The machine everyone films — and the worker nobody does

Automation is arriving fast on Bangladesh's factory floors. I'm not against it. But in a country with a labour surplus and no safety net, the real question isn't "how fast" — it's "where do the displaced people go?"

You've seen the video. A bright new machine on the floor, one operator, a smooth run, and a dozen phones held up — wow. It really is impressive. The output, the consistency, the finish.

Here's what the phones never capture: a few steps out of frame, a helper and a second operator quietly packing up their things. The machine did their work. Nobody films that part.

I've run knit factories in Bangladesh for seven years, and I want to say something carefully, because it's easy to be misheard. I am not against automation. Repeatability, quality, consistency — these are real, and manual production genuinely struggles with them. Done right, automation keeps a factory competitive and keeps orders in the country. I believe in it.

But I keep watching the same rooms as everyone else and seeing the other half of them.

The seen and the unseen

There's an old lesson from the economist Frédéric Bastiat: in economics, judge not only what you see, but what you don't. We count the machine's output — that's the seen. We don't count the two wages that just disappeared — that's the unseen. The gain is on a screen; the cost walks home.

That's true everywhere. It's just far sharper here.

But where do we live?

A machine is neutral. Context decides what it means. The same robot that is straightforward progress in a high-wage country — with unemployment insurance, retraining, other industries hiring — is something very different in a labour-surplus economy that hasn't yet built those alternatives. In one place a displaced worker has a bridge to the next job. In the other, there is no bridge yet.

So the honest question was never "how fast should we automate?" It's "where do the 20–30% of workers this displaces actually go — this year, not in a generation?" Until a country has an answer to that, speed isn't a virtue; it's just moving the pain onto the people least able to carry it.

A country of many speeds

Bangladesh runs at ten speeds at once. Automation serves the fastest lane and strands the rest. And the gains rarely spread — they concentrate, quietly, with the few who own the machines, while the cost spreads wide and thin across people who never appear in the demo. That's not a moral failing of anyone in particular; it's just what happens when a powerful tool arrives faster than the ground is ready for it.

The third path nobody films

Here's what the "wow" doesn't let you see: repeatability is not only a property of a machine. It's a property of discipline. A planned floor — proper T&A, honest OTIF, a trained and balanced line — gives you consistency and output without emptying the room. You don't have to lose a third of your people to become reliable. You have to run your process like you mean it.

So the sequence I'd argue for is simple, and it's pro-worker and pro-competitiveness:

Get more from the people you already have — through planning and discipline, not layoffs.
Automate where it protects quality — not to cut heads before there's anywhere for those heads to go.
Invest in the transition — training and reskilling, so that when automation does come (and it will), there's an on-ramp instead of a cliff.

"Be second to none"

The factories the world calls world-class didn't get there by discarding their people the fastest. They got there by raising them — by making the floor better at what it does. Bangladesh's real edge has always been its people. Automating that edge away before we've built somewhere for those people to go isn't just hard on them — it quietly cuts the branch the whole industry is sitting on.

Honestly, why I keep saying this

I say this knowing it's unpopular in a room full of raised phones. I'm not nostalgic, and I'm not asking anyone to stand still. I spent some of the best and hardest years of my life on these floors, and more than any deal I'd like to help build something here that works — for the owner, and for the people who don't make it into the video. That's the conversation I care about.

So tell me I'm wrong, if you think I am. Is it the right time? And if we automate — where do the people go? I'd genuinely like the honest answer, especially from those of you on the floor.

— A.K. · TB Textile Sourcing · Slovenia, EU
(European who ran knit factories in Bangladesh.)