March2023-GPT4 notes

GPT-4 is capable of doing amazing things (confirmed):

• Rough sketch to website • Build games without writing a single line of code • Write songs and compose music • Expose Blockchain vulnerabilities • Generate video scripts • Drug discovery • Generate entire brand design

IMO GPT-4 is a bigger leap than GPT-3 was.

• GPT-3 advanced AI from task-specific models to a single prompted model that is task-general
• GPT-4 is human-level on many hard tasks, and will signal a societal revolution where AI reaches every industry, starting with technology

Before GPT-3, we fine-tuned a task-specific neural network (e.g., BERT) for each task that we wanted to solve. With GPT-3, there was a paradigm shift to using a single large language model for any task via few-shot prompting.

Before GPT-4, AI was OK at surface-level statistical learning, but not usable for many tasks (e.g., reasoning). With GPT-4, AI reaches human-level on many intellectual tasks—a qualitative shift in the “feel” of using the technology that will unlock a new world of use cases.

Since GPT 3.5 was delivered as an intermediate step, it may seem like GPT-3 to 4 was not as big, but IMO if you consider the trajectory of developments from 2020 until now, GPT-4 is quite a lot different. Super excited for the future of AI!

GPT-4 is way better at programming than given credit for.HumanEval is a benchmark of python programming problems. With some prompt engineering, GPT-4 scores ~85%, destroying Codex’s 29% from just 2 years ago and performing much better than OpenAI’s publicized accuracy - https://twitter.com/amanrsanger/status/1635751764577361921

So what is next

• Get early access to GPT-4 https://github.com/openai/evals/blob/main/README.md#Setup
• Do you realize that GPT-5 will come before the end of the year 2023? People won’t even have time to properly integrate GPT-4 into their work or lives!