Chairman and Founder, Atypon

Interview conducted by Michael Upshall

Against the Grain V38#4 

This is a lightly edited transcript of a conversation between Michael Upshall, Editor, Charleston Briefings and Guest Host for ATG the Podcast, and Georgios Papadopoulos.  

Georgios Papadopoulos is well known as the founder and former CEO of Atypon, a publishing platform used by many of the world’s scholarly journals.  In this interview, he looks back over his education in Greece and remarkable career in the U.S.

The full interview can be viewed at:  https://www.charleston-hub.com/podcast/atgthepodcast-331-a-conversation-with-georgios-papadopoulos-chairman-and-founder-atypon/

Michael Upshall:  Georgios, thank you for joining.  Let’s start with your background.  Where in Greece were you born?

Georgios Papadopoulos:  I was born in Thessaloniki, and I studied there in the math department before I moved to the U.S. to get a PhD in computer science.  My first degree was math, although when I was in school, I wanted to study astrophysics.  But then I went to work in the summer before university.  I worked in a camp, and I met a physicist who told me in the physics department they had a lot of mandatory classes and labs, and instead, the math department didn’t have any, and these math students never went to any classes.  That’s when I switched to math.  I took advantage of it liberally.  I didn’t attend any classes at all the first two years of the math department.  And the second year I didn’t even go to take the books.  I went to take the books at the end of the year.  We didn’t have semesters in Greece.  We had whole years.  And I took them a week before the exams.  And somehow, I was always going pretty well.  And then the third year there was a transformation when I had a professor who came from the U.S., and he taught two classes in a kind of computer science.  One was numerical analysis, and the other was a computer language, and I was fascinated by it;  I decided to become a computer scientist, which was of course impossible in Greece at the time, because we didn’t have a computer science department.  That was back in 1983:  the university had a mainframe, with five punched card machines.  You would write the program, give it to the people, and the next day they would tell you that you a typo.  You had to submit all over again. 

Eventually, I tried, with a friend of mine, to buy a computer of our own, which was too expensive at the time.  But we found a computer in a physics lab, and we proceeded to learn first to type, because we didn’t even know how to type, and then to program little things.  Then, in the fourth year of school, we went to this great professor from the university, from the U.S., and we asked him to design a class for us, which he did.  It was pretty bold — never done before.  And which he did!  He designed a class for the full year.  We only had to do six projects, six pretty tough projects, and he gave us no instruction, nothing, no help of any sort.  But we discovered a stack of computer science books at this physics lab, which belonged to a professor who was previously a physicist at CERN.  So we figured out the solution to the problems.  And we took the GRE exam in order to come to the U.S. as a graduate student.  We took the GRE (Graduate Record Examination) in computer science instead of taking it in math — we thought we would show them that we actually do know computer science, even though we had no computer science classes in our degree.

We were accepted at NYU and we finished our PhD.  When I started the company (Atypon), this other guy, who was my friend from forever, became my first employee.  And he still works at Atypon, which is great.

MU:  So essentially, your whole degree was self-directed learning.

GP:  Yes, it was self-directed learning, and I think that’s the best way to do it.  If you find your calling, then you devote your entire life onto it.  I was leaving home at eight o’clock and returning at midnight, because we were programming all the time.  That’s what we were doing, seven days a week.  The university building was locked up during the weekends, and this professor, the physicist, who was also the dean, gave us the keys to the school, so that we could get in.  By the way, we were protecting a secret, which is there was an available Apple IIe in his office.  We were protecting it like gold because we didn’t want anybody else to take time on it.  So, we would do all kinds of subterfuges to try to get into his office, because we didn’t want anybody to see us.  It was quite a thing.  Anyway, we came to the U.S., and we were pretty successful in our studies.  We aced everything! 

MU:  What was the subject of your PhD?

GP:  Well, the subject was interesting.  I was fascinated by compilers, transformers essentially, compilers and data flow analysis.  That proved crucial down the line, for the first journal that went online in a very strange way.  So my PhD was in compilers, data flow analysis for logic programming languages.  It was an interesting subject at the time.

The computer language Prolog was supposed to be the AI engine of the future, of symbolic reasoning.  I was early into that, and I was fascinated by making things efficient, because that was one of the problems of adopting symbolic computing.  Of course, like every computer science graduate, even now I think, math and computer science, physics, they all write their theses in LaTeX, which was the second important thing that helped down the line.  It’s very interesting, as Steve Jobs said, you can only connect the dots looking backwards.  You don’t know whatever you’re doing at the time that is going to become important later in your life.  So I had to do everything in LaTeX and then I was very good in programming during university essentially.  And I was doing transformers.

MU:  At some point you won a scholarship?

GP:  Yes, the scholarship was from day one.  I got a few scholarships.  When I came into the country, I came as a Fulbright scholar, which had repercussions down the line for me afterwards.  The Fulbright actually paid very little.  It was just giving you $5,000 for the first year.  It was more of an honorary thing that you were selected as one of the best that Greece basically has to offer.  And the university, the computer science department, gave me a full research assistantship.  I didn’t have any tutoring duties — that was because I was already doing research as an undergrad in logic programming back in the days.  Then I got another full scholarship from the economics department.  I was building multiplayer games for them, economics games.  They needed to simulate different economic theories with real players in real time.  They had a local network at the time and computers on the network, but there was no foundation or nothing really to program these multiplayer games at the time.  So I created a platform, and I was simulating games, I was programming on top of the platform, whatever games they were asking us to do.  That was important because they were spending real money.  If there was a bug, they were out some money.

MU:  I believe your military service came after that. 

GP:  Yes, one of the requirements of the Fulbright Scholarship for all the $5,000 that they gave us was that we had to go back for two years to our country to contribute to it.  And in Greece, of course, we have military service, so I joined the Navy, or I was assigned to the Navy, which was great.  I absolutely wanted that.  I actually wanted to be in the submarines, but they didn’t allow any draftees to be in the submarines.  I had, of course, a computer science PhD and a math degree.  They put me as a secretary to the commander-in-chief of the Navy school and also teaching four hours a day of English because I also had a degree that in Greece would allow me to teach English!  Teaching English was a great experience.  Not so great for my students, unfortunately, because it was kind of harsh.  In the Navy, if you don’t have certain grades, then you cannot go out — you have to stay in.  All my students were staying in pretty much.  But they did learn English!  It was very important for them to learn English because all the manuals of the ships and stuff like that are written in English. 

MU:  Did you enjoy your Navy career?

GP:  I was very bored because I was away from home, I was away from my family, I had just married, I was away from everybody.  Because I was very bored, I was coming up with different things to take away the boredom.  One of the things was finding out my limit of going out without sleep.  As I was a petty officer, I didn’t have any guard duty, so to make myself stay awake, I would go and replace whichever guard was there.  It’s always two hours.  When the next one would come to replace me, I would say, no, no, go sleep, I’ll just stay here.  Then, during the day, I was doing everything that I needed to do.  The next night I would do exactly the same.  I found out I could go out without sleep for 72 hours, for three days, for two nights and three days.  And then I would stop functioning.  That’s an important detail later on for scholarly publishing.  Interestingly enough, that was the most important detail of my Navy duty.  I did other things, but that was probably the most extreme I had done. 

MU:  Did the Navy give you any insights into how to manage organizations?

GP:  Yes, the Navy did tell me how not to manage, so it was a very interesting experience.  The Navy taught me what not to do to motivate people.  And perhaps some things that I should do, but of course, discipline is one thing that you learn.  And I could understand why.  I came from a family that always put very high value on discipline and duty and all these things.  But the important thing was how did you actually motivate people to do things?  And what happens when people are unmotivated?  So, it was actually very important in terms of management and leadership skills.  There are good commanders, obviously, and bad commanders;  I think you learn a lot more from the bad ones, about what to avoid, because the good ones usually do things pretty well and you almost don’t notice.  That’s pretty much what goes on with leadership.  When things go well, everybody takes credit.  They don’t realize that the person who was at top is doing what he’s doing, and everybody shares credit.  And when things go badly, then the leadership weaknesses are revealed.

I finished Naval service and then came back to the U.S.  I had some ideas to start a business in Greece, but there was no financing back then.  There were no VCs (venture capitalists) in 1994! 

I joined a new startup at the time, WAIS, founded by Brewster Kahle, who you also interviewed.  At that time, the internet already existed in universities, but the Web was coming up, and WAIS was building these websites.  I worked on several websites with Brewster and eventually came to work for the first scholarly publishing website.  And it was kind of fortuitous because I had joined for less than a month, a month maybe, and I had already worked on a couple of websites.  The thing that bothered me was that I’d never built a website by itself.  I was always solving some particular problem about different things.  So Brewster called me one day and he said, let’s go for a walk.  Brewster had many good things as a manager, as a leader, I would say.  So we went for a walk, and he said, well, we have this customer, which was Stanford University, who are doing a journal for a publisher.  This is gonna be the first journal that goes online.  I mean, everything that we were doing at the time was the first.  I mean, we were doing Dow Jones with satellite data.  I worked on the first Intel site that was using SGML.  I worked on Encyclopedia Britannica, the New York Times, they were big sites.  There were so many firsts!  He said, well, you know, we put two teams on it:  I put two guys that worked on it for six months but couldn’t do it.  Then I put another two guys, and they couldn’t do it.  And do you think you can do that?  I said, okay, I’ll look at it.  And he said, well, you know, they have a big conference, and we told them that it’s ready.  Almost ready.  Now, they are very excited to show it, to show the first journal online.  And he said, you don’t need to do the whole journal, just a few articles, the whole issue, just a few articles, just to show them that we have something.

Of course, the problem back then was that we had these old browsers, right?  Mosaic, Netscape 0.9, I don’t remember now what they were.  It was early ’95, I think, because my son had just been born.  He said the problem is all these math symbols.  And I said, OK, when is this conference?  In three days.  So, it’s in three days!  I looked at what they had, and basically, they didn’t have much.  They had just found a master’s thesis by a Greek student in UK in Leeds, if I remember, who was doing a transformation of HTML to LaTeX.  It was not like that.  We were getting HTML, so I had to write basically a transformer from HTML, or from SGML to LaTeX.  I took his tool, but then I modified the LaTeX source code to output the images and create PNGs out of the formulas and symbols of math symbols, and stuff like that.  Then I stuck them back into the HTML.

I worked on it for three days, and two nights, and at the end, when I was just falling asleep on my chair, it was all done.  I went to sleep.  24 hours later, my wife woke me up and said, some people from Stanford are calling you and they invited you to a dinner.  What have you done?  I’m like, I don’t know, I did this journal.  Apparently, it was the first journal, the Journal of Biological Chemistry, the first journal that went online, with full text HTML.

I continued supporting it, but a month later, Brewster sold the company to America Online, which wanted to shut down all the projects.  I basically continued supporting it and started my own company, which is another whole story.  Atypon started as a company that was doing transformations of the SGML we were getting from the publishers to full-text HTML but for scientific articles.  I was the right person, at the right time, with all the right skills.  I had the transformer.  I could do transformers very, very fast.  I knew LaTeX very well, and I could program for three days without sleep.  That was basically how it worked out, and that’s how the first journal went online.  And after that, the platform grew to become big:  by the end, we had some 500 people, 350 product development engineers, and more than half the scholarly publications in the world being published on our platform.

MU:  So, from your very first software project you were involved in, you were offering a delivery date before you’d built the product.

GP:  Yeah, so we were doing only transformations, right?  We weren’t doing the platform yet.  I hired a salesperson with a little money that I was getting to get me more publishers to do transformations.  We moved to typesetting files because the SGML we were getting was pretty low quality.  I had only one customer at the time, Stanford (which became HighWire Press, by the way).  And he arranged a meeting with Academic Press down in San Diego.

The night before the meeting, I looked up their website, just out of curiosity.  It was like, my God, it was horrible.  So next morning, we fly down to San Diego, and we give a presentation about our services.  At the end of the meeting, I remember Larry Alexander, was, Metzner was the big guy, VP at the time.  Very nice guys.  And I said, by the way, you know, can I tell you a few things about your website?  He said, sure, go ahead, and I proceeded to tell them my vision of how a website should be made, should be done.  I had already worked on websites using WAIS, and I had developed a vision.  WAIS never went without.  And I had also developed a tool back then, a prototype of what would later be known as XSL.  Basically, for every website, we would take a piece of code, for example a registration login page.  Everything was done using print statements inside the code, so to do another registration page for another site, you had to take that code and change it and combine it all over again and go through debugging.  Every time you wanted to change something, you had to change code, which is crazy.  So, I created essentially a templating system, but I had the vision of agent-based computing and distributed modules and all of these things.  So I gave them that vision.  I didn’t even have slides!  This was less than slideware, total paperwork!  They said, when are you leaving?  I said, well, we have a flight in an hour, and they answered, can you change your flight, go and have lunch and come back after lunch?  We want to hear more about this.  So we had the lunch, came back and they said, tell us more, tell us what you would do.  I told them what I would do, and they said, by the way, we have an RFP out.  We agree, our site sucks, and we want to change it.  We sent the RFP to 32 organizations, companies.  Are you interested in receiving an RFP? 

At the time, at Atypon we were like five people, including my salesperson.  I had hired my friend, who was a PhD in computer science, I’d hired another PhD from NYU.  And I don’t remember some other PhD as well.  So it was four PhDs and another person, who also became a PhD in business eventually.  This is crazy. 

MU:  This goes against one of the fundamental rules of selling, that you don’t make sales by telling the organization what’s wrong with what they’re doing. 

GP:  I guess it doesn’t work, but it did in this case:  I guess they were eager to hear, and they liked the vision I was offering, especially about business models.  So we received the RFP, we respond, we get through the first stage, the second stage, and at the end we were against IBM.  And, of all the unlikeliest events, Atypon won.  Because we didn’t have the funding, we basically built the first platform on Academic Press’s dollar.  For which I will always be very thankful.

MU:  Did you have any idea then that this platform could be replicated across many companies?

GP:  No, I didn’t have an idea at the time that it would be applicable.  The one thing I did have, and I couldn’t do it when I was doing with Academic Press, was an idea of software as a service (SaaS).  Basically, when I created the transformation software, everybody wanted to license the software and pay me an annual fee for maintenance.  I was against it because I realized that we need more steady revenue.  So I switched:  I told them I would just sell the software by the page.  They were sending me documents;  I would do the transformations and charge them by the page.  I created essentially the SaaS model before there was one!  The SaaS model came in around 2001 I think, but I had it in 1995 or 1996. 

Actually, it only became SaaS when Blackwell Publishing came along.  Blackwell gave us the money to buy the hardware, because we didn’t have the money.  We bought the hardware with the agreement that we would be able to use it for other customers.  And that’s how we became a platform, in 2003, when we had our first other publisher.  Crossref had already come as well, and we had Annual Reviews, who was one of the publishers who was the first to come onto the Atypon platform. 

That’s when the cataclysmic event happened, at the end of 2003:  Academic Press was acquired by Elsevier, Annual Reviews came to Atypon, and HighWire basically ended our contract.  So I lost 70 % of my revenue at the beginning of that year.  I said to my employees, no problem.  I mean, we were always kind of very frugal, putting all the money in the bank.  I said, nobody gets fired.  I gave up my and my wife’s salaries completely, and I had to find another way of growing and living with my family, basically by investment.  We proceeded to grow the company and to grow it out of it, and by 2004 we were profitable again.  It was an incredible story.  It lasted until 2015, when Wiley came to ask us to acquire us, and eventually I agreed, at the end of 2016.

MU:  At the low point at 2003, how many staff did you have?

GP:  I had 12 — actually 13:  I know, because I called my father, who was a small business person, and I asked him how many employees he had at his peak, and he said 12, and I said, I have 13.  Actually, I never felt I surpassed him.  True, true stories.  I never felt I was able to surpass him, to do whatever he did.  He was one of my models of business people.

MU:  You found yourself running a company, but you didn’t go through the standard sort of management training or MBA or anything like that.  Did you feel you needed that kind of input?

GP:  No.  I attended one class in business school because I had a friend who got me into Harvard to attend the class.  No, I didn’t find it important, frankly.  I understand how business is taught in business schools, and maybe I’m full of it, but I think you can learn everything on your own.  Now, it doesn’t mean that I was a good business person.  I had to grow, and I had to read a few books.  I did find weaknesses in myself.  I realized that I was the one holding the company back.  I had hit all the points that some business people talk about, the points of growth that you hit:  that you, the manager, the leader has to grow, in order to grow the company.  I did hit all of those, but after the first one, which is 12 people by the way, because 12 people is the limit of what you can manage in a flat structure.  After that point, you go up and down, you know, and it’s actually very interesting.  After that point, I was aware, and I was able to grow, so the company would continue to grow.

MU:  At what point did you realize that it wasn’t your programming that was making the company grow?

GP:  Very early on — I mean, six months.  It was very interesting:  all these many years of computer science education, and I only programmed for two or three years in my life.  It was very soon that I realized I was a better programmer than anybody I was finding, but the company needed more than that, and I could just do the other functions, including sales, but also product management, product development essentially, and architecture was more important than me writing code myself.  So, I gave up on code very, very fast.  Which was something I loved before, but I had to do it for the good of the company and for the good of the people who are working there.  Once you have people working for you, you have responsibilities, you know, their wives and their mortgages are on top of you.  Their spouses are also women employees.  And their health incidents and their troubles, all of these things, you have to serve them.

MU:  Even though you started at a very fortunate time, the beginning of the internet, that still doesn’t explain your incredible success.  How do you explain Atypon’s astonishing rise?

GP:  Yeah, that’s a good question.  Because at every point, I was recognizing what was the problem and solving the main problem.  I realized that basically, there are three things a company offers:  efficiency, quality, and customer service.  Efficiency is like low cost essentially.  You can go with low cost, you can go with high quality, or you can go with great customer service.  You can have two out of three, but you cannot have all three.  So if you have high quality and high customer service, you cannot have a low price.  That’s kind of difficult, let’s put it this way.  It can be achieved eventually, but not right away:  you have to know what to focus on.  What I focused on in the beginning is quality of software and low price, how to make everything very efficient.  Not low price in terms of “I price my software lower.”  I had margins, good margins, but the question for me was where did you put your focus?  I put my focus on engineering, on more engineers.  I created an administrative tool so that I didn’t have to do the customer service.  Essentially, I outsourced the customer service back to my customers.  That was the efficiency I created;  it was also better for my customers, in the sense that they didn’t have to interact with the customer service person to do things on their behalf.  How would you feel if, instead of using Microsoft Word to write the document, you tell somebody what you want to write, and then he writes it, and then you check if he wrote it correctly, and then you send the corrections, and then you have to review it again and again and all  that loop.  I put my resources, engineering resources, on creating better administrative tools and better technology.  Although we were five people, when HighWire was 60 or 70, we were able to compete that way.  When we grew, I realized the next problem was all the customer requests coming in:  that’s the problem big platforms with many customers face:  they’re getting all these requests for changes, and all of them are urgent.  

Of course, everybody wants you to do something different from all the others (all of them are different, by the way).  And they all come, interestingly enough, in the last three months of the year because they have been allocated a budget, and they need to spend that budget in the last three months.  So, I created essentially a dedicated services model that solved this problem.  That was the problem that HighWire and other competitors of mine were hitting:  everything was becoming so complex that they couldn’t move forward.  They couldn’t improve the platform.  I realized that I was hitting the same ceiling, so I created dedicated services.  That took our revenues into years, and made life easier, more predictable, and more calm for the engineers and for the public.  Finding the problem that hinders your growth and figuring out how to solve it is the important thing. 

There were other things I tried do because I realized that all the work we had done was not good for science itself.  I tried to make amends, or basically change course, to create a better experience for the user.  Up to that point, we were focused on how to take care of the needs of the publishers, but we never thought so much about the user’s point of view.  How does the user want to use all these websites?  Essentially what we created was a wrong thing.  We created a balkanization of science:  journals in publisher websites, all different, all with a different set of features, with different interfaces, with a different set of demands on the user, registration, authentication everywhere.  If you want notifications from five different journals that belong to five different publishers, you have to register in all of them.  I tried to create a suite of services, but the publishers, of course, would not allow any organization to aggregate their content.  They were very much against it.  I tried to create what the first CEO of Academic Press called “Distributed Aggregation.”  Academic Press back in1998 was way ahead.  Peter Bolman, the CEO of Academic Press, created the term distributed aggregation.  He had already realized the problem of balkanization, and he said, we need distributed aggregation.  The user has a smoother experience as he navigates from website to website.

But of course, this diminishes the value of the individual websites, the value of what Atypon was building, right?  Because you have now all these services done at a level above the websites.  I was okay with that, but Wiley, not so much, so that vision never materialized.  We did create the tools, but they didn’t have the internal support;  they didn’t succeed.  Generally, these tools, the distributed aggregation, with a few exceptions, has not succeeded.  The exception is Google Scholar, for example.  If you do a search, you search in Google Scholar.  But there are few others;  maybe Altmetrics is another one.  But there are very few examples that have succeeded, and they don’t offer essentially enough value for the users.  They don’t offer enough value on top of the website.  If I could do things now, I would do them differently.  The models of these exist in the world:  look at music publishing, for example, like Spotify.  There are many platforms for independent movies, especially in Europe.  The situation in the movies here is not so clear, because it’s going towards balkanization.  Disney doesn’t license its content to other platforms, and Warner will probably start doing the same.  They are moving from licensing their content to these big platforms competing with each other with different business models, with different innovation and revenue, and different business models.  Their move to balkanization is not helpful for the users.

But most of all, it’s not helpful for scholarly users.  It’s one thing to say, if I need to see a movie, I have to find the platform.  I have to be sure I’m a subscriber to that platform to see that particular movie.  It boggles the mind, but it’s okay.  But in science, everything is interconnected through the references.  You need to go to this other website and suddenly you’re in a totally different environment.  The way I would do it is to create competing online libraries, and the publishers would license the content to these libraries.  Right now is a good time for making this change in the publishing world.  There is no place, in my view, for companies like Atypon in the future.

MU:  You mentioned Google Scholar, but of course there are two other areas where this kind of aggregation has taken place.  And both of them with question marks over them.  One is the of LLMs that are pulling together all sorts of content.  The other is pirate sites like SciHub, doing exactly what you describe.

GP:  Yes, and it tells you that there is a need.  It’s a proof of the need that exists.  And preprints is another one, if you think about it.  So there is a need out there.  The business model right now is reverting.  I think the reason the balkanization happened is because the publishers were the owners of the content, and reselling of the content was very important.  Now there is, you know, publishers have become essentially service organizations to researchers and research organizations.  Essentially, they are selling both author fees, selling services editorial and the brand of the journal, which is very important, and will continue being important.  And other services of course, such as typesetting, all of these things, the menial things that they do, are also important. 

All these things that a publisher does are covered by author fees.  Once they have the author fees, there is no incentive for the publisher to spend on putting things on a website.  They might as well give it away for free for other websites to publish it and try to figure out how to make money out of it.  The agreement would be, you have to give it away.  Okay, it has to be presented, full text and then you figure out what kind of value you want to add, and then we’re out of this game.  They’re spending good money on websites, not particularly huge money, but they could save that money.

MU:  Do you think the trend towards open science, and away from publisher control of articles, is a good thing for scholarly publishing?

GP:  Yes.  I think open science is a good thing:   a lot of the things they’re talking about are absolutely a good thing.  There is one thing I don’t agree with or I’m worried about, let’s put it this way:  how it is assumed it will be financed.  Who does the financing is, to me, the important thing.  The implication in many of these presentations is that the government pays for it.  It’s easy to fall into this trap because the government pays for research, therefore the government should pay for the publication of the research.  But the government should not be running the publication of the business.  Whoever is doing it shouldn’t be dependent on government revenue;  it should be competing private organizations, because competition is the only way to ensure innovation.  Once it goes into government, it withers and dies.  Sorry, that’s what I’ve seen in all the examples.  Otherwise, the Soviet Union would have been a superpower.  You need private competition.  So you need publishers, in my view.  Publishers playing an important role, and they should be private, profit-seeking organizations.  I believe in profit-seeking as a model for, and competition as a model for innovation.

MU:  Perhaps the closest equivalent that you described to this sort of aggregation model is Crossref, which is maybe not commercial, but certainly sustainable.

GP:  Yes, it can be owned by publishers.  If, for example, there are no private organizations that are willing to pick it up as it is, it can be funded by publishers.  They can fund two or three organizations, each doing different things and, let the best do well, and let there be an alternative.  I don’t want government involved.  Once government is involved, can be no private competition, because the government can copy you, can put any money into it.  It’s not fair competition at that point.  There needs to be a level playing field for competition.

MU:  We can’t have an interview without talking about GenAI.  Are you using GenAI?  Are you using LLMs?

GP:  Is there anybody not using them these days?  I don’t know.  I have a subscription to two of them, ChatGPT and Grok.  I use Grok a lot more than ChatGPT;  I find Grok better.  I use it heavily, on a daily basis.  I was trying to write, for example, an investment article, and I got all kinds of information from Grok.  I learn about investment analysis, publishing trends and everything new.  Right now, I am trying to build a house, and I’m learning all kinds of things from Grok.  I have certain hobbies, like cars, for example — I don’t know if you know about restomods (restored classic cars), and I’ve designed using Grok — it has an imaging component.  I tell it how to design it, and I see it before I build it.  Whenever I’m bored, I just go to Grok, and learn about something new.

MU:  It’s been remarkable talking to you:  your entire career has been about innovation, and you haven’t finished innovating now.  I look forward to some more great innovations in the years to come. 

GP:  Thank you very much, Michael.  Honored.