I think this may well be the biggest flaw in the book-as-printed. And the very sharp Arthur Goldhammer agrees, and sends a withering, accurate critique
As a veteran of 25 years at IBM Research, I felt the same disappointment. And not only is there more to the story of the Industrial research lab than Edison vs. Tesla, it's worth (and has been) the subject of a full book by itself. Some that material may have been left on the editing room floor along with the 400 pages cut from the earlier draft. Edison vs Tesla is indeed an interesting story. There was an interesting prior, private lab in Tuxedo Park that contributes some flavor to the eventual Los Alamos effort and Manhattan Project, but Los Alamos was only a part of the story. Major wartime labs were created at Hanford, Washington and Oak Ridge, Tenn. Supporting efforts at Berkeley, Livermore, and Sandia continued into the Cold War era. A parallel effort in radar and control spawned Labs in the MIT/Harvard area which survive as Draper Lab and Lincoln Labs. The launch of the computer era took place in not only Bell Labs for the transistor, (followed quickly by IBM and Texas Instruments' contributions). But computers were built at IAS, Penn, IBM, and MIT before settling into something like their current forms and architectures. Operating systems arose both in IBM and in the Boston-Berkeley axes that led to Unix and finally, through a global effort, to Linux. (Sorry, left out UK contributions, which were quite significant.) There were some major shifts in style and focus from defense emphasis to commercial ("tabulating machines") to ever smaller and more distributed computing "fabrics" of today. While these started out in industrial labs of the conventional sort, the rise of a startup economy shifted the center to firms like Apple and Google, whose dynamics are completely different from those of GE, ATT, IBM, and Microsoft..
It's a long story, not the one you started out to tell, but the change in the tech industry is an important part of your final questions about what comes next, post 2010.
I am just rereading Hobsbawm and that, together with Scott's comment below, bring up the whole issue of how much the wartime labs contributed to the take-off in technology that is so much of Brad's story of the twentieth century. The huge gaping hole in the right wing version of economic history is the role of government (and not just the modern corporation or Hayekian visions of markets) in economic takeoff. It also raises the question of how much the subcontracting of research to private corporations (as described by Scott) contributes to the takeoff of both.
It wasn't just the industrial research lab. It was also the rise of the university. One couldn't just move to the frontier by studying tea kettles. There was a lot more to learn, and it was stuff that needed to be formally taught. Innovation was increasingly driven by theory. The era of the gifted and devoted amateur was ending and a new era of collective innovation had started. It is no surprise that the modern corporation came into its own as a means of collective endeavor.
This means that you have another book to write. (You might consider looking into what mathematicians call the "coalition model of growth".)
It's a fruitful criticism if it gets us another book. Well-handled!
I'm still mulling Helen Thompson's thesis (in last winter's "Disorder" and many short pieces since) that all our technological progress gave us a new piper to pay - a never-ending energy bill, along with the never-ending problem of sourcing it. It would be interesting if you read the book, because not many economists have - yet her argument needs to make economic as well as geopolitical sense, and I'm still trying to figure out if it does. The contrast couldn't be greater though: her compressed vs your expansive style; your grand but disruptive growth story, her account of the balancing act required by multiple disruptions.
The energy problem is an old one. Wasn't one big hope for the Virginia colonies that they would use the local forests for glass manufacturing? I'm guessing Venice became a glass manufacturing center because it had good marine access to firewood.
That's why it is important that research and development of zero CO2 emitting energy continues so that while increasing the cost to the user of CO2 emitting energy. we keep reducing the total cost of energy use per capita.
Probably so. But I'd put it in the more general context of what was (NOT) driving employment of technology in every epoch. But if you do, you set yourself up for having to ask why it seemed to stop in 1973 (?)
I also found the small coverage of IRLs to be frustrating - it is such a tantalizing aspect of the 1870-onwards grand narrative arc that you set out with. The same goes for the modern corporation. As a sociology and market research professional I'm really intrigued by both of these grand themes you laid out in the first chapters of the book. Regarding the link between the industrial research labs and modern corporation, I think you are pointing to the role modern corporates played in moving high-level abstract innovation/discoveries through to "product-market fit": seeing market applications for these abstract discoveries by scientists/psychologists/etc. THAT would be a very interesting follow-up paper I would like to read in Substack when you publish it. In contrast, I absolutely engrossed in your exposition of the Hayek vs Polanyi discussion - and your points about Social Democracy vs Neoliberalism. It's very refreshing to see a critique of "the market" from someone who also sees its unique ability to generate the 1870-2010 increase in standards of living. The whole book is a masterpiece!
1. Is there a definitive work on the history of the industrial research lab? An analogue to Chandler (Visible Hand) or maybe Drucker?
2. Separately, I thought the contents of this tweet thread by Kamil Galeev deserved to be more widely known. Stalin, who recognized the possibilities of running HR in the Soviet Union, didn't read memos. He spent most of his time eavesdropping on telephone conversations of other party chiefs. From Bazhanov: "I was Stalin's secretary": https://twitter.com/kamilkazani/status/1516825897982214145
There are lots of histories of industrial labs, written from the inside with corporate support. See, for example Marc deVries on Philips and the many volumes written and edited by Emerson Pugh and Paul Ceruzzi on IBM. The chronicle of Bell Labs, from the Golden years to its eventual trivestiture and disappearance exists in the forms of several autobiographies from the Unix group. Look for Brian Kernighan. But none of these gives the God's Eye view of what allowed these worlds to flourish for several decades before turning into specialized services in the current century. The business factors that shortened the time frames in the big companies' labs, and narrowed the scope of technology explored in the startups are internal information, I suspect. I would love to see an investigative study comparing the story arc of Bell Labs and, say, Google. The time scale for the growth from pure science to research as a service has been greatly compressed, but is it the same arc? And have overlapping, disposable smaller companies provided the same function as the big ones used to?
Bill Janeway blames the death of innovation in the traditional big American corporation on a change in regulations that allowed spare internal funds to return more to the stockholders, and sooner, through stock buybacks than investment in research could. This evolved during the period when stock options became standard compensation for executives and high performers. I was at IBM through the near-death and Lou Gerstner turnaround and experienced the changes in internal culture that resulted. I think this was a powerful factor.
The short answer is that research to drive the information ecology ahead is now so expensive and specialized that it is starting to look like high energy physics, which can be done only in special places with long lead times and huge budgets. Of course, once such a facility is in place (think Google's labs, Open AI, DeepMind) then those with access to the inner sanctum can create a few amazing tools, like DeepMind's protein structure synthesis, or the various chatBOTs that are clogging the airwaves right now. Simpler things get done during public open hours, and a host of smaller innovations are happening as droids in the back rooms optimize advertising success and try to manage privacy and social interactions.
But this world is a long way from Edison, Tesla, which were run very top down, and the greater freedom that a few enjoyed in the salad days of Bell Labs and early IBM Research. IBM Research has survived by transforming itself into a consulting service during the Gerstner years, I think. I haven't worked there since 2000, so this is just an impression. Bell Labs never had much understanding or direct involvement in ATT's upper management, so they were blindsided by the various 'fectas and disappeared. I think Xerox, GE, Westinghouse, etc., never really evolved beyond the Edisonian stage.
Noah's ruminations on what it would take to create a new Bell Labs are on target. He points out that the challenges of energy and climate don't find a large protected industry that could own them, so the Bell Labs ideal doesn't fit. I think carbon extraction into a profitable product like nylon might be the exceptional case.
I clearly need to write something about the science -> IRL -> corporation -> value chain -> general diffusion transmission path over the past century and a half, paying due attention to the roles of governments, startups, artisanal specializers, and regulators...
There are two different flows to follow -- intellectual diffusion, carried out largely by people moving from one environment to another as startup investments blow up and get absorbed elsewhere; and (Janeway's point) the flow of funds from existing commercial successes, whether or not they are actual monopolies, into the Next Big Thing. There was a regulatory change behind that. I'll be interested to see if the Deep Mind protein structure predictor makes it across the walls that separate machine learning people from "real" medical biochemists in the pharma business.
As a veteran of 25 years at IBM Research, I felt the same disappointment. And not only is there more to the story of the Industrial research lab than Edison vs. Tesla, it's worth (and has been) the subject of a full book by itself. Some that material may have been left on the editing room floor along with the 400 pages cut from the earlier draft. Edison vs Tesla is indeed an interesting story. There was an interesting prior, private lab in Tuxedo Park that contributes some flavor to the eventual Los Alamos effort and Manhattan Project, but Los Alamos was only a part of the story. Major wartime labs were created at Hanford, Washington and Oak Ridge, Tenn. Supporting efforts at Berkeley, Livermore, and Sandia continued into the Cold War era. A parallel effort in radar and control spawned Labs in the MIT/Harvard area which survive as Draper Lab and Lincoln Labs. The launch of the computer era took place in not only Bell Labs for the transistor, (followed quickly by IBM and Texas Instruments' contributions). But computers were built at IAS, Penn, IBM, and MIT before settling into something like their current forms and architectures. Operating systems arose both in IBM and in the Boston-Berkeley axes that led to Unix and finally, through a global effort, to Linux. (Sorry, left out UK contributions, which were quite significant.) There were some major shifts in style and focus from defense emphasis to commercial ("tabulating machines") to ever smaller and more distributed computing "fabrics" of today. While these started out in industrial labs of the conventional sort, the rise of a startup economy shifted the center to firms like Apple and Google, whose dynamics are completely different from those of GE, ATT, IBM, and Microsoft..
It's a long story, not the one you started out to tell, but the change in the tech industry is an important part of your final questions about what comes next, post 2010.
I am just rereading Hobsbawm and that, together with Scott's comment below, bring up the whole issue of how much the wartime labs contributed to the take-off in technology that is so much of Brad's story of the twentieth century. The huge gaping hole in the right wing version of economic history is the role of government (and not just the modern corporation or Hayekian visions of markets) in economic takeoff. It also raises the question of how much the subcontracting of research to private corporations (as described by Scott) contributes to the takeoff of both.
Yes indeed: big questions, and important ones...
It wasn't just the industrial research lab. It was also the rise of the university. One couldn't just move to the frontier by studying tea kettles. There was a lot more to learn, and it was stuff that needed to be formally taught. Innovation was increasingly driven by theory. The era of the gifted and devoted amateur was ending and a new era of collective innovation had started. It is no surprise that the modern corporation came into its own as a means of collective endeavor.
This means that you have another book to write. (You might consider looking into what mathematicians call the "coalition model of growth".)
It's a fruitful criticism if it gets us another book. Well-handled!
I'm still mulling Helen Thompson's thesis (in last winter's "Disorder" and many short pieces since) that all our technological progress gave us a new piper to pay - a never-ending energy bill, along with the never-ending problem of sourcing it. It would be interesting if you read the book, because not many economists have - yet her argument needs to make economic as well as geopolitical sense, and I'm still trying to figure out if it does. The contrast couldn't be greater though: her compressed vs your expansive style; your grand but disruptive growth story, her account of the balancing act required by multiple disruptions.
The energy problem is an old one. Wasn't one big hope for the Virginia colonies that they would use the local forests for glass manufacturing? I'm guessing Venice became a glass manufacturing center because it had good marine access to firewood.
That's why it is important that research and development of zero CO2 emitting energy continues so that while increasing the cost to the user of CO2 emitting energy. we keep reducing the total cost of energy use per capita.
But fortunately, it's not a CO2 emissions bill. Properly priced by Prof. Pigou, CO2 emissions can be squashed w/o much effect on energy use/capita.
Probably so. But I'd put it in the more general context of what was (NOT) driving employment of technology in every epoch. But if you do, you set yourself up for having to ask why it seemed to stop in 1973 (?)
No problem - I'd LOVE to have you point us readers to more on that topic......Even a 5-10 page essay would be fantastic......
Oh. I see. Lowering the bar, we are...
I also found the small coverage of IRLs to be frustrating - it is such a tantalizing aspect of the 1870-onwards grand narrative arc that you set out with. The same goes for the modern corporation. As a sociology and market research professional I'm really intrigued by both of these grand themes you laid out in the first chapters of the book. Regarding the link between the industrial research labs and modern corporation, I think you are pointing to the role modern corporates played in moving high-level abstract innovation/discoveries through to "product-market fit": seeing market applications for these abstract discoveries by scientists/psychologists/etc. THAT would be a very interesting follow-up paper I would like to read in Substack when you publish it. In contrast, I absolutely engrossed in your exposition of the Hayek vs Polanyi discussion - and your points about Social Democracy vs Neoliberalism. It's very refreshing to see a critique of "the market" from someone who also sees its unique ability to generate the 1870-2010 increase in standards of living. The whole book is a masterpiece!
Yes. Skimpy coverage of IRLs is a serious flaw in the book...
Brad
1. Is there a definitive work on the history of the industrial research lab? An analogue to Chandler (Visible Hand) or maybe Drucker?
2. Separately, I thought the contents of this tweet thread by Kamil Galeev deserved to be more widely known. Stalin, who recognized the possibilities of running HR in the Soviet Union, didn't read memos. He spent most of his time eavesdropping on telephone conversations of other party chiefs. From Bazhanov: "I was Stalin's secretary": https://twitter.com/kamilkazani/status/1516825897982214145
There are lots of histories of industrial labs, written from the inside with corporate support. See, for example Marc deVries on Philips and the many volumes written and edited by Emerson Pugh and Paul Ceruzzi on IBM. The chronicle of Bell Labs, from the Golden years to its eventual trivestiture and disappearance exists in the forms of several autobiographies from the Unix group. Look for Brian Kernighan. But none of these gives the God's Eye view of what allowed these worlds to flourish for several decades before turning into specialized services in the current century. The business factors that shortened the time frames in the big companies' labs, and narrowed the scope of technology explored in the startups are internal information, I suspect. I would love to see an investigative study comparing the story arc of Bell Labs and, say, Google. The time scale for the growth from pure science to research as a service has been greatly compressed, but is it the same arc? And have overlapping, disposable smaller companies provided the same function as the big ones used to?
Bill Janeway blames the death of innovation in the traditional big American corporation on a change in regulations that allowed spare internal funds to return more to the stockholders, and sooner, through stock buybacks than investment in research could. This evolved during the period when stock options became standard compensation for executives and high performers. I was at IBM through the near-death and Lou Gerstner turnaround and experienced the changes in internal culture that resulted. I think this was a powerful factor.
Interesting. Please tell me more...
The short answer is that research to drive the information ecology ahead is now so expensive and specialized that it is starting to look like high energy physics, which can be done only in special places with long lead times and huge budgets. Of course, once such a facility is in place (think Google's labs, Open AI, DeepMind) then those with access to the inner sanctum can create a few amazing tools, like DeepMind's protein structure synthesis, or the various chatBOTs that are clogging the airwaves right now. Simpler things get done during public open hours, and a host of smaller innovations are happening as droids in the back rooms optimize advertising success and try to manage privacy and social interactions.
But this world is a long way from Edison, Tesla, which were run very top down, and the greater freedom that a few enjoyed in the salad days of Bell Labs and early IBM Research. IBM Research has survived by transforming itself into a consulting service during the Gerstner years, I think. I haven't worked there since 2000, so this is just an impression. Bell Labs never had much understanding or direct involvement in ATT's upper management, so they were blindsided by the various 'fectas and disappeared. I think Xerox, GE, Westinghouse, etc., never really evolved beyond the Edisonian stage.
Noah's ruminations on what it would take to create a new Bell Labs are on target. He points out that the challenges of energy and climate don't find a large protected industry that could own them, so the Bell Labs ideal doesn't fit. I think carbon extraction into a profitable product like nylon might be the exceptional case.
I clearly need to write something about the science -> IRL -> corporation -> value chain -> general diffusion transmission path over the past century and a half, paying due attention to the roles of governments, startups, artisanal specializers, and regulators...
There are two different flows to follow -- intellectual diffusion, carried out largely by people moving from one environment to another as startup investments blow up and get absorbed elsewhere; and (Janeway's point) the flow of funds from existing commercial successes, whether or not they are actual monopolies, into the Next Big Thing. There was a regulatory change behind that. I'll be interested to see if the Deep Mind protein structure predictor makes it across the walls that separate machine learning people from "real" medical biochemists in the pharma business.