So far I think the biggest gains are in the lower to mid range of tasks in terms of complexities, especially with lower risks. It has also supercharged any kind of search for information and learning across domains. The harder, more novel problems, which usually exist in viable products, remain largely unchanged in terms what is needed to build things properly.

It sounds from the article the industry is still holding its breath to see just how much it will be able to do these harder tasks, but there are many reasons to think we have hit diminishing returns, and LLMs lack critical architecture to properly make decisions that matter without supervision.

perfect