If you use LLM, I recommend looking into DeepSeek. A completely free alternative to ChatGPT at the GPT-4o and GPT-o1 level. Moreover, they have a VERY cheaper API compared to all other existing models, and no rate limit.

Imagine developers rushing to use this.

But the coolest thing is that their new model is completely open source. Large models will require a lot of power to run locally, but mini models can be easily run locally even on a personal MacBook.

Try DeepThink mode. It's fun to read the "thoughts" of the model. It feels like AGI has already been created. Or maybe it is...