1. Most people ignore it.
2. Some people plink around on it, and maybe learn to tap out a few bars of a song with one finger.
3. Some people put in the time and energy to play it fluently.
One of my readers is in stage three, learning how to give artificial intelligence tools the prompts and guidance it needs to increase the speed and accuracy of letters and briefs coming out of his law practice. There was a steep learning curve, he tells me, but over the course of a year of intense effort to integrate AI, it went from slowing him down to half-speed, then to break-even. Now it doubles his productivity, and it is getting better every month. He thinks the tool will revolutionize the practice of law. It is already starting to do so.
He has urged me to let AI help write my blog posts. I haven't done so. I am at stage one.
Artificial intelligence is user friendly, but there is a learning curve to using it as an effective tool in doing serious work. It requires being willing to change the process for how one works. When one switches from pen and paper to a typewriter to send a letter, one doesn't use a pen to write the letter onto the side of the typewriter, and mail the typewriter. One completely abandons the pen and no longer touches the paper. The letter-writer needs to accept using a whole new process.
I don't feel ready to sit down every morning and prompt AI to write a blog post consistent with my thoughts as revealed in prior posts. So I don't. But maybe that is the future of opinion journalism. Like my guest post author, I am glad I am retired.
College classmate Erich Almasy went from college to the Harvard Business School and straight into management consulting. He shares his observations about resistance to change.
![]() |
Almasy |
Guest Post by Erich Almasy
Nothing Changes and Everything Stays the Same
In my first year as a management consultant, I was tasked with researching examples of significant industrial change. Curiously, despite industry innovation and even invention, I found that changes in industry leaders rarely occurred. In the 1940s and 50s, DuPont developed the first water-based house paint, Lucite®. While it revolutionized the paint industry, after the dust settled, the top paint companies like Sherwin-Williams and PPG remained the same. Similarly, electric furnaces using scrap metal instead of iron ore upended steel industry technology, but the Big Three steelmakers retained their positions. Radial tires from Michelin, personal computers from Altair, and even decaffeinated coffee from Kaffee Hag did little to displace the industry leaders because market share and consumer preference were slow to change. As long as the industry leaders stay in power, they have little incentive to redefine the work they do. To move the needle, it seems that the creation of a new industry is needed, such as semiconductors, solar panels, email, and electric vehicles. Even then, widespread adoption typically takes decades.
Recently, it was reported that eight of ten companies are using AI, but that they have not noticed a jump in productivity. This parallels the early years of the so-called personal computer and Internet revolutions, both hailed as productivity game changers but subsequently seen as largely sizzle and no steak. At the heart of both these dilemmas is the simple fact that improving productivity requires significantly changing the nature of the work involved, and that task is challenging. As Ron Frank, my HBS Professor and good friend, used to say, “work is the way it is because that’s the way it is.” As a consultant, I found it’s tough to change what works, and until one does, the existing patterns, designs, and habits of work processes remain the same. Work processes are highly resistant to new technology when it is only layered on. The status quo means “the state in which,” and it requires significant change to overcome inertia (a tendency to do nothing or remain unchanged).
To make work different is very risky and expensive. Witness Ford Motor’s recent announcement that it will spend billions to end linear assembly of vehicles in favor of a “tree” assembly, making front, middle, and end portions of its cars on separate lines and bringing them together at the end. The middle will be a battery platform that will be used in all EV models Ford makes. This radical redesign of work and huge change for Ford may dramatically reduce labor and parts costs. Like the Model T assembly line, the Springfield rifle’s interchangeable parts, and the division of labor in Adam Smith’s pin factory, radical redesign of work causes radical change.
The aspirational claims for AI so far are that it will produce a fundamental change in how white-collar workers handle everyday management tasks. Estimates range up to the replacement of up to fifty percent of these workers. Careers similar to the one I had in management consulting will change dramatically. An even more spectacular opportunity is claimed for medical research and the development of new treatments or drugs. However, present AI results include gay and anti-Semitic slurs, misinformation, plagiarism, outright lies, and points of view that are bent to specific politics. To the point that it is unclear to what extent this technology can be trusted. When AI writes your research report, you may spend more time checking its facts than writing it yourself. Or not. We are in an age where children, teens, and young adults are looking at social media and texting on their smartphones rather than reading or writing. Why not let a computer app do all your communication for you? Sorry, I’m old school. I still write in cursive. I will obstinately continue to communicate using only Grammerly and spell-check. I also think that breakthroughs like Ford’s will come more from human effort than computer programs. I’m also glad that I’m retired.
Nothing Changes and Everything Stays the Same
In my first year as a management consultant, I was tasked with researching examples of significant industrial change. Curiously, despite industry innovation and even invention, I found that changes in industry leaders rarely occurred. In the 1940s and 50s, DuPont developed the first water-based house paint, Lucite®. While it revolutionized the paint industry, after the dust settled, the top paint companies like Sherwin-Williams and PPG remained the same. Similarly, electric furnaces using scrap metal instead of iron ore upended steel industry technology, but the Big Three steelmakers retained their positions. Radial tires from Michelin, personal computers from Altair, and even decaffeinated coffee from Kaffee Hag did little to displace the industry leaders because market share and consumer preference were slow to change. As long as the industry leaders stay in power, they have little incentive to redefine the work they do. To move the needle, it seems that the creation of a new industry is needed, such as semiconductors, solar panels, email, and electric vehicles. Even then, widespread adoption typically takes decades.
Recently, it was reported that eight of ten companies are using AI, but that they have not noticed a jump in productivity. This parallels the early years of the so-called personal computer and Internet revolutions, both hailed as productivity game changers but subsequently seen as largely sizzle and no steak. At the heart of both these dilemmas is the simple fact that improving productivity requires significantly changing the nature of the work involved, and that task is challenging. As Ron Frank, my HBS Professor and good friend, used to say, “work is the way it is because that’s the way it is.” As a consultant, I found it’s tough to change what works, and until one does, the existing patterns, designs, and habits of work processes remain the same. Work processes are highly resistant to new technology when it is only layered on. The status quo means “the state in which,” and it requires significant change to overcome inertia (a tendency to do nothing or remain unchanged).
To make work different is very risky and expensive. Witness Ford Motor’s recent announcement that it will spend billions to end linear assembly of vehicles in favor of a “tree” assembly, making front, middle, and end portions of its cars on separate lines and bringing them together at the end. The middle will be a battery platform that will be used in all EV models Ford makes. This radical redesign of work and huge change for Ford may dramatically reduce labor and parts costs. Like the Model T assembly line, the Springfield rifle’s interchangeable parts, and the division of labor in Adam Smith’s pin factory, radical redesign of work causes radical change.
The aspirational claims for AI so far are that it will produce a fundamental change in how white-collar workers handle everyday management tasks. Estimates range up to the replacement of up to fifty percent of these workers. Careers similar to the one I had in management consulting will change dramatically. An even more spectacular opportunity is claimed for medical research and the development of new treatments or drugs. However, present AI results include gay and anti-Semitic slurs, misinformation, plagiarism, outright lies, and points of view that are bent to specific politics. To the point that it is unclear to what extent this technology can be trusted. When AI writes your research report, you may spend more time checking its facts than writing it yourself. Or not. We are in an age where children, teens, and young adults are looking at social media and texting on their smartphones rather than reading or writing. Why not let a computer app do all your communication for you? Sorry, I’m old school. I still write in cursive. I will obstinately continue to communicate using only Grammerly and spell-check. I also think that breakthroughs like Ford’s will come more from human effort than computer programs. I’m also glad that I’m retired.
[Note: To get daily delivery of this blog to your email go to: https://petersage.substack.com. Subscribe. Don't pay. The blog is free and always will be.]
6 comments:
The first field of white collar work that AI will massively disrupt is web development and coding. It already is in fact. I have multiple friends who work in this field and to a person they have all seen their work outputs increase by drastic amounts thanks to integrating AI into their work flow. 50% increase in productivity is the minimum from any of my 5 buddies who do this. Others are reporting increases of more like 3-5x. And this is all literally in the last 6 months that the AI Agents have gotten significantly better.
The first people to not get jobs will be all the kids going to school right now studying web development and coding. There will not be an entry level white collar job on the other side of their degree. They will however have immense student loans that will need to be repaid.
Then it will happen in other fields as well. It is almost impossible to understate just how disruptive AI will be in many job fields in the extremely near term future. And you can be certain but definitely not reassured that Trump will have zero fucking clue how to handle it.
Using AI for creative work is like paint by number, and I think it risks draining the individuality of expression, and may limit the incidence of true insight. MEANWHILE MORE VICTIMS ARE COMING FORWARD AND THE DISTRACTIONS ARE REALLY STARTING TO LOOK DESPERATE. FLAG BURNING? ARE YOU KIDDING? THE DOJ (TRUMP CRIMINAL DEFENSE LAWYERS) RELEASED 30K DOCS WITH LITTLE NEW RELEVANT FACTS. STONEWALL!!
I’m a believer that the effect of AI on our lives is underestimated in the next ten years. If I’m still alive in ten years, I’m guessing I will have a robot of some kind helping out, medical advances will be profound, and something I have no clue of will be significant. The spike of change is vertical.
A nice piece of economic history by Jeremy Greenwood, "The Third Industrial Revolution", shows that in the previous two cases of railroads and electrification, economy-wide productivity did in fact increase, but only with a long lag. The IT revolution, he argues, is similar.
https://ideas.repec.org/a/fip/fedcer/y1999iqiip2-12.html
Prior to a wide measurable increase, productivity was reflected in growing economic inequality, the salaries of those who work *with* the new technology rising faster than others, while many being are actually displaced by it, and thus poorer.
Essentially, new productivity earnings flow to a few, and away from the many who are replaced. So it more or less evens out for no overall GDP gain, until the new technology becomes much more widely disseminated and integrated into most people's work.
Yet another reason for young people to hate or be jealous of Boomers:
https://www.wsj.com/economy/jobs/ai-entry-level-job-impact-5c687c84?mod=hp_lead_pos9
I had a great career as a software developer. If I were 20 now, I don’t know what would be a good field for me to go into. I think I would seriously consider going into the military, probably the Air Force.
Post a Comment