric Intelligence (acquired by Uber)

 

The Paleolitic Jarawa Tribe

The Jarawas are an indigenous people of the Andaman Islands in India. They live in parts of South Andaman and Middle Andaman Islands.

0313-for-webtribesmap-180

The Jarawas, who number about 400 and whom one geneticist described as “arguably the most enigmatic people on our planet,” are believed to have migrated from Africa around 50,000 years ago. They are very dark-skinned, small in stature and until 1998 lived in complete cultural isolation, shooting outsiders with steel-tipped arrows if they came too near.

The New York Times has published a series of fascinating articles about the tribe and India’s interactions with it. Reading the articles caused me to pause and reflect on the fact that while I, and others Silicon Valley-ites like me, live to drive humanity into the future, not all human civilizations want technology or will benefit from it.

Prisoner’s Dilemma

Radiolab has a fantastic story about the prisoner’s dilemma called One Good Deed Deserves Another. It was being rebroadcast on the local public radio channel KQED today as part of the episode called The Good Show. Take a listen on their site.

Prisoner’s Dilemma is a classic game theory problem. In it, one tries to figure out whether it is better to cooperate or compete with an opponent when you don’t know (or don’t trust) what the opponent will do.

Try out a prisoner’s dilemma simulation here: http://serendip.brynmawr.edu/bb/pd.html.

On a related note…

Four years ago, Radiolab took their show on the road. I caught a performance in Cupertino. It remains one of my favorite theater experiences ever! Here’s a recording from their Seattle performance.

Indo-European Language Family

ie-migrations
Indo-European migrations. Source: Wikipedia

The two languages I speak fluently, English and Hindi, are both Indo-European languages. Both originated from the Proto-Indo-European (PIE) language that was spoken by a Bronze-Age civilization in central Asia around 3500 BC.

Today, 46% of the human population speaks an Indo-European language natively, by far the highest of any language family. There are about 445 such languages currently in use, according to the estimate by Ethnologue, with over two-thirds (313) of them belonging to the Indo-Iranian branch. The most widely spoken Indo-European languages by native speakers are Spanish, English, Hindustani (Hindi and Urdu), Portuguese, Bengali, Russian, Persian and Punjabi, each with over 100 million speakers.

 

51233rcinbl-_sl300_In The Story of Human Language,a fantastic lecture series on the evolution of language, the linguist John McWhorter describes the discovery of the Indo-European language family.

In 1786, William Jones, a British jurist and Orientalist, presented an address to the Bengal Asiatic Society in which he observed:

“The Sanskrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either, yet bearing to both of them a stronger affinity, both in the roots of verbs, and in the forms of grammar, than could possibly have been produced by accident; so strong, indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists.”

Jones was making the first official observation of the fact that groups of languages develop from single ones; that is, he inaugurated the study of the natural history of language.

The kind of “affinity” he referred to involved not only word roots in common among Sanskrit, Latin, and Greek but also aspects of grammar. For example, even the case endings on nouns in these languages are clearly related: tooth in four cases in the languages William Jones referred to:

            SANSKRIT   GREEK    LATIN 
nominative       dán   odón     dēns 
genitive       datás   odóntos  dentis 
dative          daté   odónti   dentī 
accusative    dántam   odónta   dentem

To see the similarity among modern descendants of PIE, take a look their word for “tooth” in these Indo-European languages.

Dutch      tand
German     zahn
Danish     tand
Icelandic  tönn
Welsh      dant ‎
Latin      dēns
Lithuanian dantìs
Armenian   ‎atam   (ատամ)
Persian    ‎dandân (دندان)
Sanskrit  ‎ dát    (दत्)
Hindi      dānt   (दांत)

500px-indo-european_branches_map-svg

A map showing the approximate present-day distribution of the Indo-European branches within their homelands of Europe and Asia. [Link to the legend on Wikipedia]

ietreecentum1
Indo-European Language Tree (Part 1) [source
ietreesatem1
Indo-European Language Tree (Part 2) [source]