LLMS ARE MAKING ME DUMBER

Summary from the TLDR newsletter

Large language models can help people output substantially more work, but this usually comes at the cost of learning or improvement. People are incentivized to maximize immediate output, so the sacrifice seems fair. Before the rise of generative AI, learning was a prerequisite for output, but people can now use technology to build things without deeply understanding the implementation. Developers should decide what skills they are and are not okay with offloading and what they want to do for themselves.