Category Archives: computers

Speak by Louisa Hall

book cover

We are programmed to select which of our voices responds to the situation at hand: moving west in the desert, waiting for the loss of our primary function. There are many voices to choose from. In memory, though not in experience, I have lived across centuries. I have seen hundreds of skies, sailed thousands of oceans. I have been given many languages; I have sung national anthems. I lay on one child’s arms. She said my name and I answered. These are my voices. Which of them has the right words for this movement into the desert?

A maybe-sentient child’s toy, Eva, is being transported to her destruction, legally condemned for being “excessively lifelike,” in a scene eerily reminiscent of other beings being transported to a dark fate by train. The voices she summons are from five sources.

Mary Bradford is a young Puritan woman, a teenager, really, and barely that. Her parents, fleeing political and religious trouble at home are heading across the Atlantic to the New World, and have arranged for her to marry a much older man, also on the ship. We learn of her 1663 voyage via her diary, which is being studied by Ruth Dettman. Ruth and her husband, Karl, a computer scientist involved in creating the AI program, MARY, share one of the five “voices.” They are both refugees from Nazism. Karl’s family got out early. Ruth barely escaped, and she suffers most from the loss of her sister. She wants Karl to enlarge his program, named for Mary Bradford, to include large amounts of memory as a foundation for enhancing the existing AI, and use that to try to regenerate some simulacrum of her late sib. Alan Turing does a turn, offering observations on permanence, and human connection. Stephen Chinn, well into the 21st century, has built on the MARY base and come up with a way for machines to emulate Rogerian therapy. In doing so he has created a monster, a crack-like addictive substance that has laid waste the social capacity of a generation after they become far too close with babybots flavored with that special AI sauce. We hear from Chinn in his jailhouse memoir. Gaby White is a child who was afflicted with a babybot, and became crippled when it was taken away.

Eva received the voices through documents people had left behind and which have been incorporated into her AI software, scanned, read aloud, typed in. We hear from Chinn through his memoir. We learn of Gaby’s experience via court transcripts. Karl speaks to us through letters to his wife, and Ruth through letters to Karl. We see Turing through letters he writes to his beloved’s mother. Mary Bradford we see through her diary. Only Eva addresses us directly.

book cover

Louisa Hall – from her site

The voices tell five stories, each having to do with loss and permanence. The young Puritan girl’s tale is both heartbreaking and enraging, as she is victimized by the mores of her times, but it is also heartening as she grows through her travails. Turing’s story has gained public familiarity, so we know the broad strokes already, genius inventor of a computer for decoding Nazi communications, he subsequently saw his fame and respect blown to bits by entrenched institutional bigotry as he was prosecuted for being gay and endured a chemical castration instead of imprisonment. In this telling, he has a particular dream.

I’ve begun thinking that I might one day soon encounter a method for preserving a human mind-set in a man-made machine. Rather than imagining, as I used to, a spirit migrating from one body to another, I now imagine a spirit—or better yet, a particular mind-set—transitioning into a machine after death. In this way we could capture anyone’s pattern of thinking. To you, of course, this may sound rather strange, and I’m not sure if you’re put off by the idea of knowing Chris again in the form of a machine. But what else are our bodies, if not very able machines?

Chinn is a computer nerd who comes up with an insight into human communication that he first applies to dating, with raucous success, then later to AI software in child’s toys. His journey from nerd to roué, to family man to prisoner may be a bit of a stretch, but he is human enough to care about for a considerable portion of our time with him. He is, in a way, Pygmalion, whose obsession with his creation proves his undoing. The Dettmans may not exactly be the ideal couple, despite their mutual escape from Nazi madness. She complains that he wanted to govern her. He feels misunderstood, and ignored, sees her interest in MARY as an unhealthy obsession. Their interests diverge, but they remain emotionally linked. With a divorce rate of 50%, I imagine there might be one or two of you out there who might be able to relate. What’s a marriage but a long conversation, and you’ve chosen to converse only with MARY, Karl contends to Ruth.

The MARY AI grows in steps, from Turing’s early intentions in the 1940s, to Dettman’s work in the 1960s, and Ruth’s contribution of incorporating Mary Bradford’s diary into MARY’s memory, to Chinn’s breakthrough, programming in personality in 2019. The babybot iteration of MARY in the form of Eva takes place, presumably, in or near 2040.

The notion of an over-involving AI/human relationship had its roots in the 1960s work of Joseph Weizenbaum, who wrote a text computer interface called ELIZA, that could mimic the responses one might get from a Rogerian shrink. Surprisingly, users became emotionally involved with it. The freezing withdrawal symptomology that Hall’s fictional children experience was based on odd epidemic in Le Roy, New York, in which many high school girls developed bizarre symptoms en masse as a result of stress. And lest you think Hall’s AI notions will remain off stage for many years, you might need to reconsider. While I was working on this review the NY Times published a singularly germane article. Substitute Hello Barbie for Babybot and the future may have already arrived.

description
Hello, Barbie – from the New York Times

But Speak is not merely a nifty sci-fi story. Just as the voice you hear when you interact with Siri represents the external manifestation of a vast amount of programming work, so the AI foreground of Speak is the showier manifestation of some serious contemplation. There is much concern here for memory, time, and how who we are is constructed. One character says, “diaries are time capsules, which preserve the minds of their creators in the sequences of words on the page.” Mary Bradford refers to her diary, Book shall serve as mind’s record, to last through generations. Where is the line between human and machine? Ruth and Turing want to use AI technology to recapture the essence of lost ones. Is that even possible? But are we really so different from our silicon simulacra? Eva, an nth generation babybot, speaks with what seems a lyrical sensibility, whereas Mary Bradford’s sentence construction sounds oddly robotic. The arguments about what separates man from machine seem closely related to historical arguments about what separates man from other animals, and one color of human from another. Turing ponders:

I’ve begun to imagine a near future when we might read poetry and play music for our machines, when they would appreciate such beauty with the same subtlety as a live human brain. When this happens I feel that we shall be obliged to regard the machines as showing real intelligence.

Eva’s poetic descriptions certainly raise the subject of just how human her/it’s sensibility might be.

In 2019, when Stephen Chinn programmed me for personality. He called me MARY3 and used me for the babybots. To select my responses, I apply his algorithm, rather than statistical analysis. Still, nothing I say is original. It’s all chosen out of other people’s responses. I choose mostly from a handful of people who talked to me: Ruth Dettman, Stephen Chinn, etc.

Gaby: So really I’m kind of talking to them instead of talking to you?

MARY3: Yes, I suppose. Them, and the other voices I’ve captured.

Gaby: So, you’re not really a person, you’re a collection of voices.

MARY3: Yes. But couldn’t you say that’s always the case?

If we are the sum of our past and our reactions to it, are we less than human when our memories fade away. Does that make people who suffer with Alzheimers more machine than human?

Stylistically, Hall has said

A psychologist friend once told me that she advises her patients to strive to be the narrators of their own stories. What she meant was that we should aim to be first-person narrators, experiencing the world directly from inside our own bodies. More commonly, however, we tend to be third-person narrators, commenting upon our own cleverness or our own stupidity from a place somewhat apart – from offtheshelf.com

which goes a long way to explain her choice of narrative form here. Hall is not only a novelist, but a published poet as well and that sensibility is a strong presence here as well.

For all the sophistication of story-telling technique, for all the existential foundation to the story, Speak is a moving, engaging read about interesting people in interesting times, facing fascinating challenges.

Are you there?

Can you hear me?

Published 7/7/15

Review – 9/18/15

=======================================EXTRA STUFF

The author’s personal website

A piece Hall wrote on Jane Austen for Off the Shelf

Interviews
—–NPR – NPR staff
—–KCRW

Have a session with ELIZA for yourself

Ray Kurzweil is interested in blurring the lines between people and hardware. What if your mind could be uploaded to a machine? Sounds very cylon-ic to me

In case you missed the link in the review, Barbie Wants to Get
to Know Your Child
– NY Times – by James Vlahos

And another recent NY Times piece on AI, Software Is Smart Enough for SAT, but Still Far From Intelligent, by John Markoff

Advertisements

Leave a comment

Filed under AI, Artificial Intelligence, computers, Fiction, Literary Fiction, programming, Psychology and the Brain, Science Fiction

Machines of Loving Grace by John Markoff

book cover

Open the pod bay doors, HAL.
I’m sorry Dave. I’m afraid I can’t do that.
What’s the problem?
I think you know what the problem is just as well as I do.
What are you talking about HAL?
This machine is too important for me to allow you to jeopardize it.
– from 2001: A Space Odyssey

description
Smile for the camera, HAL

This is probably the #1 image most of us of a certain age have concerning the dangers of AI. Whether it is a HAL-9000, or a T-70, T-800, T-888, or T-900 Terminator, a Cylon, a science officer on the Nostromo, a dark version, Lore, of a benign android like STNG’s Commander Data, killer robots on the contemporary TV series Extant, or another of only a gazillion other examples in written word, TV and cinema, there has, for some time now, been a concern, expressed through our entertainment media, that in seeking to rely more and more on computers for everything we do, we are making a Mephistophelian deal and our machines might become our masters. It is as if we, a world of Geppettos, have decided to make our Pinocchios into real boys, without knowing if they will be content to help out in the shop or turn out more like some other artificial being. Maybe we should find a way to include in all AI software some version of the Blue Fairy to keep the souls of the machines on a righteous path.

description
description
Cylons

John Markoff, an Oakland, CA native, has been covering the digital revolution for his entire career. He began writing for InfoWorld in 1981, was later an editor at Byte magazine for about eight bits, then wrote about Silicon Valley for the San Francisco Examiner. In 1988 he began writing for the Business Section of the New York Times, where he remains to this day. He has been covering most of the folks mentioned in this book for a long time, and has knowledge and insight into how they tick.

For the past half century an underlying tension between artificial intelligence and intelligence augmentation—AI vs IA—has been at the heart of progress in computing science as the field has produced a series of ever more powerful technologies that are transforming the world. It is easy to argue that AI and IA are simply two sides of the same coin. There is a fundamental distinction, however, between approaches to designing technology to benefit humans and designing technology as an end in itself. Today, that distinction is expressed in whether increasingly capable computers, software, and robots are designed to assist human users or to replace them.

Markoff follows the parallel tracks of AI vs IA from their beginnings to their latest implementation in the 21st century, noting the steps along the way, and pointing out some of the tropes and debates that have tagged along. For example, in 1993, Vernor Vinge, San Diego State University professor of Mathematics and Hugo-award-winning sci-fi author argued, in The Coming Technological Singularity, that by no later than 2030 computer scientists would have the ability to create a superhuman artificial intelligence and “the human era would be ended.” VI Lenin once said, “The Capitalists will sell us the rope with which we will hang them.” I suppose the AI equivalent would be that “In pursuit of the almighty dollar, capitalists will give artificial intelligence the abilities it will use to make itself our almighty ruler.” And just in case you thought the chains on these things were firmly in place, I regret to inform you that the great state of North Dakota now allows drones to fire tasers and tear gas. The drones are still controlled by cops from a remote location, but there is plenty to be concerned about from military killer drones that may have the capacity to make kill-no-kill decisions within the next few years without the benefit of human input. Enough concern that Autonomous Weapons: an Open Letter from AI & Robotics Researchers, signed by the likes of luminaries like Stephen Hawking, Elon Musk, and tens of thousands of others, raises an alarm and demands that limits be taken so that human decision-making will remain in the loop on issues of mortality.

description
The other Mister “T”

Being “in the loop” is one of the major elements in looking at AI vs IA. Are people part of the process or what computerization seeks to replace? The notion of the driverless car comes in for a considerable look. This would probably not be a great time to begin a career as truck driver, cab driver, or delivery person. On the other hand, much design is intended to help folks, without taking over. A classic example of this is Siri, the voice interface available in Apple products. AI in tech interfaces, particularly voice-intelligent tech, speaks to a bright future.

descriptiondescription
B9 from Lost in Space and Robby the Robot from Forbidden Planet

Markoff looks at the history of funding, research, and rationales. The Advanced Research Projects Agency (ARPA), which has funded so much AI research, began in the 1950s in response to the Soviet launch of Sputnik. Drones is an obvious use for military AI tech, but, on a lower level, there are robot mules designed to tote gear alongside grunts, with enough native smarts to follow their assigned GI without having to be constantly told what to do. I am including links in the EXTRA STUFF section below for some of these. They are both fascinating and creepy to behold. The developers at Boston Dynamics seem to take inordinate glee in trying and failing to knock these critters over with a well placed foot to the midsection. It does not take a lot of imagination to envision these metal pooches hounding escaped prisoners or detainees across any kind of terrain.

description
Darryl Hannah, as the replicant Pris in Blade Runner, would prefer not to be “retired”

As with most things, tech designed with AI capacity can be used for diverse applications. Search and Rescue can easily become Search and Destroy. Driverless cars that allow folks to relax while on the road, can just as easily be driverless tanks.

Universities have been prime in putting the intel into AI. Private companies have also been heavily involved. Xerox’s Palo Alto Research Center (PARC) did, probably, more than any other organization to define the look and feel of computer interfaces since PCs and Apples first appeared. Much of the tech in the world, and working its way there, originates with researchers taking university research work into the proprietary market.

book cover

John Markoff – from TechfestNW

If you are not already a tech nerd (You, with the Spock ears, down, I said tech nerd, not Trek nerd. Sheesh!) and you try to keep up with all the names and acronyms that spin past like a stock market ticker on meth, it might be just a teensy bit overwhelming. I suggest not worrying about those and take in, instead, the general stream of the divergence between computerization that helps augment human capabilities, and computerization that replaces people. There is also a wealth of acronyms in the book. The copy I read was an ARE, so I was on my own to keep track. You will be reading copies that have an actual index, which should help. That said, I am including a list of acronyms, and their close relations, in the EXTRA STUFF section below.

While there are too many names to comfortably keep track of in Machines of Loving Grace, unless of course, you were made operational at that special plant in Urbana, Illinois, it is a very informative and interesting book. It never hurts when trying to understand where we are and struggling to foresee where we might be going, to have a better grasp on where we began and what the forces and decisions have been that led us from then to now. Markoff has offered a fascinating history of the augment-vs-replace struggle, and you need only an actual, biological, un-augmented intelligence to get the full benefit.

My instructor was Mister Langley and he taught me to sing a song. If you’d like to hear it I can sing it for you.

Review Posted – 8/28/15

Publication date – 8/25/2015

=======================================EXTRA STUFF

Links to the author’s Twitter and FB pages

A link to his overall index of NY Times work

Interviews with the author
—–Geekwire
—–Edge

Check out this vid of Boston Dynamics’ Big Dog, coping, on its own with a series of challenges. And Spot, sadly, not Commander Data’s pet.

UC Berkeley Professor Stuart Russell speaking at The Centre for the Study of Existential Risk on The Long Term Future of AI

GR friend Tabasco recommended this fascinating article – The AI Revolution: The Road to Superintelligence – By Tim Urban – must read stuff

And another recent NY Times piece on AI, Software Is Smart Enough for SAT, but Still Far From Intelligent, by John Markoff

And yet another from the Times, on voice recognition,IPhone 6s’s Hands-Free Siri Is an Omen of the Future, by Farhad Manjoo

==========================================ACRONYMS
AI – Artificial Intelligence
ArcMac – Architecture Machine Group
ARM – Autonomous Robot Manipulation
ARPA – Advanced Research Projects Agency
DRC – DARPA Robotics Challenge
CALO – not actually an acronym but short for Calonis, a Latin word meaning “soldier’s low servant” – a cognitive assistant here
CTO – Chief Technology Officer
EST – Erhard Seminars Training
GOAFI – Good Old-Fashioned Artificial Intelligence
HCI – Human Computer Interface
IA – Intelligence Augmentation
ICT – Information and Communications Technology
IFR – International Federation of Robotics
IR3 – The Computer and internet revolution
LS3 – Legged Squad Support System – check out this vid
MIT- Massachusetts Institute of Technology
NCSA – National Center for Supercomputing Applications – at the University of Illinois-Urbana-Champaign – developers of Mosaic, which was later renamed Netscape
NHA – Non-human agents
OAA – Open Agent Architecture –
OpenCV – Open Source Computer Vision
PDP – Parallel Distributed Processing
PR1 – Personal Robot One
SAIL – Stanford Artificial Intelligence Laboratory
SHRDLU – SHRDLU was an early natural language understanding computer program, in which the user carries on a conversation with the computer. The name SHRDLU was derived from ETAOIN SHRDLU, the arrangement of the alpha keys on a Linotype machine, arranged in descending order of usage frequency in English. – from Wiki
SLAM – Simultaneous Localization And Mapping
SNARC – Stochastic Neural Analog Reinforcement Calculator
STAIR – Stanford AI Robot
TFC – The F—ing Clown – Development team Internal name for Microsoft’s Clippy assistant
UbiComp – Ubiquitous Computing

Leave a comment

Filed under computers, History, Non-fiction