Keeping the (idiosyncratic) human in the loop.

I spend a lot of my time at Mozilla these days building prototypes that incorporate some aspect of artificial intelligence and, subsequently, inescapably, considering the impact of such developments on humans. My thoughts often wander, unsurprisingly, to my own future as a developer and reluctant tech user.

There has been plenty written about how AI will replace or shift jobs, and my own career as a mid-level designer / dev hybrid is certainly far from safe. I could expound on this more, but James Somers’ New Yorker Article “A Coder Considers the Waning Days of the Craft” captures it with impressive eloquence. In particular this sentiment resonates:

“Maybe the next cohort [of Developers] will spend their late nights in the guts of the A.I.s their parents once regarded as black boxes. I shouldn’t worry that the era of coding is winding down. Hacking is forever.” - James Somers

Now, more than ever, is the time to deeply examine what we want our relationship to technology to be. Somers notes that his collaborator, less the skilled programmer, but all-in on AI, began to align his own neural network with GPT-4’s. “I would have said that he had achieved mechanical sympathy.” And while such sympathy (empathy?) with machines is perhaps admirable, there are plenty of us for whom the superiority of AI in certain tasks registers as defeat. Somers highlights legendary Go player Lee Sedol’s defeat at the hands of AlphaGo. “Perhaps what pushed Lee Sedol to retire from the game of Go was the sense that the game had been forever cheapened,” he writes.

This strikes a chord - I like hacking - even in my acceptance that in many ways my skill might be inferior to the endless resources of AI coding assistants like Github Copilot. My friends, the out of work copywriter and the successful but financial constrained digital artist, love the daily, active, humanistic participation in their craft, however non-performant or imperfect.

Now, more than ever, is the time to deeply examine what we want our relationship to technology to be. Maybe the thing to teach isn’t a skill but a spirit.

Relucant acceptance is not our only option. Humanity’s role in digital systems (and natural ones as well) is, in my opinion, the thing we should be dropping everything we’re doing to have a good answer to. The constant push of capitalism, with profit and productivity as its ultimate aim, allows little time or incentive for this sort of deliberation. What can we bring to the table?

Somers argues, “Computing is not yet overcome. GPT-4 is impressive, but a layperson can’t wield it the way a programmer can.” This might provide some relief for the nervous mid to senior dev, but what about the rest of us “laypeople”, unordained in the church of GPT? I took much greater comfort in the sentiments of Glenn McDonald, a self described and recently laid-off “Data Alchemist” at Spotify:

“The most transformational ideas begin in individual human hearts.❤️”

He dissects this further in a blog post:

“It’s easy to find advice for how to modulate your attachment to work but good luck finding books that advise a corporation on how to encourage reliance on unsupervised inspiration. It needs to be able to treat employees as interchangeable, to periodically layoff 17% of the workforce to cut overhead by 1% and boost stock price by 5%. The key is to turn skills into roles symbolic & anonymous within the system. When people are units in slots, the machinery doesn’t need to care who they are. If you bring a new idea to a well-organized machine, it will efficiently route you into a pipeline.”

This is vastly different than the instinctual approach of a human, however. McDonald: “If you came to me with a new idea or an unanswered problem, I would listen, ask questions, pivot…I could do new things because my inquiries weren’t prescribed. No advice on corporate agility advises finding people with random skills who do unique work on the fly out of passion & stubbornness. You can’t sell a methodology where a crucial step is ‘Luck into anomalous contributors’. Anomalies are what processes try to preclude. But great companies exists because a few people once had unruly unsupervised impulses that the status quo couldn’t accommodate. Companies are better to ignore caution with the exuberant disregard of those doing their jobs with inadvisable devotion.”

Reflecting on the reasons he was fired in his blog, McDonald writes, “I don’t really think I was laid off with grand purpose. My guess is that my name got picked by somebody who didn’t understand the scope of my work, as opposed to somebody who did and objected to it.”

McDonald’s niche at Spotify seemed elusive but crucial. As a recommendation on his LinkedIn page reads, “Glenn is one of the most creative people I’ve ever worked with. His contribution to any team for any project can not be over-estimated. I’m not sure what the hell it is he does, but he does it incredibly well.”

This encapsulates perfectly what so many of us bring to work each day, with capacities, methodologies, personalities, and experience that can’t be programmatically extracted into an annotated data set or squeezed into the context window of a large language model. Glenn was an anomaly. Perhaps all of us are anomalies, and our unpredictability is precisely what makes us unique and essential to the system.

Our first computer, The Turing Machine, whose architecture and logic sits at the root of all our modern computing, was designed to be analytical, binary, 1s and 0s. True or false. However, as I learned in James Bridle’s excellent “Ways of Being”, Turing noted that another machine was possible: “A choice machine”, or “oracle”.

“Unlike the [analytical] machine, which steps through its instructions relentlessly until they are complete, this [oracle machine] pauses at critical moments in its computation to ‘await the decision’ of ‘the oracle’.

I’ve written about this before - and in that frame of mind the oracle was the world itself. And the system was an “ecology of technology.” But if the world is the oracle and the computer is (increasingly) the analyzer or executor, what does that make us? The medium? The conductor? The artist? James Bridle: “This is the dance of anthropocentrism: transparency does not equate to understanding; seeing does not mean knowing or dominating.”

What role in technology, then, do we want to fulfill?. We need to decide before the choice is made for us.