Machines will take over the world, but not without our help
After two years at IBM, I’ve observed our major institutions failing workers, especially the already vulnerable
Among software engineers, working for MAANG (Meta, Apple, Amazon, Netflix and Google) means you’ve “made it,” not in the least because of the income prospects. Joining an engineering-first organization promises agency and autonomy that is not easily matched. Lower on the ladder are organizations seeking to modernize but without a history or commitment to good software, and lowest of all are consulting or contracting agencies like IBM. There are simply two kinds of “developers” in the labor market: The engineer in the best sense, and the code monkey lent out for a pretty penny.
Driven by the impulse to modernize, the banks and insurance companies that fall into the middle group have been rewriting their software. As a technical consultant at one of the largest banks in the U.S., I was a traffic cop for some of that chaos, often digging through codebases and slapping together documentation. What we aimed to accomplish eluded me. Were we trying to make our data movement faster, cheaper, more secure? Were we offering customers a better experience? Perhaps because I was a contractor from IBM I was never clued in. The most compelling visions seemed to be driven from the ground up by strong individuals circumventing cumbersome obstacles. The product my team stewarded excised the need to code altogether to escape the bureaucracy of software deployment at this bank: trails and trails of paperwork, and monthly releases that had to be locked in two months prior.
I started at IBM in the inconspicuous role of managing processes to bring on new talent to the same client I mention above, whether IBM full time hires (in the precise sense, contractors) or “subcontractors,” recently landed immigrants who contracted through multiple middlemen to find work at IBM. Each person had a gross profit margin: the difference between revenue and cost, divided by revenue. We were all grubbing for the cheapest way to do the work, whether it was the client insisting on rock-bottom hourly rates, or IBM refusing inflation-mitigating raises for subcontractors to preserve a sizable gross profit margin. Once on the ‘inside,’ no longer staffing new talent but embedded on a team doing the abovementioned work, I counted my hours against a ‘project id’ in a staffing system meant to visualize how dollars piled up against each initiative – for both IBM and the client. In light of all this, it was hard to believe we did little more than pinch pennies trying to appear in pursuit of a lofty but ultimately amorphous objective: rebuild… things… slightly… better?
My official role on the client team puzzled me. Before IBM, I had never heard of the title “business systems analyst,” though it was commonspeak to IBM managers, most of the financial institutions IBM contracted with, and the staffing agencies that contracted with IBM. At first, I did not mind the somewhat odd demands on BSAs: We wrote technical requirements that shaded into the work developers did. I could read Kafka mapping documents in my sleep. I was spelling out if-then statements and designing the flow of data, such as when to filter out on irrelevant messages. The developers, meanwhile, took the requirements and translated it into code. They defended against the unusual data input, the corner cases, and knew how to say it in Java.
But why are there the two of you? My husband, a software architect in-house at a software company, asked me one day. If you’re deciding the logical design, shouldn’t you be paid like a software engineer? Why are they reducing their software engineers to ‘code monkeys’? The meat of engineering is in the engineering, he exclaimed, not in translating what was already designed into the same code components again and again.
In my remaining year at IBM, I observed that BSAs and developers played different roles depending on the team. A developer griped with me when I didn’t provide the data value used to name his HTML file, while another developer knew the business requirements better than his BSA did. Some BSAs had technical acumen while others were glorified project managers. Still worse, some BSAs never unmuted in our calls, were asked to investigate technical matters but came up empty-handed during the next week’s check in, and then they were shuffled away to other teams… or maybe other companies. When asked what exactly I did for work, I took to explaining that my role “filled the cracks” between what software engineers did. I was caulk.
The fact that such a role existed meant that someone, somewhere deemed such ‘caulking’ worth it enough to hire the extra head, and the cost that came with it. This seemed at odds with the client’s hardpressed stance. But my role was, in fact, born of cost-saving measures. Contracting out or offshoring a job traditionally done by expensive, U.S. educated talent, companies hoped to pay less per head. But the talent struggled with the engineering demands, whether because cultural gaps led to communication issues, or schooling never emphasized that engineering involved more than writing code, or because, having never worked ‘in-house’ or on a gig that lasted longer than six months, people never benefited from proper mentorship and training. They threw more cheap people at the work. That was me, the BSA. Two people came to shoulder the job of one person, unevenly and with great inefficiency. The numbers still told a hopeful story, however, as it was more people for less buck.
This offshoring and outsourcing, dubbed “fissuring” by economist and professor David Weil, entailed the calculated movement of jobs outside the legal and social jurisdiction of a company, which absolved said company of any responsibility for the people who filled those jobs. By contracting with IBM, my client unburdened itself of needing to provide healthcare and retirement benefits, mentorship or work satisfaction, and managerial support to IBM consultants; IBM, by contracting with staffing agencies, avoided all of the complications that came with visas and work sponsorships, not to mention the need to pay a competitive wage.
People were expensive. This seemed like a fair compromise. Yes, “Rockstar” developer Eddie would continue to earn a poverty-line wage for the next four years, hired through a client innovation center in one of the highest crime cities in the world, but he worked at a desk, logged off around six, and built America’s newest software infrastructure. Yes, Senior Developer Jay’s neighborhood was prone to petty theft, but when his house was broken into and his laptop stolen, express-shipping him another laptop was a minute cost and he was back at work within the week. Yes, David received a greatly reduced paycheck after all the contracting agencies skimmed off the top, and he had no negotiation power being an immigrant tied by sponsorship to a third-party employer, but he had a chance to live in Canada instead of China. Yes, Bommu would never receive the mentorship that would graduate him from a code monkey to a skilled architect, but we paid him better than a job back in the motherland.
I read about fissuring in the recent book of a labor and employment law professor, Cynthia Estlund, who introduced it as the precondition for automation. You can only replace employees if their work is machine-replicable in the first place. When I mentioned offhandedly the technical way in which I wrote my requirements, she exclaimed “that is exactly the kind of language easily understood by machines.” When companies adopt tools like Chat GPT, they will displace the developers who are already no more than ‘code writing machines.’ The reverberations will be felt not only by our developers at home, but by countries like China and India that rely on the information economy.
Stripping the human out of the job, companies eliminate the top-heavy managerial chain that sits above him, the need to pay for his unproductive days, his sick days, his holidays, his medical bills, his retirement dues, his socialization on the company’s dime, to heed his aches and pains, complaints and depression, dreams and aspirations, need for appreciation, desire to be fulfilled. This has thrown employment law, which has regulated fissuring by raising bottom-line mandates, for a loop. The more we raise the bar on what employers need to provide for workers, the more likely they will be replaced with robots.
In the early 1900s, computers rang up census data using hole-punched cards at a fraction of the speed it took humans to record and compile that information– I am describing none other than IBM’s first venture, though the company was then called “Computing-Tabulating-Recording Company.” Only fifty years had passed since the first Industrial Revolution rendered man’s physical labor obsolete. Anticipating that a second would trivialize human thinking, J.P. Guilford, President of the American Psychological Association said: “It would be necessary to develop an economic order in which sufficient employment and wage earning would still be available. This would require creative thinking of an unusual order and speed.”
Such a statement belied not only the economic reordering already taking place, but who led the charge. The lineage of creative thinking can be traced to mechanical and electrical engineers presenting and publishing about it in the 1940’s and 50’s. Guilford read all those papers. They sought, at that time, to distinguish engineering from science. While scientists holed away hashing out theories with no direct significance to the world, engineers applied those theories to make the world go round. America was built on the backs of hardworking mechanics, said President of the American Society of Mechanical Engineering Robert H. Thurston, not via the abstract intellectualism that characterized the Europeans.
Factoring in business interests was therefore a virtue, not an adulteration, of the discipline. The truly creative engineer lifted his eyes from the design and assembly of a particular product to the “total economic, industrial, and mechanical system” of which his product was a part. This broader perspective, spoken by the President of Teletype Corporation T. E. Shea, resounded with the other presenters at the ASME conference, most of whom owned their own enterprises. Eyes on the broader industrial complex, they successfully scaled their companies by specializing their factories and speeding up production. Turnstiles, conveyor belts and rotating devices piped machine parts from one unskilled worker to another, each of whom repeated a standardized procedure, replacing the work of a skilled craftsman. People became no more than well-oiled parts of a machine-like production facility.
A look at history exposes the human agency that led us to where we are now, but as always, that history is complicated. These businessmen laid the foundation of our economy. Making order out of chaos, they released and enforced code on ubiquitous manufacturing parts like the screwthread, socket, or meter. Their gains in productivity allowed them to pay their workers more and charge their customers less. Yet the quality of most jobs fell significantly. The profitability of the enterprise determined what the human did in the workplace— not whether he was fulfilled, participated in cultural heritage through the learning of trades, or had meaningful interactions with other people. They orchestrated a widespread deskilling that underlies the fissuring and labor conditions we see today.
It is self-evident that the human was left behind in this quest for profit. We should be people-wise and business smart– pursuing these in tandem. Running a business virtuously is no matter of theory, however. When I think about whether it can be better, I get snagged on the question, Who? Is there anyone who can restructure the IBMs of the world, who can steer the billions of dollars and tens of thousands of people toward good? Does the momentum of a company barrel past anyone’s attempt at reform? Is the ship already sinking as we rearrange the deck chairs?
I am no optimistic humanist… but I have a foolish hope for better things. Has someone preserved, all along, the pockets of goodness and joy that we desperately seek out? Who holds together the systems of the world whose parts don’t quite fit together right? Who is able to restructure and rectify them? I call that someone God.