3 Comments
User's avatar
Moritz Wallawitsch's avatar

The conclusion here seems wrong:

> They found that ~70% copy-pasted the first output without editing it at all. .... This means that ChatGPT can immediately substitute for some workers. It’s a story of replacement rather than augmentation.

1. If they only accepted 70% and not 100% isn't that an argument for augmentation not automation?

2. Who wrote the prompt? We always need someone to write the prompt as these models are no generally intelligent agents.

I wrote a more nuanced take on this topic here: https://scalingknowledge.substack.com/p/why-job-displacement-predictions

Expand full comment
Kenn So's avatar

Love a different take on it. For the 1st one, the 'right answer's is it's a blend of both. In my job, these tools augmented me. I get more stuff done. But there's a flip side to that. My cousin runs a small marketing agency and was thinking of hiring another person to help service clients. After learning how to use these tools, she opted not to. Her current team is augmented , replacing the need for a new employee. It's not a direct replacement aka firing, but I know of 2nd hand anecdotes of firing because they're made redundant because of AI. Stories of both augmentation and replacement will play out across society, but on the balance I still lean replacement of existing jobs. But there will be net new jobs that we don't know of yet.

On your 2nd point, I would have agreed a few months ago. But seeing how fast agentic AI is developing, a single human can ask AI to self-prompt. A human will still have to prompt, but there's net less total need for humans to prompt.

Expand full comment
Anonymous's avatar

Cool stuff, mate

Expand full comment