Moving Forward

We tend to react to new or unknown things emotionally; it’s hard not to. The initial response is defensive.

Later, our analytical mind assesses the situation: how likely is this to be a real risk, what are the consequences of inaction, and what steps could mitigate it.

Then we decide: act or do nothing — doing nothing is still a choice.

Where is the Bear?

This probably sounds overdramatic; it’s not as if we’re being chased by a bear. (If you literally are, please stop reading this drivel and get to safety.)

But in the longer view, this could have a profound and lasting effect on our careers, at least in how we conduct them today.

Past Disruptions

Throughout history, inventions have repeatedly removed the need for human intervention or manual labor. Each time such advances arrived, they were perceived as a threat to the livelihoods of those affected.

In some cases, the impact was real: farmers no longer sow and reap fields by hand; switchboard operators are no longer needed to route calls; clerks no longer push carts of files through rows of filing cabinets. These jobs are gone or are now very different.

We are web farmers

Professional web developers no longer write all the code deployed to a website by hand. It’s still possible — like tilling soil without a tractor — but technology now removes many low-level, routine, and uninspiring tasks and allows us to focus on more interesting and impactful work.

We’ve come to accept this with web frameworks, libraries, content management systems, and npm packages. Some advances never gain traction, however, and ultimately the difference comes down to a few factors:

• Safety — Does the advancement’s benefit outweigh its drawbacks or harms?

• Reliability — Can we count on it to perform consistently?

• Effectiveness — Does it actually do what it claims?

Technologies like WYSIWYG editors proved unreliable at consistently producing compatible, extensible code. Flash advanced interactivity but alienated users on assistive technologies, low-end hardware, or alternate screens. Java applets enabled serious in-browser applications, but their bandwidth and processing overhead was excessive.

And now, AI

At first I worried: after nearly thirty years in the field, would my career last another three? Would I be replaced by someone with less experience who subscribes to more/better AI services?

Now, moving past that initial reaction, I see that AI is a tool. As with any tool, evaluate it by three criteria:

• Safety — Does using AI-generated code introduce unacceptable risks?

• Reliability — Can AI produce the results we want consistently?

• Effectiveness — Does the code meet our standards and perform as required?

We accept web frameworks, content management systems, and npm packages because they are largely safe, reliable, and effective. We should evaluate AI and other new tools on the same basis.

Moving Forward

While there will certainly be more disruption, AI is not a bear. It’s a tool we should evaluate and test — and if it proves safe, reliable, and effective, we should consider adopting it.

It can replace parts of our workflows that are tedious, repetitive, or uninteresting.

Moving forward, we should evaluate AI code generation for use in rote tasks where it helps, freeing us to focus on parts of projects that require human judgment and care.

But if you choose to do nothing — not even consider it — this disruption might feel like a bear.

🐻