> People like to say, "But those people don't even create anything, they're just taking credit for other people's work!" But just letting smart people gather and work on pet projects doesn't seem to work particularly well either. Maybe having a headstrong product oriented person upstairs is useful/necessary.
I mean, isn't that what Bell Labs basically was? Put smart people in a supportive playground and harvest the cream off the top? Lasers, MOSFETs(?), UNIX, C, awk, all those came in one way or another from there.
Methinks Google got too impatient. Something like "hey we launched Google X, build it and the inventions will flow" but the Bell Labs model requires something Google was and is constitutionally incapable of rewarding: iteration, time passing, just plain putting in the boring work, year after year.
Google is forever more in hyper growth mode, like birthing a baby that doubles its size on X months. But an 8 year old is not double the size of a 6 year old like babies do. And same for a 10 year old, you need another 8 years (16 year old) to approximately double the size again, and that process assumes constant focus and maintenance of growth attempts so that the body doesn't have stunted growth.
Google seems to be great at launching new products, then utterly fails to iterate them, because of new shinyism or internal promotion incentives or something. I suspect that bled over into Google Labs / X / other company initiatives.
> I mean, isn't that what Bell Labs basically was? Put smart people in a supportive playground and harvest the cream off the top?
Was it? A lot of people here think the transistor was one of the most important inventions to come out of Bell Labs. Here's Wiki's description of its creation:
> In 1945, Bell Labs reorganized and created a group specifically to do fundamental research in solid state physics, relating to communications technologies. Creation of the sub-department was authorized by the vice-president for research, Mervin Kelly. An interdisciplinary group, it was co-led by Shockley and Stanley O. Morgan. The new group was soon joined by John Bardeen. Bardeen was a close friend of Brattain's brother Robert, who had introduced John and Walter in the 1930s. They often played bridge and golf together. Bardeen was a quantum physicist, Brattain a gifted experimenter in materials science, and Shockley, the leader of their team, was an expert in solid-state physics.
> According to theories of the time, Shockley's field effect transistor, a cylinder coated thinly with silicon and mounted close to a metal plate, should have worked. He ordered Brattain and Bardeen to find out why it wouldn't. During November and December the two men carried out a variety of experiments, attempting to determine why Shockley's device wouldn't amplify.
It hardly sounds like they just gave a bunch of smart people a playground and told them to have fun. The problem with many historical examples is that people sometimes only pay attention to the fun parts they want to hear (a lot of focus was put on engineers and they were given some leeway) and often ignore the hard work that went into it (a lot of management, structure, and focus). We get left with overly simplistic solutions that sound wonderful, but then fail when people try to implement them.
One thing everyone here always forgets about places like Bell Labs and Xerox parc, was that it was NOT a monoculture of developers working on things. These places encouraged collaboration between many different disciplines. They encouraged diversity.
What turned me off tech culture is not just that tech companies became more like banks. It's that there is a monoculture of thought, and highly arrogant thought at that. It's the place where developers are royalty and everyone else is regarded as mouth breathing servants. Of course, most companies, for damn good reason, are not structured with developers running everything, and so you get a constant stream of toxic comments complaining about how every other role at the company is useless and holding back the oh-so-brilliant developers.
That's not diversity, it's not inclusion, and it's not a recipe for actually changing the world for the better.
> I mean, isn't that what Bell Labs basically was? Put smart people in a supportive playground and harvest the cream off the top? Lasers, MOSFETs(?), UNIX, C, awk, all those came in one way or another from there.
In a somewhat-parallel to early Google, it's worth noting that Bell Labs was funded by, and operated in service to, a near-total national-scale monopoly. When that monopoly was deconstructed, it triggered a decline in the output of the research organization as well (there's a clear decline in the ground-breaking output of Bell Labs after the 1984 Ma Bell breakup).
The book "The Idea Factory: Bell Labs and the Great Age of American Innovation" - https://en.wikipedia.org/wiki/The_Idea_Factory describes how the Labs was working.
The last chapter tries to figure out why it was working and one of the conclusion is that the Labs was always guided by the long term idea that it might help the telephone company (and that the breakup removed a) the long-term funding b) some kind of direction )
Bell Labs worked during the Cold War, in a Keynesian economic system. They were focused at fundamental technology, not in creating an end-user service to conquer the world.
Bell Labs did that, but they worked in service to the rest of the Bell system that was precisely targetting "creating an end-user service to conquer the world". Citing Bell Labs as if they were "an entire company" is misleading - they were the R&D wing for what was essentially a monopoly in telecoms.
> I mean, isn't that what Bell Labs basically was?
More like Xerox PARC. I wish we could re-create the secret to how productive that engineer playground was. Yes, Xerox HQ (across the country) had little interest in monetizing their output but as an invention factory it was an immense success and we use their ideas every day.
But did Bell Labs put those out as products? There are things that have come out Google (Kubernetes, their contributions to Linux and containers, gRPC, HTTP/2-3, Bazel, Go, compression, etc.) that have been similarly useful for others to build businesses on, the same way Bell and Xerox's research stuff became the foundations for other things outside of those companies.
But the discussion here is around Google's inability to create good products, so I don't think the comparison is apt.
This is the real core of the problem. Bell Labs was populated by hackers who were perfectly content to earn double to triple what an average factory worker was getting for less toil while doing cool shit. FAANGs became too profitable and attracted people who would have been finance bros in the 90s. Engineers expected to be compensated like quants and directors like portfolio managers. You don't get a Skunk Works or DARPA from that. Moonshot basic research like that may or may not make money. Your fundamental motivating force has to be doing cool shit, not getting rich. If you want something like Google X to work, the culture has to be such that no one is going to jump ship when Netflix offers to double your cash compensation.
How do you get this? I don't even know. As far as I can tell, a lot of 20th century Skunk Works and Bell Labs type of stuff was motivated by heavy government investment and individual patriotism in trying to outdo Nazis and Soviets for the sake of free civilization. Bell himself was an interesting case. He only invented the telephone in the first place as part of his research into helping deaf people and he invested most of his riches from the invention back into research for helping deaf people. He rather presciently never even owned a telephone and refused to allow one in his personal workspace because of fear it would distract from the work.
Well one component that has changed there is that the marginal tax rate on the inflation-equivalent of a Google Director pay back then was 80-90%, so there wasn't really a point in enticing people with large salaries.
Also, the factory workers you mentioned could afford houses in a big city on a single income. Now a single FAANG eng struggles to do the same unless they get N promotions, and even moreso if you tack childcare costs on top.
You make a lot of good points, but was the failure of Google moonshots really the fault of the workers who were easily lured by comp, or management who were impatient about seeing material results? Seems like either could have been chasing higher returns.
Initiatives like Google Fiber flaming out were certainly not the fault of anyone at the company. Maybe for drastically underestimating the difficulty of the problem. Actually, maybe that’s a reoccurring issue with these moonshots.
I try to look at systems level more than blaming certain individuals. I'm sure a lot of engineers working on Google X projects honest to God cared deeply about the specific problem they were tackling. My own wife was once pretty close to taking a job there. But unless they're going to donate labor by working for free, they can't fund the projects themselves. The company has to be willing to lose money on most if not all of their work. I guess that worked for a while when ads were so profitable and no one else did them well enough. I may be misremembering my history, but as far as I understood, Bell Labs had to do this kind of public good work as a condition of being granted a national monopoly.
At the cultural level, though, during periods like the space race and cold war, there seemed to be a national zeitgeist such that companies, their owners, their directors, were all willing to sacrifice some profit to serve a common national mission. Maybe it was just the fact that the CEOs had grown up in a time when food was rationed and factories were ordered by federal fiat to produce war materiel.
I don't actually think this type of culture is completely dead, for what it's worth. But it's not exactly in vogue. I personally work in defense, I served in the Army, and I make more than enough money but quite a bit less than Netflix likely would have paid me, but I work on products that serve US strategic interests. Entire companies exist explicitly for this purpose, but they're not Google scale and never will be. I don't think very many people on Hacker News would even consider this good, though. Witness the backlash to OpenAI even being willing to work on military applications at all. The readership is so cynical that they can't imagine a military serving any purpose aside from valueless destruction. In the 20th century when militaries saved the world from German and Japanese conquest and mass genocide, that kind of thinking was a lot less prevalent.
I also don't think these are the only two options, either. People can also be motivated purely by the science itself or by the global wellbeing of all mankind. I just don't know that, historically speaking, those bring large-scale resource mobilization to bear quite as well as getting rich and ensuring the continued strategic dominance of a way of life represented by a country of birth you deeply believe in. As people come to believe less and less in their own countries, we seem to be left mostly with getting rich as the only viable option.
If I had to guess, it's because from the Iraq War to modern day border theater, popular image of the military ain't what it used to be. It also seems like tech intended for defense is adjacent to civilian intelligence and law enforcement use, of which there is also cynicism and fears of surveillance- see Palantir viewed as the commercial equivalent to PRISM, for instance.
The public has had substantial loss of faith in its institutions in the past couple of decades, and the military is not exempt from it. America's foreign policy has come a long way since WWII.
But I don't think all is lost. I'm pretty sure solidly civilian organizations like United States Digital Service and 18F have good reputations, much like the Postal Service does. But all of them receive a pittance of investment compared to defense spending.
And growth is not just in physical size but in intelligence. 40 year old me knows a whole lot more even than 20 year old me thought he did.
Sustainability is hard. Would Google still be a favorite if it had never gone beyond a handful of products? Well never know, but has chasing new shiny worked out?
I mean, isn't that what Bell Labs basically was? Put smart people in a supportive playground and harvest the cream off the top? Lasers, MOSFETs(?), UNIX, C, awk, all those came in one way or another from there.
Methinks Google got too impatient. Something like "hey we launched Google X, build it and the inventions will flow" but the Bell Labs model requires something Google was and is constitutionally incapable of rewarding: iteration, time passing, just plain putting in the boring work, year after year.
Google is forever more in hyper growth mode, like birthing a baby that doubles its size on X months. But an 8 year old is not double the size of a 6 year old like babies do. And same for a 10 year old, you need another 8 years (16 year old) to approximately double the size again, and that process assumes constant focus and maintenance of growth attempts so that the body doesn't have stunted growth.
Google seems to be great at launching new products, then utterly fails to iterate them, because of new shinyism or internal promotion incentives or something. I suspect that bled over into Google Labs / X / other company initiatives.