Predictions for AI in 2024

In 2024, every application you use will have AI. If it doesn’t, you will probably use an alternative in sometime.

Here are my predictions for AI in 2024 -

1. OpenAI is still at large but LLMs by other providers will catch up

Most LLM models by other providers like Google, Anthropic and others are catching up to OpenAI. GPT-4 is still the best for most generic applications like chatbots but others are not far behind. They will eventually catch-up and probably be great at specific applications.

2. Open Source Models will catch up with commercial ones

Some Open & “Smaller” Language Models are faster, cheaper and safer to run on-prem. And sustainable. They will be good for variety of use cases and shorter generations that will improve over time.

3. RAG will rule

RAG is a term for using your own information to generate results otherwise unknown to LLMs. In the first phase of AI adoption, RAG will become an invisible layer powering most application - new and old. It remains one of the best ways to adopt AI in enterprises.

4. AI Talent Shortage will continue

AI is a generational shift in technology and it will take a while to train and re-train talent. As AI adoption becomes mission critical for companies, talent shortage will become acute.

5. AI use cases will explode

We are beginning to scratch the surface on how AI can be used. There will be new applications powered by AI challenging the incumbents and novel applications breaking boundaries. It will be an exciting year for AI.