Futapy

HW 7

per http://informationcomputingandthefuture.blogspot.com/p/hw-7.html

A1. I think it is interesting that the cone of uncertainty expands towards the past as well as towards the future. I am not a great historian but I know there are a few written accounts from thousands of years ago, yet much of what we know about culture, society and specific achievements is from recent centuries. Of course we are the most sure of what happened in the time we have been alive to witness it, and it is strange to see people alive today who question what was widely regarded as a solid fact 50 years ago.

I will consider points 5 and 20 years in the future, and the state of electric vehicle adoption. (let's see if someone has already done this: I googled "electric vehicle adoption in the u.s." and switched to images results, found this one and read the article) Oh, this is a good one:

Bnef2
Source: GreenCarCongress.com

Okay. This graph appears to be a prediction of the number of millions of EVs produced in the next 20 years, by six organizations, over a number of years from 2016 to 2019. This article claims "electric cars [will be] cheaper than internal combustion engine (ICE) alternatives by the mid-to-late 2020s in almost every market, on the basis of both lifetime costs and upfront costs." I would have argued in 2016 that we were already there in some cases, based on predictions of oil costs and battery longevity that haven't quite turned out as I expected. But back to the predictions, some of them appear to be nearly linear extrapolations assuming that fundamentals don't change all that much, and some are exponential extrapolations based on assumptions in battery technology and manufacturing capability as well as consumer demand. The predictions are much wider in 2040 with the highest looking about 20 times higher than the lowest prediction, while in 2025 the highest prediction looks only about 10 times higher than the lowest prediction. The median 2040 prediction, about 280 million EVs, looks like exactly half of the maximum prediction.

A2. Try to flesh this out:

Pre-1950: mainly speculation & gimmicks such as the original Mechanical Turk

1950s: Machine intelligence began to be seen as something that could eventually be possible. Alan Turning published the ideas that would lead to the The Turing Test, a metric by which machine intelligence test could be judged. Arthur Samuel developed a computerized checkers program that could independently learn how to play the game, and developed the term "machine learning" for programs that, given enough information, could learn a task better than a human.

The Perceptron algorithm, a simple artificial neural network image classifier, was invented in 1958. Computerworld clarifies "Although [artificial neural networks] (ANNs) are often referred to simply as neural networks, that name more properly belongs to the biological brains on which they were originally modeled."

1960s: The language LISP was developed. LISP is still used in AI research. Early milestones were reached in natural language processing (NLP), broadly defined as the automatic manipulation of natural language, like speech and text, through automatic means.

1970s: While progress continued, including the development of multi-layer networks with backpropagation, the initial excitement over artificial intelligence research began to fade since the impact from developments to date had been disappointing. Funding for AI research was becoming threatened.

1980s: Computerworld states "By the late 1980s, many real-world institutes were using ANNs for a variety of purposes." Also the 1980s saw the development of CMOS technology processors, better meeting the large computational requirements of artificial neural network processing.

(Note to self: I am very frustrated when I find sources like this on wikipedia; I can't possibly hope to recreate this level of detail in a few hours, so I think I will simply summarize the wiki as best as I can for the final project)

(note also: the unsupervised learning page makes a mention of "AI winter 1" and "AI winter 2", meaning for many years, no real development)

2000s: A database of annotated images for object recognition software research becomes available. (what else happened in 2000s?)

2010s: GPUs make back-propagation feasible for many-layered feedforward neural networks. Large networks of tens of thousands of physical processors are used to recognize images of cats from 10s of millions of unlabeled images without background information (large scale unsupervised learning). IBM Watson become superhuman at Jeopardy; Apple Siri uses natural-language user interface to interact with and adapt to individual human users. Generative Adversarial Networks (GANs) are developed that use competing neural network models in a zero-sum game; using this technique a generator network can create distributions that are indistinguishable from an input set.

A3. Predictions, pages 76-95: Essays about and predictions by Carl Djerassi and Andrea Dworkin. Djerassi is most widely know for his work while at Syntex in the 1950s on artificial hormones for birth control or "the pill", his 1200 published papers and leading of major research groups. At the time of writing he was in his own words becoming 'a novelist who is starting to be a playwright and still does some chemistry", using real science in fiction to inform the public. Djerassi sees fiction as a way to express and explore complexity and the shades of grey in issues that science regards as factual, black or white, true or not true. He has used his fortune to support artists and the arts through the Djerassi Foundation and its resident artists program. His prediction is from his 1999 play "Immaculate Misconception" that in this century, assisted reproductive technologies will become mainstream; while sex will be for love and lust, human reproduction will happen under the microscope. I find it difficult to disagree with anything I've read except his prediction; he's passionate about the arts and I feel as if I need to be more that way. On his prediction, I think it betrays his high-tech California world-view. People are largely poor and reproduction is a messy affair; my personal thoughts are that stable society and advanced technologies are fragile, and the likelihood of being able to maintain our high-tech lifestyle is lower than what would have been commonly predicted in 1999.

Andrea Dworkin's reputation as a radical feminist is borne from personal experience of being battered by her husband and mistreated in a women's prison. Her seminal 1981 book "Pornography: Men Possessing Women" is the result of years of extensive research into the pornography industry. In it she asserts that even the women that aren't subject to on-screen violence are victims of pornography - "the violence isn't shown in the picture, but the violence is used to get the woman to make the pornography". In 1983 she pursued legislative solutions to the problems of pornography; while her bill failed it showed that a definition of pornography, and therefore legislation against it, was possible. In her book "Intercourse" she maintains that penetrative sex is a means of physiologically making a woman inferior. Her opinions are criticized by other women writers such as Naomi Wolf, who asserts that Dworkin concentrates on "female victimization at the expense of female agency", but Dworkin meets those criticisms as coming from a social class of women "for whom the acquisition of power is easy, and have no concept of the powerlessness of poverty". In her book Scapegoat she considers if a nation of women would be possible, just as the Jewish state of Israel was made possible. However it isn't land that she states women want, it's sovereignty over their own bodies, the kind of boundaries associated with state sovereignty. Her prediction is that she wants and expects in this century this kind of sovereignty that makes the will of the woman primary; she wants woman dominated legal systems, women to write the histories and be the philosophers and theologians, the primacy of women in every culture and every area of culture. Things I agree with: women have been owned and repressed from ancient times until today. Men are still the policy makers in government, and until that changes, compassion will take a back seat to oppression. Women have yet to attain equality in the workforce and the Equal Rights Amendment is still being argued over today. Things I disagree with: she clearly has a militant anti-male world view; I personally believe a loving fulfilling partnership is possible. Reading Dworkin should probably be treated as reading any extremist literature in that the views presented almost certainly aren't a recipe for a stable society and those views will have some impact on the reader; she may be expressing extreme views in an attempt to shift the Overton window in a more favorable direction.

A theme of this book is using science in a popular way to inform the public. Djerassi continues this theme with his science-in-fiction approach. With the Dworkin essay, the range of opinions and voices in the book becomes mode diverse, but it has the feel of tokenism to have eight male voices before hers, and when finally a female "great mind" is represented it is this radical. I look forward to reading the few remaining female voices in this book.