← Back to archive

A year of LLM milestones, trends, and surprises

6 January 2025 · 34 views · 0 likes · 0 reposts · 173 blog views

AI automation neuroscience product research

What a year for Large Language Models. Here’s a quick rundown of the biggest milestones, trends, and surprises:

🔹 Breaking Barriers: The GPT-4 benchmark shattered by 70+ models, many outperforming it in key areas. 🔹 LLMs on Laptops: Some GPT-4-level models now run on personal devices. 🔹 Price Wars: Costs for LLM usage dropped dramatically, making advanced AI more accessible. 🔹 Multimodal Magic: Models now handle text, images, audio, and video seamlessly. 🔹 Voice & Vision: Chat with AI like it's sci-fi. Live video and realistic voice interactions became a reality. 🔹 Commoditized Tools: Building apps with prompts is now a common feature in many LLM platforms. 🔹 Environmental Impact: Mixed news - training is more efficient, but datacenter expansion raises concerns. 🔹 Synthetic Data Revolution: AI training on synthetic data proved shockingly effective.

Despite leaps forward, challenges remain: "Agents" haven't materialized as expected, usability for non-experts is still tricky, and the field desperately needs better public understanding and critical analysis.

Originally posted on X

PS: I'm pretty active on 𝕏 if you'd like to follow along

Related posts

Archive

2026

2025

2024

2018