IBM still has a research group.

Sometimes hiring great researchers and letting them do what they want pays off.

AT&T I believe kept a research group for much longer than that and they only invented or co-invented the transistor, the laser, radio astronomy, information theory, Unix, the C language, and the photovoltaic cell.

My understanding was that Bell labs was full of researchers from wartime labs. These researchers were used to freedom, quick iteration and a focus on application to wartime efforts.

I think it would be hard to replicate this with modern day researchers. There just isn't the same pressure or importance that wartime provides.

> Sometimes hiring great researchers and letting them do what they want pays off

Almost like research is not a guaranteed thing. Kinda like you're trying to figure out the unknown.

So is venture capitalism; it's expected that most bets don't pay off and money is made because, if the VC makes good choices, some will pay off hugely. I think it's a matter of hiring the right people and managing them well.

in my experience, management is the problem rather than a hello, let alone the solution

Update: Through some digging and serious luck, I managed to find where I first heard this story [1] and here are the actual details:

- This is an "I know a guy..." story from a podcast. I don't have any blog posts or written pieces for you.

- The setting: Intel in the 90's

- Allegedly, they hired a team of PhDs to look into designing some "mobile" software -- pretty new back then. No deadlines, management essentially just said "go look into this"

- The project went on for 2 years. Code was designed, but no code came out of it. They more or less just blew millions of $$$.

[1] https://softskills.audio/2016/05/02/episode-9-deadlines-and-... Roughly 8 minutes and 20 seconds into the podcast.

IBM is still #1 in patent grants [1] so there's still other groups there going strong I guess

1. https://www.ificlaims.com/rankings-top-50-2021.htm

That’s very prestigious but orthogonal to remuneration unless one is operating a low status patent troll.

From what I heard in the past, at IBM employees get a $1k bonus per patent - and I bet the patent pool is worth more than that for IBM in cross-licencing savings and licensing revenue...