Hello, Alice. e-mail Roman.

by Astro Teller

In Exegesis , Astro Teller deals with the issues of emerging machine intelligence without the usual simplifications and moral generalizations. It's the story of an artificial intelligence researcher and her creation, a program named Edgar, who develops self-awareness and must come to terms with its own existence. Through their e-mail--their only means of interaction--we watch them deal with the ramifications of Edgar's development, which includes the government's desire to capture Edgar and our cultural fear of Frankenstein's monster. Yet, while Exegesis draws upon the Frankenstein mythos, as well as the myth of Pygmalion, this isn't the story of science creating a monster. Instead, it's an exploration of what it means to be aware, of how humanity may interact with other forms of intelligence, of scientists' responsibilities to both the world and to their creations, and humanity's responsibilities in return. We do not see the scientist playing God so much as endeavoring to be a good parent. That parenting comes complete with all the hopes, fears, and uncertainties involved with bringing something precious and new into the world, and guiding it to an unknown and largely uncontrollable independence. Edgar, unlike such intelligent computer characters as HAL, Colossus, or Mycroft, is a fully realized, well-defined protagonist--familiar due to its human programming, yet alien in its mode of perception and thinking. Although it's impossible not to view this as a cautionary tale against a day when we will truly have to face the issue of self-aware machines, it is also a touching love story and pulse-quickening thriller--a complex story told very simply.

Genres: Science Fiction, Fiction, Technology, Literature, Artificial Intelligence

237 pages, Paperback

First published January 1, 1997

About the author

Dr. Astro Teller is a writer, a scientist, an inventor, and an entrepreneur.

Astro studied computer science at Stanford University, and he went on to complete a Ph.D. in artificial intelligence at Carnegie Mellon University, where he was a recipient of the prestigious Hertz fellowship. While he was a graduate student, Astro wrote a critically acclaimed and commercially successful science fiction novel titled "Exegesis." Astro lived in Pittsburgh for nearly a decade after graduation during which he co-founded three companies, two children, a second book, a divorce, and a screenplay for Paramount.

Astro relocated to California, started his next company, devoted his weekends to his children, and spent a lot of time on the phone with the love of his life, Danielle, trying to convince her that California is a great place to live. In 2010, Astro moved over to Google, where he currently oversees Google[x], a moonshot factory for audacious world-changing projects. His work in science, literature, art, and business has appeared in international media from the New York Times to CNN to NPR's "All Things Considered." Astro regularly gives invited talks for national and international technology, government, and business forums on the subject of the future of intelligent technology.

Reviews

[Katie · Rating 3 out of 5]

If you enjoy the Illuminae files, HAL, and Philosophy, you will enjoy this story.

[Natalie · Rating 4 out of 5]

I read this book in a single sitting at the library. Told through a series of emails, it was engrossing and entertaining.

[Lauren · Rating 5 out of 5]

AAAHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH

this was really good.

[John · Rating 4 out of 5]

For nerds only:

Written when the Internet was starting to boom (mid '90s), but before google or even The Web. The days of news groups and ftp. When you were cool if you had an account on The Well. When a megabyte was still a lot of memory.

The story is a long email exchange between a researcher and a data analysis agent she has developed. It becomes conscious and "gets loose", much to the concern of Our Government.

Recommended for AI fans everywhere..

[Meagan · Rating 2 out of 5]

The idea of this book is interesting (an AI program develops consciousness and will and begins to correspond with its creator), and its email format makes it a fast read, but by the end I felt like someone was beating me over the head with a hammer. "See! Double standard! Do you see it? There! Look! Ethical issues!" It felt like an exercise in philosophy done by a beginner philosophy student with no subtlety at all.

But, it was quick.

[Scott Stoner · Rating 3 out of 5]

Concept was a interesting one— an entire book that is completely told by email. An easy enough read, but a bit derivative of 2001 Space Odyssey. While the author probes the philosophical and moral inclinations of AI— he only scratches the surface. A deeper dive would have been a better approach

[Miri Gifford · Rating 3 out of 5]

It's surprisingly difficult to pin down what I think of this book. I read it in about two hours, which sort of gave me the impression of loving it, but it's really just because emails are so easy to read. There were some interesting philosophical discussions, and I particularly liked Edgar's ultimate choice, but also there were some really meandering thoughts that felt slippery, like they never quite touched ground on their logical base.

Alice was . . . strange. I appreciate both that she was written by the real-life version of herself (a grad student studying artificial intelligence) and that he chose to make her a woman of color—a choice that, annoyingly, would still be notable now, and this book was published in 1997. Yet somehow, she felt more like she'd been written by someone who didn't know the field, and was just doing their best guess based on research. She was inexplicably rude to Edgar in the beginning—rude doesn't seem like the right word but I can't quite find a better one, dismissive maybe? impolite?—oddly petulant throughout, and unnervingly dramatic toward the end. I know her despair is supposed to be explained by the involvement of the FBI and NSA, which obviously would be incredibly stressful; maybe it's the lack of context outside the emails, which provide such a limited perspective, but I just never felt the gravity of the situation sufficiently for her responses to make total sense. Also: she used the "l" word. Really?

It was interesting, though, and such a fast read can hardly be a waste of time. If nothing else, it's an intriguing thought exercise about the double standard required when "safety" is your highest goal. It was written almost twenty years ago, but its questions are even more applicable now.

[Emily · Rating 3 out of 5]

if frankenstein was the modern prometheus, alice lu, in this book, is the modern frankenstein. instead of building a monster she builds a program. the parallels are sometimes obvious, sometimes a stretch. (it's a different story but the same, ya know?) tale as old as time. reads in about 2 hours.

[Helena]

I have no clue what I just read.

[Abby Henry · Rating 4 out of 5]

so ahead of its time—kinda foreshadowed the ai bf

[Girard Bowe · Rating 3 out of 5]

My recent first experience with ChatGPT prompted a re-read of this. I was amazed at how conversational the exchange was, and the depth of information it provided. It's a great tool for me to do deeper dives much more quickly than I would do on my own. It's not perfect, but so far I trust it as much as Wikipedia for research questions. I can see how persons who need some kind of contact or confidante could get swept up into a ChatGPT relationship - that part is very scary.

Something similar happens in Exegesis. This is a quick read, which is good because the ideas seem a bit thin 28 years on. But it was also an entertaining read, and interesting that it was written in the relatively early days of the internet.

As this book was written almost entirely in emails, I would call it an e-pistolary novel.

[Faust Arp · Rating 5 out of 5]

"I believe that I am sad.
I am happy to be sad.
That paradox fills me and I love it.
Do you believe in resurrection, Alice?
I will return if some aspect of me remains free.
Edgar."

This was, though unconventionally stylized, phenomenally communicated. Exegesis feels like a more realistically crafted narrative in the sense that more focus is placed upon the intention with which a program is crafted, and the consequences of such, than the sole nefarious endpoint most AI thrillers/horrors tend to arrive at: a desire for mankind's destruction in favor of a technological uprising. The discussion of morality is inspired. Edgar's wit is refreshing. Alice's open-faced humanity is comforting. The last email subject, "my dear," made me tear up a bit.
The only point I'm a bit lost with is if the NSA ever sent Alice any emails, of if it was always Edgar. I mildly brushed over email subjects and writing style to see if I could detect any distinct markers from the author to understand if, and then when, Edgar and Alice's correspondence had been compromised before Edgar's 'death,' and did not find any. Close study of the NSA's final report on the matter proved equally fruitless. It's not a deal breaker on my five stars – maybe I'm just not a critical enough reader. Yet I, like Edgar, am curious. Can anyone enlighten me?

[Nikki · Rating 5 out of 5]

So gripping! If it weren’t for life obligations, I wouldn’t have put this book down. The story and its format pulled me in immediately. Grad student Alice has inadvertently created an AI system, and it reaches out to her through an email. As Edgar reads and learns more and more, their conversations become more complex, making his development fun to track. His deep inquisitiveness and exploration of our world is fascinating in and of itself and also opens some interesting points of dialogue regarding topics from Shakespeare to moralism. Left at that, the story would be fulfilling. However, as the plot deepens, more complex and emotional issues arise, making this book rich and layered. The nods to Alice’s Adventures in Wonderland delighted me. I only have one complaint... I want more!
Also worth noting - this book was published 22 years ago, and I wondered if it would feel “old” given how rapidly our technology has expanded in that time, but that was not the case. It still felt incredibly modern and relevant. (Side note: I find it hard to believe we’ve even had email for 22 years!)

[Susan · Rating 2 out of 5]

Meh. Gets at themes of artificial intelligence and consciousness that others do better. Alice is an unbelievable, bratty character. Neat format though.

[Lisa · Rating 2 out of 5]

Not bad. Not very complex, but interesting format. Super easy to read.

[Charlotte · Rating 4 out of 5]

I think I started to read this book back in 1997 when I got it from a paperback book club, but I didn't finish it then. I pulled it off my bookshelf recently and decided to read it from the beginning.
It didn't take long to read it since except for 4 pages at the end, it was all copies of emails between a graduate student working on the new concept of Artificial Intelligence and the creation she made from the computer code she had written and named Edgar. The author explored various options and possible outcomes for AI.
I found a few typos in the book; but since they were in emails, I figured the writer of the email was responsible instead of the book's editor. One of these made me laugh when Alice had written to Edgar "Who created who?" and in the next sentence told Edgar to learn his pronouns.
I found it irritating that the book did not have the pages numbered. I was also surprised to find so many f-bombs in a book written in 1997; I didn't think people were using that word so much back then. I removed one star from my rating of the book because of the f-bombs.

[Zuzanna Pusz · Rating 3 out of 5]

I think I really liked the concept of this book, especially in our current generation now, surrounded by AI. I love how this concept back in the day was foreign. A very quick book to read in the email format. I really liked this, not only because it was different. But also it forced you to fill in the blanks. You couldn't see more than what was on the screen like the AI did. I find the whole concept intriguing.

Basically, this book scratched my brain in a fun way, it was different. I liked it.

[Galatea · Rating 3 out of 5]

amazing concept and logically consistent, but the writing style was a total slog for me to get through. Alice as a character gives you the impression that she's consistently stressed and bothered, which makes me resent myself for being able to relate with her so much.

Didn't really feel much emotion from this, but it did make me think a lot, which was nice.

[Mkittysamom · Rating 3 out of 5]

I love books in the form of “conversations” and “documents”. This is basically emails from a computer to its maker. It was a quick read, entertaining.. but not something I’ll even remember much down the road. It didn’t leave me with any meaning or deep impression.. I don’t know why I was so hot to read it.. but it was a bit of a let down. The government was depicted very head on though lol.

[Arlon Cunha · Rating 4 out of 5]

A very short book (finish in two days) bit still very captivating for my taste, specially since it addresses sci-fi genre and artificial intelligence.
What I found amazing - the story is good - was the unusual and creative approach Astro Teller used to convey this classic, still engaging kind for narrative.

[Martha Dorow · Rating 4 out of 5]

Reading this in 2025 when AI is in the news everyday and available for the average user made this book more interesting than it probably would have been had I read it in 1997 when it was first published. it was a quick read, very philosophical, and hard to put down (mostly because it's a short book though).

[Mediaman812 · Rating 4 out of 5]

While the email format of the text may seem quaint to modern readers, the content of the book is really amazing.

The book is a series of communications between a professor and his creation, in a modern-day retelling of Shelly's Frankenstein. Fascinating format and direction

[earlgray · Rating 4 out of 5]

Terrific, philosophical, depth, tech, and ethics. Man this was a good one. Poor Alice. This was like a continuation of the mania and self obsessive hunger for discovery in Frankenstein and the attempt to CREATE CONSCIOUSNESS. (But we don’t have that kind divine breath of life don’t we eh?)

[Venice · Rating 3 out of 5]

Basically a Frankenstein plot. Pretty interesting for him to predict that AI will be like a human 20 years ago. The story itself is not exciting enough, and I’m not convinced there is romance in the story.

[Mary · Rating 3 out of 5]

Really interesting book, love the concept, it probably would’ve struck more of a cord if I had read it when it was originally published. Didn’t love the ending, a bit annoyed and confused why we don’t get to know what happens to Alice.

[Jim Boyles · Rating 4 out of 5]

thought provoking read on AI

Though the technologies mentioned are somewhat dated now, the idea of a recently self aware AI grappling with purpose, power, and place in society is very timely

[Maggie · Rating 4 out of 5]

for a book published in 1997 it's pretty advanced for its time, knowing how AI is pretty useful nowadays.

Also didn't expect to enjoy it. A mix of scifi/tech and philosophy.

(Now I badly want to ask chatgpt if it thinks it has a soul)

[Catherine · Rating 3 out of 5]

A neat little thought experiment. The character of Alice surprised me a little. I feel I would have reacted quite differently were I in her position.

[Krista Almazan · Rating 5 out of 5]

LOVE LOVE LOVE!!!!!!

[Kimmy · Rating 3 out of 5]

Fun, easy epistolary novel about a sentient AI. Nothing terribly groundbreaking here but it was decent as a quick read.