New AI Trained Only on Pre-1930 Data Speaks Like the Most Old-Timey Guy Imaginable
TL;DR
We do declare that this talking machine is rather of its time. The post New AI Trained Only on Pre-1930 Data Speaks Like the Most Old-Timey Guy Imaginable appeared first on Futurism. We do declare that this talking machine is rather of its time. The post New AI Trained Only on Pre-1930 Data Speaks Like the Most Old-Timey Guy Imaginable appeared first on Futurism.
Nauti's Take
The experiment is genuinely valuable for understanding bias in LLMs – it shows just how heavily training data shapes the output. The catch: the same dynamic runs in reverse, and modern production models are nowhere near as easy to de-bias.
Nauti's view: a perfect teaching tool for devs and AI literacy – not proof that bias problems are solvable in real-world models.
Sources