Ask HN: Will Large Language Models be like another industrial revolution
1 year ago I thought GPT style models were cool and useful, but nothing too revolutionary. I had tried copilot but that too felt like not an extreme advantage for typed languages over something like IntelliJ.
However, I have been trying GPT4 and seeing the rate of progress, and it all seems too much progression too fast. GPT4 is crazy good on all kinds of things (e.g. I asked it to make changes to python C internals to support some new language features and it came up with sensible changes). Unlike GPT3.5, I have so far not seen GPT4 just make up code functions. It hallucinates but the frequency seems much lesser. And with the pace that OpenAI is developing new things, it seems like it is even hard to predict what these models will be doing in 2 years.
Do you think we are witnessing the start of something with similar impact as the industrial revolution, but happening even faster and will dramatically change human civilisation?
However, I have been trying GPT4 and seeing the rate of progress, and it all seems too much progression too fast. GPT4 is crazy good on all kinds of things (e.g. I asked it to make changes to python C internals to support some new language features and it came up with sensible changes). Unlike GPT3.5, I have so far not seen GPT4 just make up code functions. It hallucinates but the frequency seems much lesser. And with the pace that OpenAI is developing new things, it seems like it is even hard to predict what these models will be doing in 2 years.
Do you think we are witnessing the start of something with similar impact as the industrial revolution, but happening even faster and will dramatically change human civilisation?