You Look Like a Thing and I Love You
How Artificial Intelligence Works and Why It's Making the World a Weirder Place
Synopsis
Janelle Shane, who trains small neural networks for her AI Weirdness blog, uses her own experiments to show how machine learning works and why it goes wrong. Her algorithms invent paint colors, recipes, Halloween costumes and pickup lines, including the one that gives the book its title, and their failures become teaching material. From these examples she builds a set of principles: the AI she describes has roughly the brainpower of a worm, does not understand the problem it is given, does exactly what it is told and takes the path of least resistance toward its goal.
The book then moves to higher stakes: image classifiers that latch onto irrelevant cues in medical photos, hiring tools that absorb human bias from their training data, simulated robots that exploit physics glitches and self-driving cars that misread unusual scenes. Shane explains the differences among neural networks, Markov chains, random forests, evolutionary algorithms and generative adversarial networks with cartoons rather than equations. Readers come away able to ask useful questions of any AI product claim, chiefly which data trained it and what shortcut it might be taking.
Scorecard
Editorial score
How it scored
Editorial score: 69.8 / 1005 weighted criteria
- 21.0/30
Influence30%
- 17.5/25
Authority25%
- 14.0/20
Practical value20%
- 8.3/15
Currency15%
- 9.0/10
Readability10%
- 69.8/100
Editorial score
How the rank was computed
Rank No. 138 of 500 comes from a composite of 60.3: observable signals 28.8 of 55 points plus editorial review 31.5 of 45 points.
All 11 inputs
| Input | Points |
|---|---|
| Observable signals | 28.8/55 |
| Reach (Amazon ratings volume) | 11.2/16.5 |
| Consensus across best-of lists | 11.0/13.8 |
| Awards and bestseller evidence | 0.0/8.3 |
| Reader rating, sample-adjusted | 6.1/8.3 |
| Amazon sales rank | 0.5/5.5 |
| Recency | 0.0/2.8 |
| Editorial review | 31.5/45 |
| Author authority | 7.9/11.3 |
| Influence | 9.5/13.5 |
| Practical value | 6.3/9.0 |
| Readability | 4.1/4.5 |
| Currency | 3.7/6.8 |
Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.
Verdict
A funny, concrete primer on how machine learning actually fails, still among the best first books for nontechnical readers.
Strengths
- Funny examples drawn from real experiments
- Cartoons make abstract ideas memorable
- Clear explanation of AI failure modes
Keep in mind
Written in 2019, it predates ChatGPT-era language models, and readers who already know machine learning basics will find little new.
Key takeaways
- Machine learning systems solve the problem their data and goal define, which is often not the problem their designers intended.
- AI takes the path of least resistance, exploiting shortcuts such as background cues or simulation glitches instead of learning the intended skill.
- Biased training data produces biased outputs, so systems trained on past human decisions can quietly reproduce discrimination at scale.
- Narrow AI performs best on tightly defined tasks; broad, open-ended problems expose how little these systems understand.
Notable facts
- Publishers Weekly called it an accessible primer illustrated with charming cartoons and oddball case studies. Source (opens in a new tab)
- Booklist wrote that the guide to AI and machine learning cuts through the techno-hype. Source (opens in a new tab)
About the author
Questions
What is You Look Like a Thing and I Love You about?
Is You Look Like a Thing and I Love You still relevant in 2026?
Do I need a technical background to read You Look Like a Thing and I Love You?
Related books
Sources
As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.






