[avanai]

The most impressive thing about this story is that they figured out the answer. They did the research, and nailed down that it was Nyquist who was was the productivity booster. It’s the exact opposite of the OP’s story, where management tried to fire the Nyquist-equivalent.

[Tarq0n]

They found an answer that felt right to them. The reseachers weren't blinded to the context they were working in, and their hypothesis is essentially unfalsifiable so I would take it with a grain of salt.

[liquidpele]

All they figured out was that the smart people hung out together.

[sound1]

IMHO this is the right answer :)

[Strilanc]

Honestly, I'm kind of skeptical of the answer. I'm not saying that talking with Nyquist wouldn't be useful, probably it was, but what's stopping a dozen other things at least that useful from being part of the answer?

[jacquesm]

> I'm not saying that talking with Nyquist wouldn't be useful

No, you probably shouldn't be saying that:

https://en.wikipedia.org/wiki/Harry_Nyquist

https://en.wikipedia.org/wiki/Nyquist_frequency

> but what's stopping a dozen other things at least that useful from being part of the answer?

Because someone needs to act, and that's exactly what Nyquist did, in a very unobtrusive and non-confrontational manner.

[edgyquant]

Posting his achievements does nothing to prove speaking with him during lunch was useful in this context

> Because someone needs to act, and that's exactly what Nyquist did, in a very unobtrusive and non-confrontational manner

Seems like your headcanon. It reality they ate lunch together and passed ideas around.

[Strilanc]

For the record, I did recognize the name. That's why I believed talking to them was useful.

[OJFord]

Concretely I'd suggest that Nyquist was probably most interested in lunching with other smart people who had interesting things to talk about.

I.e. there's no check or control on their output without lunch or breakfast with him, maybe it'd have been little different.

[solumunus]

It’s a lovely story but a prime candidate for “correlation !== causation”.