The future isn't in writing new AI code, it's in refining the old code and benefitting from the vast computation powers of the future.

We've already seen the advent of chatgpt, Claude, Gemini, Llama, Perplexity and many more. <br><br>Unlike the previous year when GPT-4 was released and everyone's mind was blown, the current year is not bringing new ideas and is instead focused on improving what's been done before and augmenting it to work better with more powerful hardware. <br><br>The pattern in new releases is, take an old model and improve it with better hardware. <br><br>This is not only benefitting OpenAI but also other companies like Google and Anthropic who are getting exponentially better by exploiting the same idea.