AI Philosophy
What Is AI Really? Three Essays That Challenge the Name Itself
The label “artificial intelligence” may be the most successful marketing term in tech history. These three pieces argue the name obscures more than it reveals.
There Is No A.I.
TLDR: Jaron Lanier argues that what we call “AI” is not autonomous intelligence but aggregated human collaboration laundered through statistics. The framing erases the millions of human contributors whose data made the system possible.
Key Insight: Reframe AI as “social collaboration at scale” rather than a thinking machine.
ChatGPT Is a Blurry JPEG of the Web
TLDR: Ted Chiang frames LLMs as lossy compression of the internet. They produce plausible approximations with fabricated details filled in, much like a JPEG introduces artifacts. This is fundamentally different from understanding.
Key Insight: Treat LLM outputs like JPEG artifacts, not primary sources.
AI Is Not Intelligent
TLDR: The foundation of the AI industry rests on a semantic sleight of hand. CEOs claim models “understand” when LLMs perform statistical pattern completion. The gap between the marketing and the mechanism is enormous.
Key Insight: The word “intelligence” in “artificial intelligence” is doing billions of dollars of rhetorical work.
What does this mean for how we think about AI?
These three essays converge on a single point: the language we use to describe AI systems shapes our expectations, our policies, and our willingness to defer to them. Calling pattern completion “intelligence” inflates trust beyond what the technology warrants. Precision in naming is the first step toward precision in use.