Smith breaks the illusion that AI understands anything, showing it's just pattern matching that fails at basic common sense.
The AI Delusion: Reader Reviews & Ratings
Loading…
Write a Review
Your Review
0/500
Other reviews
17 total
Smith effectively exposes how brute-force computation tricks us into thinking machines understand, exposing real dangers in blind trust.
Gary Smith brilliantly exposes how AI fails at common sense, like confusing a cart for a store, proving it just matches pixels instead of understanding the world like humans do. He uses real examples like the dead salmon study and the stork-baby myth to show how data mining creates fake patterns that look meaningful but are just random noise, warning us that trusting these black-box algorithms without human oversight leads to dangerous mistakes in fields like healthcare and hiring where context matters far more than raw data, reminding us that true intelligence requires embodied experience and logical reasoning that machines simply cannot replicate despite their computational power.
Solid examples; much better than vague advice.
Smith exposes how Watson failed on Jeopardy by matching keywords instead of understanding context like humans do.
The examples of data mining traps feel too convenient and ignore real-world complexity.
The dead salmon study example perfectly shows how data mining creates fake patterns from pure noise without any real scientific meaning.
Gary Smith finally exposes the hype around AI and shows why calling machines intelligent is just a dangerous lie.
Gary Smith convincingly dismantles the hype around AI by showing how systems like Watson and AlphaZero rely on brute-force computation rather than true understanding, using clear examples like the Jeopardy keyword matching failure to prove that mimicking intelligence isn't being intelligent. This perspective helps readers see the danger in trusting algorithms for complex decisions without human oversight, making the distinction between pattern matching and actual wisdom feel urgent and necessary for navigating our tech-filled future.
This book completely changed my perspective on what AI actually can and cannot do.
Gary Smith proves Watson isn't smart, just fast at matching keywords.
The book exposes how data mining creates fake patterns from random noise, showing that algorithms often find coincidental correlations instead of real insights, which makes us trust dangerous conclusions based on meaningless statistics rather than actual logic or common sense that humans possess naturally.
I found the explanation of why AI fails at common sense tasks really helpful.
Smith exposes how AI hype masks simple pattern matching tricks.
Smith brilliantly proves Watson failed Jeopardy by matching keywords instead of truly understanding clues.
Smith argues that AI lacks common sense, using examples like the dead salmon MRI study where algorithms found brain activity in noise. This reveals how data mining creates false patterns without real meaning, making us trust tools that cannot actually understand the world like humans do. The book warns that without human oversight, AI will make dangerous mistakes based on coincidental correlations rather than logic, urging us to demand transparency and theory-driven research instead of blind faith in black-box algorithms that prioritize flashy results over genuine insight.
While the logic about pattern matching is sound, the tone feels overly skeptical and dismissive.