Vintage chatbot lives in the past like an elderly relative
Talkie is a vintage language model trained exclusively on English-language texts published before 1931, making its knowledge cutoff the end of 1930. Its creators aim to use it to study AI behavior and test long-term forecasting and scientific discovery capabilities. While it can solve simple programming tasks, its capabilities are limited compared to modern models.
- ▪Talkie is a 13-billion-parameter language model trained solely on texts published before 1931.
- ▪The model's training data includes books, newspapers, journals, patents, and case law from before 1931, chosen because works from that period are in the U.S. public domain.
- ▪Talkie's creators hope it can help evaluate AI's ability to predict the future and replicate historical scientific discoveries using period-limited knowledge.
- ▪The model generated only basic, one-line solutions to Python programming problems, indicating significant limitations.
- ▪Researchers suggest vintage models like Talkie could advance understanding of AI by simulating historical knowledge constraints.
Opening excerpt (first ~120 words) tap to expand
(function() { let windowUrl = window.location.href; windowUrl = windowUrl.substring(windowUrl.indexOf('?') + 1); let messageElement = document.querySelector('.shareableMessage'); if (windowUrl && windowUrl.includes('code') && windowUrl.includes('expires')) { messageElement.style.display = 'block'; } })(); AI + ML Vintage chatbot lives in the past like an elderly relative Talkie's training data stops at the end of 1930, and its creators hope it'll help us better understand how AI thinks Brandon Vigliarolo Brandon Vigliarolo Published tue 28 Apr 2026 // 18:51 UTC If you're tired of interacting with a bot that spews Nazi propaganda or refers to itself as MechaHitler, you could sign off of Elon Musk's xAI.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at The Register.
Discussion
0 commentsMore from The Register
