The Atlantic · Ted Chiang ·

A look at AI consciousness debates; LLM conversations are cleverly disguised examples of sentence continuation, but that doesn't deny how impressive LLMs can be

Anthropic is regarded as a giant among AI companies, but perhaps what it really excels in is anthropomorphism.

A look at AI consciousness debates; LLM conversations are cleverly disguised examples of sentence continuation, but that doesn't deny how impressive LLMs can be

Lead Source

How this story grew

Coverage · 0 Discussion · 0
Jun 4Jun 5

More

The Verge: The Verge

Discussion

TechSnif Coverage

Ted Chiang: LLM Chats Are Just Fancy Sentence Completion

Writing in The Atlantic, Ted Chiang argues AI conversations are elaborate sentence continuation — not consciousness.

Ted Chiang has a sharp take on the AI consciousness debate. Writing in The Atlantic, he argues that conversations with large language models are essentially cleverly disguised sentence continuation exercises. Not thinking. Not understanding. Just really, really good pattern completion.

That's not a dismissal. Chiang acknowledges LLMs are genuinely impressive technology. But impressive and conscious are very different things.

His real target? Anthropic. Chiang suggests the AI company — widely considered a heavyweight in the industry — might actually excel most at something unexpected: anthropomorphism. In other words, making machines feel more human than they are.

The piece lands at a moment when AI consciousness debates are intensifying across the industry. As models get more conversational and persuasive, the line between sophisticated autocomplete and genuine understanding keeps getting blurrier — at least in the public imagination.