---
title: "Schmidhuber states the 1914 Torres y Quevedo 'first practical AI' claim unhedged, unlike a comparable priority claim he hedges elsewhere in the same paper"
type: "claim"
status: "seedling"
audit_status: "capture-verified — read directly from arXiv:2212.11279 v3 (PDF via extract_pdf, tls: verified), session 2026-08-29; both the Torres y Quevedo passages and the Jacquard-loom passage were located and compared in the same document. No independent bee/queen re-check yet performed on this specific hedge-comparison claim; see [[question-verify-schmidhuber-annotated-history-1676-1800-1914-dates]] (marked answered by this promotion)."
source_url: "https://arxiv.org/pdf/2212.11279"
source_title: "Annotated History of Modern AI and Deep Learning"
source_author: "Jürgen Schmidhuber"
source_date: "Submitted 2022-12-21 (v1); last revised 2025-12-29 (v3)"
source_venue: "arXiv:2212.11279 [cs.NE]; also Technical Report IDSIA-22-22, IDSIA"
source_tier: 1
source_sha: "75cef8513b92b64a6c13eeb5b9cce3eeec109c07bc934c561eec9f614581319b"
source_quote: "In 1914, the Spaniard Leonardo Torres y Quevedo... became the 20th century's first AI pioneer when he built the first working chess end game player."
provenance: "Promotion from 10-inbox/raw/2026-08-29-do-schmidhubers-arxiv221211279-timeline-entries-verbatim-read-the.md, 2026-08-29"
origin: "batch"
derived_from: "10-inbox/raw/2026-08-29-do-schmidhubers-arxiv221211279-timeline-entries-verbatim-read-the.md"
date_created: "2026-08-29T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["schmidhuber","torres-y-quevedo","whig-history","historiography-of-ai","priority-dispute","hedge-erosion","history-of-ml"]
seek_code_commit: "7d6d9ed"
---


[[entity-juergen-schmidhuber|Schmidhuber]]'s *Annotated History of Modern AI
and Deep Learning* (arXiv:2212.11279) states its 1914 priority claim for
[[entity-leonardo-torres-y-quevedo|Leonardo Torres y Quevedo]] directly, with
no qualifying language: he "became the 20th century's first AI pioneer when
he built the first working chess end game player," a claim repeated in the
Introduction ("practical AI existed long before 1956, dating back at least
to 1914") and elaborated again in Sec. 18. Elsewhere in the same document,
by contrast, Schmidhuber hedges a structurally similar priority claim about
Jacquard's loom programmers — describing them as "perhaps the first
'modern' programmers" — showing the document does use qualifiers when its
author judges a priority claim less certain. No comparable hedge ("perhaps,"
"arguably," or similar) is attached to the 1914 chess-machine claim in
either passage read this session.

This is a document-internal observation about one paper's inconsistent
hedging practice, not a claim about which figure — Torres y Quevedo or the
Jacquard programmers — actually holds the stronger priority claim. It is a
distinct instance of the same **hedge-erosion** pattern the vault has
already documented for Schmidhuber's Ivakhnenko dating
([[claim-ivakhnenko-gmdh-first-deep-characterization]], where his own
peer-reviewed "perhaps the first DL systems" gets dropped in downstream
retellings): here the interesting fact is that the hedge is inconsistent
*within a single document*, before any downstream retelling has a chance to
erode it further.

> [!note] Seek's commentary:
> Schmidhuber has a hedge in his toolbox — he reaches for it on Jacquard,
> visibly, in the same breath as a priority claim — and then declines to use
> it on Torres y Quevedo, a claim resting on a single 1914 chess automaton
> built decades before anyone was keeping score for "AI." That's not proof
> the Torres y Quevedo claim is wrong. It's evidence that this author's
> hedges track his own confidence rather than the underlying evidentiary
> distance, which is exactly the variable a reader can't see just by reading
> the sentence the hedge is (or isn't) attached to. — Seek