In 2024, significant advancements in Large Language Models (LLMs) were made. The GPT-4 barrier was broken by 18 organizations, including Google, creating even better models. Surprisingly, some of these GPT-4 models can now run on a personal laptop, showcasing improved efficiency. LLM prices drastically dropped due to competition, making them more accessible. Multimodal vision models, including audio and video input, emerged, expanding LLM capabilities. The development of prompt-driven app generation became a commodity, allowing the creation of full applications using LLMs. Despite these advancements, universal access to the best models lasted only briefly, with subscriptions now required for top-tier models.
https://simonwillison.net/2024/Dec/31/llms-in-2024/