Claim · 2019-09-10 · publication · Stated in source · Supported

Marcus and Davis's excerpt of September 2019 states that GPT-2 can generate strings of words that sound humanlike without real coherence, framing its text generation as statistics of which words occur together rather than understanding.

One recorded claim. The claim panel gives its qualification, the source passages with their place in each source, and how it was checked.

What it connects

If Computers Are So Smart, How Come They Can't Read? → articulates → Language models lack agency or understanding

Where it appears

Arguments about catastrophic risk from advanced AI →

Claim and evidence

If Computers Are So Smart, How Come They Can't Read?articulates ↓Language models lack agency or understanding

Machine-checked, not reviewed by an editorStated in sourceSupported

Supporting evidence: primary · 1 source

Date: 2019-09-10 · publication Records dated 2019 →
Role: Earliest located
Register (kind of text): Nonfiction

Scope and limitsThe authors' own argument about GPT-2, adapted from their book; it is not a test of every language model. Earliest located among the texts the searches found, after a cross-family check located it; it is not evidence of coinage.

The evidence 01

primary sourceSupports the claim

If Computers Are So Smart, How Come They Can't Read?

Wired (wired.com) · 2019-09-10

Cite the original source

Cite the original document, not Selekoa. Check these details against it.

Source excerpt

“This means you can generate strings of words that sound humanlike, but there’s no real coherence there.”

Location in source Discussion of GPT-2, paragraph before a generated continuation

How it was read Retrieved directly from the publisher

Machine-checked · 2026-10-31

Source details and check notes

Authors: Gary Marcus; Ernest Davis
Version: The excerpt as served by Wired, dated 10 September 2019 and marked as adapted from Rebooting AI (Pantheon, 2019). Its wording can change; earlier wording is not authenticated. Retrieved at admission on 2026-10-31 (SHA-256 de6db7699906d03aa1f8ff8ce589bd6dd9a4f79cb2454a80a74c18f8e21d2977).
Canonical address: Open the canonical source ↗
Accessed: 2026-10-31
Language: English (en)
Attribution: Gary Marcus, Ernest Davis, “If Computers Are So Smart, How Come They Can't Read?”, 2019-09-10.
Reuse: No open redistribution licence asserted. Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution.
Check notes: Added at the admission of the risk-debate pack (research/risk-debate). The quotation (17 words) matched a fresh retrieval of 2026-10-31 (SHA-256 de6db7699906d03aa1f8ff8ce589bd6dd9a4f79cb2454a80a74c18f8e21d2977). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms. The cross-family check of 2026-10-31 proposed the record and read this passage. No human editorial review.

Status and meaning

Editorial status. Not reviewed by an editor. An exploratory record that no editor has accepted. This status does not say whether a person has read any linked passage. Each evidence record above shows how it was checked.

Relation. A passage of the document meets the proposition’s criterion. The record documents what the text states, in its register (fiction and satire are labelled); it is not evidence that the proposition is true and implies no influence on or from another text.

Date. Date of the Wired excerpt; the book appeared in 2019.

Record ID rd-r14-marcus-davis-2019-computers-cant-read