Comment Bell vs. Academic R&D (Score 4, Interesting) 41
One thing Bell Labs did do correctly was by being part of a company, they had the needs and desires of a market and consumers to act as a pull, a "Why" so to speak. Even with some of the more speculative research they did, there was always an eye towards how would this get deployed to a market as a product and thus become a technology available to the broader world. Academia does not have this; most academia lives in a bubble. WHen you read academic papers, they all follow the same outline: 1) there's a problem in the world, 2) a new technology is hypothesized to solve it, 3) here is an experiment and data to show it cna solve it, and 4) that problem might be solved by this new technology.
While generally most papers I read have mid- to good- experimental design and data and decent conclusions, it's the first part about a problem that is almost entirely made up from nothing. In my field at least, when I read an academic paper that is supposed to relate to my field, the problem they are identifying isn't even a problem at all; it has no relevance to the field or what people want. As such the technology is funded, experiments are done, a paper is published, and it dies on the vine because it turns out no one wants it because it was never a problem in the first place.
It's noticeable that when it comes to silicon and chip-based technologies, academia does very little; the silicon industry basically took it all over through the formation of SRC. SRC unfortunately just got gutted by the Trump administration, but for a long time SRC and the industry ran all of the new technology development; there wasn't anything like chip architectures or new interconnects or anything coming out of academia at all; it was all industry led and academia only did niche stuff like graphene and organics and the like.
Industry at least provides a pull for a new technology academia has to push it and it rarely results in anything useful. It's different for different fields, but in my field at least I haven't found a single university program that's even remotely relevant; the Bell Labs model is sorely missed.