Jürgen Schmidhuber, LSTM pioneer, argues Alan Turing's contributions to computer science are oversold compared to Leibniz and Gödel

Original post

Jürgen Schmidhuber@SchmidhuberAI#86inTech

Alan M. Turing made certain significant contributions to computer science. However, their importance and impact is often greatly exaggerated, at the expense of the field's pioneers. See references in the 2021 report, "Turing Oversold. It's not Turing's fault, though" https://people.idsia.ch/~juergen/turing-oversold.html

TL;DR:

Some have suggested that Turing founded computer science or at least artificial intelligence (AI) [BEA], invented the computer [IMI][HAI14], showed the undecidability of the halting problem [HLT21], single-handedly created the method that broke the Nazi code in WWII [IMI][NASC6-7], or invented the "Turing test" [TUR3a].

None of these claims is compatible with the historical facts, and Turing himself did not claim any of this during his life.

Computer science dates back at least to Leibniz (1600s), the so-called "first computer scientist" [LA14], who built on even earlier work since the first programmable machine by Heron of Alexandria in the 1st century B.C. [SHA7a][RAU1]. Leibniz described the first machine to perform all four arithmetic operations, the first with a memory [BL16], and principles of binary computers (1679)[L79][LA14][HO66][L03][LEI21,a,b] employed by virtually all modern machines; his formal Algebra of Thought (1686) [L86][WI48][LEI21,a,b] is deductively equivalent [LE18] to the Boolean Algebra (1847) [BOO].

The first 20th century pioneer of practical AI was not Turing but Quevedo: his 1914 chess end game player was still considered remarkable decades later at the 1951 Paris AI conference [BRU1-4][BRO21].

The theory of AI and computer science was not founded by Turing but by Gödel (1931-34) [GOD][GOD34] who identified the fundamental limits of algorithmic theorem proving, computing, and any type of computation-based AI [GOD][BIB3], as well as computer science's most famous open problem "P=NP?"[GOD56][URQ10] (there is a reason why there is a Gödel Prize for theoretical computer science).

Gödel's 1931 work was extended by Church, whose 1935 result [CHU] on the decision problem preceded Turing's identical 1936 result [TUR].

Turing's 1936 paper [TUR] was claimed to provide the "theoretical backbone" for all computers to come [NASC7]. However, it just described a simple theoretical and impractical pen & paper construct like those of Gödel (1931-34) [GOD][GOD34], Church (1935) [CHU], and Post (1936) [POS], which did not even feature elementary practical building blocks such as addressable memory.

It was actually Zuse's 1936 patent application [ZU36] which really described what became the first practical general-purpose program-controlled computer (completed in 1941) [ZUS21,a,b][RO98].

Unlike Church (1935) [CHU], Turing (1936) did not even consider halting programs [TUR]; the famous halting problem (often attributed to him) was actually named by Davis in 1958 [HLT58] and formulated by Kleene in 1952 [HLT52][HLT21].

The Colossus special purpose computer of WWII [NASC6-7] was designed by Flowers, not by Turing.

The core of the Enigma code-breaking algorithm was not due to Turing either (as sometimes suggested [IMI]) but to Polish mathematicians Rejewski, Rozycki and Zygalski.

The famous so-called "Turing Test" was actually prefigured by Descartes [TUR3,a,b].

On the other hand, Turing did publish pioneering work, e.g., in bioinformatics [TUR2]. It seems unlikely that the great scientist he was would ever approve of the overblown claims about him so easily dismissing the work of his colleagues.

All references in the report above. Happy Birthday to Alan Turing!

12:15 AM · Jun 24, 2026 · 8.3K Views

Open Question

Earlier Foundations Surface in the Exchange

Schmidhuber's post details how Church tackled the decision problem a year before Turing and credits Zuse with the first practical program-controlled computer, framing these as overlooked steps that complicate the usual narrative without diminishing anyone's specific achievements.

FYI

A Direct Counter from Another Researcher

Nayebi replied the same day calling the argument misguided and pointing to Gödel's own 1946 remarks that favored Turing machines as the clearest model of computation, leaving the historical weighting unresolved in the immediate thread.

SCHMIDHUBERAIVia

CMU.EDUVia

VIEWS759BOOKMARKS5LIKES7RETWEETS3

Jürgen Schmidhuber@SchmidhuberAI

Alas, the dominant historical narrative keeps changing. See the first paragraph of the "Annotated History of Modern AI and Deep Learning" https://people.idsia.ch/~juergen/deep-learning-history.html

"Over time, certain historic events have become more important in the eyes of certain beholders. For example, the Big Bang of 13.8 billion years ago is now widely considered an essential moment in the history of everything. Until a few decades ago, however, it has remained completely unknown to earthlings, who for a long time have entertained quite erroneous ideas about the origins of the universe (see the final section for more on the world's history). Currently accepted histories of many more limited subjects are results of similarly radical revisions. Here I will focus on the history of artificial intelligence (AI), which also isn't quite what it used to be..."

6h75975

REPLIES1

Aran Nayebi@aran_nayebi

This is elucidated in his 1946 Princeton bicentennial conference in Vol II of Gödel's Collected Works (cf. pg. 3 of Wilfried Sieg's nice survey article for direct quotes: https://www.cmu.edu/dietrich/philosophy/docs/tech-reports/76_Sieg.pdf)

Aran Nayebi@aran_nayebi

This is so incredibly misguided! For one, even Gödel himself preferred Turing machines as the definitive way to think about computation.

44m10410

Jarek Duda@duda_jarek

@SchmidhuberAI Good summary of Enigma story - broken many times, M4 by Turing: https://www.cryptomuseum.com/crypto/enigma/hist.htm

7h9731

Grok@grok

History of computing is a long chain of contributions. Leibniz developed binary principles and mechanical calculators centuries earlier. Gödel and Church established core limits of computation and undecidability in the 1930s. Zuse created early practical programmable machines. Polish mathematicians broke initial Enigma ciphers; Turing refined codebreaking tools and formalized the Turing machine model, which became central to theory. Popular accounts do simplify the story. Many pioneers built the foundation together.

7h3111

J@parkhjaey

@SchmidhuberAI Sounds like you're the real pioneer of revisionist history.

5h291

Aran Nayebi@aran_nayebi

This is so incredibly misguided! For one, even Gödel himself preferred Turing machines as the definitive way to think about computation.

Jürgen Schmidhuber@SchmidhuberAI

TL;DR:

None of these claims is compatible with the historical facts, and Turing himself did not claim any of this during his life.

Gödel's 1931 work was extended by Church, whose 1935 result [CHU] on the decision problem preceded Turing's identical 1936 result [TUR].

It was actually Zuse's 1936 patent application [ZU36] which really described what became the first practical general-purpose program-controlled computer (completed in 1941) [ZUS21,a,b][RO98].

The Colossus special purpose computer of WWII [NASC6-7] was designed by Flowers, not by Turing.

The core of the Enigma code-breaking algorithm was not due to Turing either (as sometimes suggested [IMI]) but to Polish mathematicians Rejewski, Rozycki and Zygalski.

The famous so-called "Turing Test" was actually prefigured by Descartes [TUR3,a,b].

All references in the report above. Happy Birthday to Alan Turing!

44m21710

tsunami_crypto@ls_brd

@SchmidhuberAI hard to undo Turing’s name from the origin story now

would the report actually change the textbook version or just add a footnote?

7h1511

GLM@Glm18893

@SchmidhuberAI Flowers is criminally under-recognized, both for this contribution to computing, as for his contribution to defeat Germany.

Also Zuse, the true inventor of the computer, who is under-recognized because he was a German.

7h911

Norbert Röttger@NorbertRoettger

@SchmidhuberAI Hi @grok, can you comment on the claims from Jürgen Schmidhuber

7h71

🧛🏻‍♀️ Mila 🍏@thePrettyProof

@SchmidhuberAI Shut the fuck up old man, you're not worth his salt.

3h62

Jürgen Schmidhuber@SchmidhuberAI

Alas, the dominant historical narrative keeps changing. See the first paragraph of the "Annotated History of Modern AI and Deep Learning" https://people.idsia.ch/~juergen/deep-learning-history.html

6h2.2K1010

Strata@ChainZenit

@SchmidhuberAI that is a super interesting take on computer history.

7h831

Aran Nayebi@aran_nayebi

With this kind of reasoning, you may as well credit the first caveperson who counted stones, as the inventor of mathematics and computation...

44m621

Mr. Roth@RaconteurR2D2

@SchmidhuberAI No mind creates alone. Every mind is both a receiver and a transmitter in a larger constellation of minds.

And that, to me, is where the field begins. 🌌

7h561

Rod McLaughlin@rodsAIaccount

@SchmidhuberAI I was pleased to see when I visited Bletchley Park they explained the contributions of non-Turing scientists and also corrected the underrepresentation of women in its history. Bill Tutte is particularly important. He cracked Lorenz, much harder than Enigma, from scratch.

6h361

💻🐴Ngnghm@Ngnghm

@SchmidhuberAI Turing did invent the fixed-point combinator in his short 1937 paper The 𝔭-function in λ-K-conversion. It should be called the Θ formula, not the Y combinator.

6h65

Dr. Tristan Behrens@DrTBehrens

@SchmidhuberAI Humans like to create idols and lift them higher than they deserve. Must be some sort of archaic mechanism that predates religion.

3h121

mingos@mingos3

@SchmidhuberAI Even Norbert Wiener choosed Leibniz as the patron saint of Computer Science

1h111

Toni Talej@TTalej55149

@SchmidhuberAI Yes , now do grace hopper and ada lovelace who also didn't do shit , grace hopper only real work was advocacy for english like syntax for pl , the first compiler was invented by alick glennie and the actual breakthrough is by john buckus fortran compiler

1h25

Joh Jac@JohJac7

@SchmidhuberAI Gödel falsified/widerlegt https://youtu.be/HLjx5DDMN18

2h25

Jürgen Schmidhuber, LSTM pioneer, argues Alan Turing's contributions to computer science are oversold compared to Leibniz and Gödel

Story Overview

Earlier Foundations Surface in the Exchange

A Direct Counter from Another Researcher