Both Bell Labs and AT&T had a Buxton Index of a few decades - they both thought about problems on the timescale of decades.
On AT&T’s side, their government-approved monopoly and the mandate that allowed them to maintain it both gave them the Slack - concept to plan on the scale of decades and the incentive to do so.
I’ll make a hand wavy argument for Bell Labs’ Buxton index by using the transistor as an exemplar. It officially took eight years - from 1946 to 1954 to create the silicon transistor. However, work on the transistor began in the 30s before WWII (and was paused during the war) and the first electronic switching centers came on line in 1964. All together, the transistor timescale was maybe 20 years.^1 It seems reasonable to use this 20 year mark as a rough approximation for Bell Labs’ Buxton Index.
It’s hard to imagine the rough equivalence between AT&T’s Buxton Index and Bell Labs’ was a coincidence.
This alignment changed after AT&T’s breakup. Bell Labs’ Buxton index didn’t change, but thanks to competition AT&T no longer had the luxury of a multi-decade Buxton index.
^1:I have very little evidence, but my hunch is that 20 years is roughly the timescale to go from the hunch that a phenomenon can be exploited to exploiting it at scale. Phenomena-based cycles. This timescale is probably compressed for electronics and software because of Moore’s law and development speed (though perhaps not as compressed as many people intuit.) 20 years is shorter than the 50 year Buxton index that EW Dijkstra suggests (is/should be) the norm in Academia. The Strengths of the Academic Enterprise.