Oh, IBM from the tabulator rental era, with the Census Bureau as the biggest customer.
Here's the shock said to have put IBM into the computer business.
I once had a summer job at the Census Bureau, and got this story from someone who had a "Univac I - I was there" certificate on their wall. Census had been IBM's biggest customer all the way back to Herman Hollerith in 1890. They had acres of tabulating machines in Suitland, Maryland. All were rented by the month, because IBM refused to sell them.
Census was the first customer for the UNIVAC I computer.[1] The UNIVAC I had a long row of magnetic tape drives. At long last, they could move from warehouses of punched cards to magnetic tape. Sort without manual card handling. Converting from cards to tape took several years. It wasn't clear it was going to work. But it did.
Then the Census Bureau terminated the monthly rentals on several acres of IBM punched card equipment.
That was a huge shock to IBM. Their biggest customer for 70 years was done with them. Until then, IBM's electronic computer efforts were mostly for the military and the Atomic Energy Commission, and the machines were not cost-effective for commercial data processing. T.J. Watson Sr. had said he saw a market for about five electronic computers. IBM was thinking of computers as being for scientific computation, and not cost-effective for business data processing. But Census was ordinary batch data processing, just on a large scale. The same kind of thing their other large tabulating machine customers did. Such as U.S. Steel, MetLife, DuPont, and General Electric, all of which soon bought UNIVAC I machines.
That's what kicked IBM into making business-type computers, with tape drives and more peripherals.
With a bit of looking, you can find A Business and Its Beliefs by Thomas J. Watson, Jr., published as part of the McKinsey lecture series. He does cite the UNIVAC as the shock that made IBM understand what it had to do.
Unfortunately, the lectures were delivered before the S/360 was announced. It would have been interesting to read what he had to say about that.
Yes. Federal Office Building #3, Suitland, Maryland, sounds like an address from fiction. But it was quite real.[1][2] Built in 1941, looked like a reinforced concrete parking garage inside. Demolished some time around 2000. The computers were in the leftmost wing, as seen from the front.
I wasn't alive during the period of IBM dominance, but I have a certain romantic view of it. The rapid advancement of technology by leaps and bounds, the ability of a single person to write operating systems and compilers at the seat of a terminal, even the branding and design is something I associate with the heart of computing. Ok, maybe I'm more in love with Bell Labs than IBM, but there's a part of me that wishes I could have been there and experienced the age of the terminal and the beginnings of the PC.
My generation will probably never experience the same magic. We've reached the end of Moore's law and each year brings uninspired iterations on the major successes of the past. We're born with smart phones in our hands, it's something that we've never not known, and that destroys our ability to appreciate its value.
By the way, if anyone wants to recommend some good late 1970s/1980s fonts, keyboard reproductions, wallpapers, etc. I'm really into collecting those things. I use IBM Plex Mono quite a bit already.
When I first learned about OS internals I felt the save way. All the cool stuff has been done. All the good ideas were discovered many years ago. If only I was born 20 years earlier...
But actually, there has been a lot of progress since then. There were plenty of cool ideas and innovations being made. It's true that it may not all be ground breaking, but the actual rate of innovation is increasing, not decreasing...
Just think about what your phone can do. It's amazing.
Yeah there's still a lot happening. On the other hand, the modern smartphone is really only a bit over 10 years old in terms of mainstream impact and there probably hasn't been anything else in the tech space that's so directly had a major impact on consumers since. (Probably the Web before that and the PC before that--to a lesser degree.)
Quantum computers and quantum AI are the next architecture/paradigm revolution with that feel.
We know there’s some quantum accelerated AI work, eg by Microsoft — and we’ve seen increased quantum and photonic circuits. China has demonstrated light table and distributed quantum computing, so we know it “scales”.
But we’re not quite sure which of many approaches will get it from bench top to “working” — much less who will be dominant in a decade.
Quantum accelerated AI is that revolution in our time.
> IBM's sales centric approach naturally assumed a slower industry cadence than the one that developed in the 80s. Rapid growth in low-end computers reduced the relevance of their competitive advantage in large-scale, long-term deals, …
I remember one small incident working for IBM as a very junior employee in 1984. A customer saw the PC and Lotus 1-2-3 in a demo and wanted to buy it there and then - literally got their cheque book out. The experienced IBM salesman stepped in and tried to convince them to buy a System/36. Result no sale.
I wonder to what extent IBM never really came to terms with the success of the PC.
Well, they tried to lock things down again to a certain degree with the PS/2 and OS/2.
Of course, pretty much none of the older generation of system companies really got the PC and PC compatibility. (HP probably came closest. Being mostly compatible was still such a dramatic shift from the situation in the minicomputer space that they largely felt that it was good enough.
Indeed. The PS/2 was a symptom of their unease with the PC. Had they really pushed their own true PC compatibles much harder they would have done much better - Compaq had the first 386 I think which is extraordinary.
Just not sure that IBM could adjust itself quickly enough. The PC was after all created in a skunkworks.
"Nobody predicted what would happen next. The first shipments began in October 1981, and in its first year, the IBM PC generated $1 billion in revenue, far exceeding company projections. IBM's original manufacturing forecasts called for 1 million machines over three years, with 200,000 the first year. In reality, customers were buying 200,000 PCs per month by the second year." [0]
I suspect those sales were made in spite of much of the IBM hierarchy.
Another example IBM in the UK where I was working had a headcount limit imposed by Armonk. No matter what their revenues they couldn’t take any more people on. Did this limit their ability to sell lower value systems - seems a distinct possibility.
One of the rationales behind the PS/2 line (and OS/2) is that IBM senior management was trying to put the genie back in the bottle by designing hardware and software to work as thin clients for IBM mainframes. At the time they still thought they had tremendous leverage over the PC market, and were pretty clueless about what customers actually wanted, which cluelessness was relentlessly exploited by the clone makers.
I worked at IBM for a year out of college before getting laid off as apart of a round of mass layoffs that the company regularly does. Joining the company definitely felt cool, especially as a new grad. The history alone was pretty neat and I was super happy to be a part of something that had had such a big impact on the world. After spending some time at the company though, the biggest thing that struck me was the quality of the people and the mindset of management. IBM pays poorly, raises are nonexistent and so are bonuses. My coworkers were not particularly smart nor passionate. I honestly believe that a major factor in the decline of IBM is the refusal to invest in their people. Most engineers and designers are just cogs for the many layers of management to shuffle around and use, and the managers are more interested in playing the office politics game and micromanaging than actually building cool stuff. It's a very top-down sort of place but the people at the top and the bottom are not very smart.
Most tech services companies optimize for the median: they don't want low performers (low quality work), but they also don't want high performers (v expensive). They're built to provide good-but-not-great services.
I'm curious when it was you were there. I was in IBM from 1985-1992, and it felt like the ending of an era. At first, it was a place that invested heavily in its people. Private offices, nice facilities, very good pay, and a _requirement_ that you had to take classes related to your field every year. IBM would hire excellent faculty and instructors from universities and other companies to teach them. They sponsored technical conferences that were really good (actual learning, not just sales). When I was in graduate school, my director had me focus on doing my homework and projects as 80% of my time -- literally getting paid to be a student. My colleagues and I were proud to be "IBMers" at that time.
But things started to change in 1990-91. The company lost money for the first time in its history, and then divided into separate parts. My division became what is now the horrible IBM Global Services. In those last 2 years it became clear to me that it was time to branch out of there. Since then, I've know no IBM employee at all who is happy, much less proud.
All these great old institutions ultimately fade or fail. It used to seem more sad to me, but I've grown to favor "fail fast" more. IBM should just go...sell all the IP and branding off and call it a day. Then, new innovation will happen faster.
> The company lost money for the first time in its history
Seems to be a recurring theme. In my city there's a large steel industry. The companies used to be quite benevolent, they cared about employees and invested in them, and everyone was proud to work for them. It all ended the second they took losses and it never came back. Now they're just normal corporations, the employee pride is gone.
I would be sad to see it go, if only for the fact that they are (afaik) the only ones keeping mainframes alive as a modern thing.
Idle speculation, but I wonder if that could save IBM, if and when the market starts to sour on cloud computing (for cost and/or control reasons). Mainframes have a lot of overlap in terms of functionality (VMs, reliability, similar purchasing model for computing power) with cloud services.
I worked there for a number of years and I can say with certainty that what you're saying exists within the company. However, this is, like any sufficiently sized company, down to the culture of the given group you're working for.
One thing that IBM is very good at IME is pushing out employees from acquisitions that represent the "old guard" as it were.
The main thing it has going for it these days is inertia with banks and governments and the research arm.
The best people I ever met while working at IBM were the ones who got there via acquisitions. Also there were grumblings about how employees hired directly by IBM appeared to be promoted more often than acquisition employees. Something about them being "true blue", which if real, tells you everything you need to know about the upper management and their opinions of subordinates there.
From an outside of IBM point of view IBM limited much of what smart engineers could do in life, you couldn't write OSs, design computer hardware, etc etc those were all things that (because of the way things were structured) were IBM's pervue, it's why CDC/etc anti-trust suits occurred ....
Arrival of the microprocessor, Unix escaping AT&T and some other changes in that era changed how all that worked, now we could write our own OSs, linux came about, a host of new little semi startups started making RISCs - and Intel/AMD (and I guess Microsoft, though OS has muted that) sort of took over ownership of owning the big gatekeeper ....
Nowdays we see RISCV coming along as a disrupter and a 1000 new little flowers blooming
> Enterprise: buying IBM meant (a) job security and (b) high cost. AWS is that today.
My experience with Enterprise is the complete opposite, especially Enterprise that makes its money anything directly or peripherally related to retail.
Around 1990, I remember seeing a book in the cut-out bin of a bookstore about IBM and how it was an inevitable juggernaut that would consume the world. Wish I'd bought it :-)
All through the 1980s, I heard the same thing - IBM was an unstoppable monopoly. It's the exact same thing today, just the company is a different one. It changes every 10 or 20 years.
But the fall of the previous "unstoppable juggernauts" never deters pundits from anointing the current top dog as an unstoppable juggernaut.
I actually think that facebook is better at surviving then for example google. They always were in a anxiety spot, chasing the crowd which tried to shake the adds of in a rodeo.
Google is not really prepared for a world were they don't have a platform for search.
Regardless of my prediction, all of them have spawned Eco-systems holding offspring with their cultural"DNA", so we will never get rid of them entirely.
Additionally, IBM made several good pivots in the 90s and later. IBM embraced the world wide web and supported many web products. IBM did well for a while with RISC processors, AI, and cloud computing. However, I think the expenses out-paced the revenues.
Many years ago (pre-Fbook/Google) I recall thinking “Most of the smartest technic people I know are going to Apple” and “The ones at IBM are business majors latching onto standards.”
After Gerstner saved them from the dead, it took much longer than I expected for IBM to flounder. Perhaps this is because they were smart enough to not invest (waste money) in new technology that they were incapable of building. Yes, for them it was smart. They created much more shareholder value buying back stock and auditing their existing customer base as they wind the company down. It’s awful to be a customer of theirs when your predecessors let them in the door, but that’s how they keep the machine running.
As an intern I worked in an IBM hardware lab for consumer PCs (Aptiva) and electronics. The lab had all sorts of goodies including IBM's recently discontinued X10 line of home automation products, Home Director. We had a brand new iMac OG (G3). We hit it with bones and sticks but didn't quite know what to do with it.
IBM salesmen were very iffy at the best. But the two best computers I used in the 1980s to 1990s were the 4381 super-mini, and the various IBM PC models. Great hardware engineering, and in the case of the 4381 great software engineering.
I was part of Techstars some years ago and had some chats lined up cloud architects from Azure, AWS and IBM. Azure and AWS guys were super nice, we talked architecture, cloud benefits, possible use cases, offering free credits etc.
Then the IBM discussion came up and within' 2 minutes, I was already getting pitched to buy stuff vaguely related to what we were doing. Also some offering of stuff that made no sense whatsoever for us, and yearly packages, and free credits if I do this or that or pre-sign for this; maybe the worst discussion I had in 3 months over there.
And all I got was a knobbly biro in the form of a Charlie Chaplin walking stick. Part of the PC Junior launch, I think.
Oh and (full disclosure) a meal in the canteen at the IBM Greenford site, and one at their South Bank site. No alcohol. Hey, if you want to sell things to programmers, provide alcohol (yes, I know it was IBM policy not to do so).
I interviewed at IBM in 1997 or 98 for a job developing their mainframe processors. They made me an offer, but I told them I had a better offer from SGI, and asked if they could match it. They were so offended by this that they never replied to me. I never heard another word from them. Very strange people.
an interesting beer discussion ia what would have happened to Linux if IBM hadnt very publically announced their backing in the early 2000s. In doing do it made enterprises start to consider that Linux might be a suitably safe choice to depend on, as opposed to Solaris or HPUX or even AIX. Useful support and backing at a time when Microsoft was profoundly anti Linux.
39 comments
[ 2.1 ms ] story [ 90.9 ms ] threadHere's the shock said to have put IBM into the computer business.
I once had a summer job at the Census Bureau, and got this story from someone who had a "Univac I - I was there" certificate on their wall. Census had been IBM's biggest customer all the way back to Herman Hollerith in 1890. They had acres of tabulating machines in Suitland, Maryland. All were rented by the month, because IBM refused to sell them.
Census was the first customer for the UNIVAC I computer.[1] The UNIVAC I had a long row of magnetic tape drives. At long last, they could move from warehouses of punched cards to magnetic tape. Sort without manual card handling. Converting from cards to tape took several years. It wasn't clear it was going to work. But it did.
Then the Census Bureau terminated the monthly rentals on several acres of IBM punched card equipment.
That was a huge shock to IBM. Their biggest customer for 70 years was done with them. Until then, IBM's electronic computer efforts were mostly for the military and the Atomic Energy Commission, and the machines were not cost-effective for commercial data processing. T.J. Watson Sr. had said he saw a market for about five electronic computers. IBM was thinking of computers as being for scientific computation, and not cost-effective for business data processing. But Census was ordinary batch data processing, just on a large scale. The same kind of thing their other large tabulating machine customers did. Such as U.S. Steel, MetLife, DuPont, and General Electric, all of which soon bought UNIVAC I machines.
That's what kicked IBM into making business-type computers, with tape drives and more peripherals.
[1] https://www.census.gov/history/www/innovations/technology/un...
Unfortunately, the lectures were delivered before the S/360 was announced. It would have been interesting to read what he had to say about that.
[1] https://www.census.gov/history/www/census_then_now/suitland_...
[2] https://www.census.gov/history/www/innovations/technology/un...
But actually, there has been a lot of progress since then. There were plenty of cool ideas and innovations being made. It's true that it may not all be ground breaking, but the actual rate of innovation is increasing, not decreasing... Just think about what your phone can do. It's amazing.
You might enjoy browsing some of the old Byte magazines on the internet archive which give quite a good feeling for the era.
We know there’s some quantum accelerated AI work, eg by Microsoft — and we’ve seen increased quantum and photonic circuits. China has demonstrated light table and distributed quantum computing, so we know it “scales”.
But we’re not quite sure which of many approaches will get it from bench top to “working” — much less who will be dominant in a decade.
Quantum accelerated AI is that revolution in our time.
I remember one small incident working for IBM as a very junior employee in 1984. A customer saw the PC and Lotus 1-2-3 in a demo and wanted to buy it there and then - literally got their cheque book out. The experienced IBM salesman stepped in and tried to convince them to buy a System/36. Result no sale.
I wonder to what extent IBM never really came to terms with the success of the PC.
Of course, pretty much none of the older generation of system companies really got the PC and PC compatibility. (HP probably came closest. Being mostly compatible was still such a dramatic shift from the situation in the minicomputer space that they largely felt that it was good enough.
Just not sure that IBM could adjust itself quickly enough. The PC was after all created in a skunkworks.
[0] https://spectrum.ieee.org/how-the-ibm-pc-won-then-lost-the-p...
Another example IBM in the UK where I was working had a headcount limit imposed by Armonk. No matter what their revenues they couldn’t take any more people on. Did this limit their ability to sell lower value systems - seems a distinct possibility.
Sounds like you noticed this pretty quickly.
But things started to change in 1990-91. The company lost money for the first time in its history, and then divided into separate parts. My division became what is now the horrible IBM Global Services. In those last 2 years it became clear to me that it was time to branch out of there. Since then, I've know no IBM employee at all who is happy, much less proud.
All these great old institutions ultimately fade or fail. It used to seem more sad to me, but I've grown to favor "fail fast" more. IBM should just go...sell all the IP and branding off and call it a day. Then, new innovation will happen faster.
> The company lost money for the first time in its history
Seems to be a recurring theme. In my city there's a large steel industry. The companies used to be quite benevolent, they cared about employees and invested in them, and everyone was proud to work for them. It all ended the second they took losses and it never came back. Now they're just normal corporations, the employee pride is gone.
Idle speculation, but I wonder if that could save IBM, if and when the market starts to sour on cloud computing (for cost and/or control reasons). Mainframes have a lot of overlap in terms of functionality (VMs, reliability, similar purchasing model for computing power) with cloud services.
One thing that IBM is very good at IME is pushing out employees from acquisitions that represent the "old guard" as it were.
The main thing it has going for it these days is inertia with banks and governments and the research arm.
Arrival of the microprocessor, Unix escaping AT&T and some other changes in that era changed how all that worked, now we could write our own OSs, linux came about, a host of new little semi startups started making RISCs - and Intel/AMD (and I guess Microsoft, though OS has muted that) sort of took over ownership of owning the big gatekeeper ....
Nowdays we see RISCV coming along as a disrupter and a 1000 new little flowers blooming
Consumer: the Sears catalog was known to be the most expansive catalog-of-items to buy from. Amazon.com is that today.
TL;DR - Amazon has won both the enterprise and consumer market.
My experience with Enterprise is the complete opposite, especially Enterprise that makes its money anything directly or peripherally related to retail.
All through the 1980s, I heard the same thing - IBM was an unstoppable monopoly. It's the exact same thing today, just the company is a different one. It changes every 10 or 20 years.
But the fall of the previous "unstoppable juggernauts" never deters pundits from anointing the current top dog as an unstoppable juggernaut.
Google is not really prepared for a world were they don't have a platform for search.
Regardless of my prediction, all of them have spawned Eco-systems holding offspring with their cultural"DNA", so we will never get rid of them entirely.
After Gerstner saved them from the dead, it took much longer than I expected for IBM to flounder. Perhaps this is because they were smart enough to not invest (waste money) in new technology that they were incapable of building. Yes, for them it was smart. They created much more shareholder value buying back stock and auditing their existing customer base as they wind the company down. It’s awful to be a customer of theirs when your predecessors let them in the door, but that’s how they keep the machine running.
Then the IBM discussion came up and within' 2 minutes, I was already getting pitched to buy stuff vaguely related to what we were doing. Also some offering of stuff that made no sense whatsoever for us, and yearly packages, and free credits if I do this or that or pre-sign for this; maybe the worst discussion I had in 3 months over there.
Oh and (full disclosure) a meal in the canteen at the IBM Greenford site, and one at their South Bank site. No alcohol. Hey, if you want to sell things to programmers, provide alcohol (yes, I know it was IBM policy not to do so).
This was in London, in the mid/late 80s.