{"id":10687,"date":"2026-05-02T13:45:00","date_gmt":"2026-05-02T13:45:00","guid":{"rendered":"https:\/\/musictechohio.online\/site\/old-timey-ai-pre-1930\/"},"modified":"2026-05-02T13:45:00","modified_gmt":"2026-05-02T13:45:00","slug":"old-timey-ai-pre-1930","status":"publish","type":"post","link":"https:\/\/musictechohio.online\/site\/old-timey-ai-pre-1930\/","title":{"rendered":"New AI Trained Only on Pre-1930 Data Speaks Like the Most Old-Timey Guy Imaginable"},"content":{"rendered":"
\n

Tired of your AI chatbot\u2019s constantly-glazing therapy-speak? You could instead try striking up a conversation with \u201cTalkie,\u201d an old-timey AI model which is trained purely on books, newspapers, and other text sources from before the year 1930.<\/p>\n

With its thirteen billion parameters, the researchers behind Talkie say it\u2019s the largest \u201cvintage\u201d model they\u2019re aware of, capable of holding down a conversation as if truly stuck in a past when movies with sound in them were still a novel phenomenon, and when news announcers rattled off the latest signs of tumult in the world in a bouncy Mid-Atlantic accent.<\/p>\n

Intriguingly, Talkie is \u201cbasically\u201d unaware of the fact it\u2019s limited to pre-1930 times, according to David Duvenaud, an associate professor of computer science at the University of Toronto. The AI, he explained in a tweet<\/a>, \u201cdoesn\u2019t have a system prompt and they\u2019re not smart enough yet (as far as we can tell) to introspect well enough to figure out their cut-off date.\u201d<\/p>\n

\n
\n
\n

Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data. Try it below!<\/p>\n

with @AlecRad<\/a> and @status_effects<\/a> \"\ud83e\uddf5\" pic.twitter.com\/kThUESG13e<\/a><\/p>\n

\u2014 David Duvenaud (@DavidDuvenaud) April 27, 2026<\/a><\/p><\/blockquote>\n<\/div>\n<\/figure>\n

Talkie isn\u2019t perfect. The researchers note that it exhibits signs of \u201ctemporal leakage,\u201d in which it produces clearly anachronistic answers, such as knowing that \u201cFranklin D. Roosevelt was president of the United States from 1933 to 1937.\u201d This shows the difficult of\u00a0keeping its data set pure.<\/p>\n

Nonetheless, it raises fascinating questions. What is an LLM\u2019s ability to predict the future? Can the nearly-century-old AI learn a modern programming language? Better yet, can it make scientific discoveries?<\/p>\n

\u201cAs Demis Hassabis has asked,\u201d the researchers wrote in a blog post, referring to the Google DeepMind CEO, \u201ccould a model trained up to 1911 independently discover General Relativity, as Einstein did in 1915?\u201d<\/p>\n

These questions remain hazy, but early tests showed that Talkie was able to create one-line programs, but still has \u201ca long way to go before this capability is notable.\u201d\u00a0<\/p>\n

It\u2019s also not quite a soothsayer: tests performed by the researchers showed that the\u00a0AI found historical events summarized in the New York Time<\/em>\u2018s \u201cOn This Day\u201d section to be more \u201csurprising\u201d after the knowledge cutoff, especially in the 1950s and 1960s.<\/p>\n

That doesn\u2019t mean its extrapolations aren\u2019t amusing. In one user\u2019s testing<\/a>, Talkie predicted that another World War would break out in 1936, and that \u201cflying machines\u201d would be in everyday use for transport. Though it also strangely predicts that by 1999, \u201cthe sun will have ceased to shine,\u201d perhaps reflecting contemporaneous anxieties over the dawn of a new millennium.<\/p>\n

In another test by the researchers, Talkie called talking pictures \u2014 the slang for them being its namesake \u2014 \u201coverrated.\u201d<\/p>\n

\u201cThey will never replace silent films, but may supplement them, and possibly in the future they may be shown in the same theater at the same time,\u201d it confidently predicted in its long-winded style. \u201cAt present, however, they are interesting chiefly as a novelty.\u201d<\/p>\n

More on AI:<\/strong> Police Are Using AI Camera Networks to Stalk Women<\/em><\/a><\/p>\n

The post New AI Trained Only on Pre-1930 Data Speaks Like the Most Old-Timey Guy Imaginable<\/a> appeared first on Futurism<\/a>.<\/p>\n<\/div>\n

<\/div>","protected":false},"excerpt":{"rendered":"

Tired of your AI chatbot\u2019s constantly-glazing therapy-speak? You could instead try striking up a conversation with \u201cTalkie,\u201d an old-timey AI model which is trained purely on books, newspapers, and other…<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[177],"tags":[],"class_list":["post-10687","post","type-post","status-publish","format-standard","hentry","category-artificial-intelligence"],"_links":{"self":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/posts\/10687","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/comments?post=10687"}],"version-history":[{"count":0,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/posts\/10687\/revisions"}],"wp:attachment":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/media?parent=10687"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/categories?post=10687"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/tags?post=10687"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}