Lots of middlebrow dismissal in the comments on this article, but none that actually engage with its content, at least not at the time I’m writing. Some of the dismissals seem like they only read the headline, not the article. (Shocking, I know.)

Personally, I found the article well written, and describing a very thorough set of research. I don’t doubt that there are leaders right now ceding their decision-making to LLMs, and they would do well to read this article.

I'm missing some data to evaluate this article. If you engage a set of humans on these seven tensions, what range of bias do you get?

What else would they expect? Next token prediction virtual requires this outcome.

>An LLM is not the colleague who critically evaluates current ideas, looks into the contextual specifics, stress-tests assumptions, and pushes back when everyone gets comfortable

Well, not if you don't ask it to, it won't.

So ask it to.

(The article, which is better than it sounds like it's going to be at first, actually does advocate that towards the end. It's worth a read.)

Well of course they did! "AI" can't possibly replace Accenture or McKinsey, or lawyers!

"Harvard Just Caught AI Lying to Every Executive in America" - https://youtu.be/pd1Km6bT104

[dead]