Two metaphors for letting go of what you’ve learned.

The Buddha, in the Alagaddupama Sutta:

This raft has been very helpful to me… Why don’t I beach it on dry land or set it adrift on the water and go wherever I want?

The teaching is a raft. It’s for crossing over, not for holding on.

Wittgenstein, at the end of the Tractatus:

My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them – as steps – to climb up beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.)

The ladder gets you somewhere. Then you kick it away.

The question I kept hearing

I spent much of 2025 talking to software companies about AI. Back in March, the pitch was that these tools are actually usable right now. By autumn, that was obvious. The conversations changed.

But one question came up constantly, from people who’d started to get it: will developers ever learn to code the hard way again?

The parallel

We’re mechanising coding. It’s worth thinking about what happened when we mechanised farming.

Most people worked in agriculture at the start of the last century. Now almost no one does. Growing your own food is a hobby, something you do for reasons other than necessity.

But to mechanise agriculture, we had to master it first. You didn’t just need the calories to sustain civilisation through to the Industrial Revolution. You needed an intimate understanding of the craft in order to build machines that could do it.

The amateurization

Software engineering already has an expertise problem. Or rather: the industry has been filling up with people who were never really professionals in the deep sense.

Alan Kay, in a 2004 ACM Queue interview:

I fear — as far as I can tell — that most undergraduate degrees in computer science these days are basically Java vocational training.

Jonathan Blow, on generational knowledge loss:

The first generation works on thing X. After X is done and becomes popular, time passes and the next generation of programmers comes and works on Y, based on X. They do not need to know, exactly, how X is built, why it was built that way, or how to write an alternative X from scratch… The biggest a-ha moment was that if you are working on Y and Y is based on X, that does not imply automatically that you would know X also.

Ryan Dahl, creator of Node.js and Deno, January 2026:

The era of humans writing code is over. Disturbing for those of us who identify as SWEs, but no less true. That’s not to say SWEs don’t have work to do, but writing syntax directly is not it.

Think of the equivalent of the Beatles in the historic lineup of programmers and computer scientists. Von Neumann, Dijkstra, Grace Hopper, Licklider. How many names would ring a bell for your average team of web developers? Have they heard of the Mother of All Demos?

The industry has been doubling every five years for the longest time. That means fewer people joined out of love for the craft. The rungs of the ladder were always just stepping stones for a lot of the workforce.

For professionals, the metaphors apply. They’ve crossed something, climbed something. For amateurs, there’s nothing to let go of.

The difference

Software is different from cutting down corn. At its core, we’re producing coherent, complex systems in spite of the obstacles that reality throws at us. The craft isn’t just executing tasks. It’s managing complexity, holding structure in your head, making trade-offs that compound.

A 21st century farmer doesn’t have the same expertise as a 19th century farmer. Things have changed. Some knowledge has fallen away, replaced by different knowledge.

I don’t think this is true yet for software. I’m sympathetic to people like Robert Martin, who argue that software is fundamentally about managing details, and since you can’t delegate responsibility for details, you’ll always have to worry about them. Although I’m not confident I agree.

The temptation

The biggest temptation is to just ask the AI to complete your ticket.

You’ve arrived in your software job, there’s a Jira board, your name is on a task. The dopamine hit from completion, the clout from your non-technical product manager, is roughly proportionate to how quickly you get things done.

But that’s only the short-term payoff. The longer-term payoffs are the health of the codebase and your professional development, your abilities as a craftsperson. Those are affected by what worthy detours you choose to make along the way. How much you actually build a grip on how things work and why.

There’s a principle I’ve heard in hiring: look for people who go one level of abstraction lower than the thing they’re asked to do. These are often the most useful people. I wouldn’t expect a web developer to dive into electrical engineering. But a passing interest in how we got from machine code to talking about React components seems valuable.

The trap

I think one of the traps with AI, because we’re all lazy beings, is to think: at last. At last a raft we can discard. At last a ladder we can kick away.

But for most of the industry, this is backwards. You can’t discard a raft you never built. You can’t kick away a ladder you never climbed.

And there are natural questions that follow about how juniors will ever truly learn anything.

I think the truth is more complicated. The juniors of today will be far more adept at the current level of abstraction. But the less diligent among them will always feel that something’s wrong. They’ll feel like imposters. This is something you can explain to people, and some fraction will take up the challenge and learn deeper.

The last seventy years of software industry history suggest it will only be a fraction.

The tutor

AI is fantastic as a tutor, and only getting better. But the horse has to want to drink.

For professionals, there’s a real question here. When do you let go of hard-won knowledge that’s becoming obsolete? That’s the raft. That’s the ladder. The metaphors apply.

For everyone else, the question is different. You can’t let go of something you never held. The work isn’t discarding expertise. It’s acquiring it in the first place.