The last few years, cycling and triathlon have been experimenting with upto 120g carbs intake per hour. Last year, Cameron Wurf ate 200g carbs per hour when he broke the world record for fastest bike split ever in a triathlon (which was broken again a few months later).
a 2025 look at elite triathletes fueling at https://www.triathlete.com/nutrition/race-fueling/ironman-wo... shows that norwegian athletes are ingesting higher amounts of carbs (~180g/hr bike, ~120g/hr run - 2 males, ~150g/hr both run & bike - 1 female) especially for the bike portion.
This might seem weird coming from a car manufacturer but Skoda is a big sponsor of cycling races, most notably of the Tour de France and other ASO races. And as explained in the footer, they started out with building bicycles in the 19th century.
Your analogy to computer architectures doesn't make sense, unless comparing GPT-like LLMs to different LLM architectures like Mamba or RWKV. It indeed wouldn't make sense to not teach about Mamba or RWKV in an introductory AI or LLM course.
AI is much broader than LLMs alone. Computer vision, RL, classical ML, recommender systems, speech recognition, ... are still part of AI, just not very visible to the average consumer.
It's not an actual file but as a variable in a js file. The last link in the blog post does link to a commit with a file that contains the instructions for Claude, lines 129-737.
reply