Ask HN: Will the future have several more programmers or a few excellent ones? | Hacker News Reader

Ask HN: Will the future have several more programmers or a few excellent ones?

3 pointsanimeshk10 comments

Hey guys.

Just curious about this. Currently there seems to be an increasing trend in the number of people who are becoming programmers. But there also seems to be an increasing trend in auto-programming software and apps coming out. Now as machine learning becomes more and more effective, these programmes may be heavily driven by AI.

That said, it's interesting to think which trend will evolve over the other. What do you guys think? Will there be more programmers in the future or will there be very few, excellent programmers and building things would not require any programming knowledge at all?

Thanks.

View on HN

There will be just as many or more people using technology to solve problems, having access to a large number of methods.

But programming for user interfaces, networking and databases might not be the bread and butter of the future computer engineer, just like coding in assembly is rarely used today, even though all programs are eventually compiled into it.

We'll move higher on the abstraction stack. That will make programming both simpler and more complex, as C++ made programming both easier than assembly, and much more complicated, when you consider OOP.

That's an interesting take. It's also interesting to imagine what the higher abstraction layers would be like, say a hundred years from now..

You know, when humans find it hard to make sense of software requirements expressed by other humans in order to produce usable software, I am not sure if AI can. Also, all this AI thing as it stands now is mostly about data processing. So until NLP becomes a reality, in that Turing test is smashed to pieces, I am sure programmers won't become irrelevant.

However, I see programming becoming more of a DIY thing. In fact, one might argue that it started out as DIY, with scientists and engineers hacking software for their own use. Still, the task of building non trivial software, based on off the shelf tools or custom made, will need professional programmers (aka software engineers), much like how despite people increasingly DIYing many things there still are carpenters, electricians, automechanics, tailors etc. So I would say there will be more.

About excellent programmers, I don't think everyone can agree on what parameters constitute excellence, and it is quite arbitrary, so that part of the debate will rage on even in the future.

How far in the future?

We're a long way from AI or building-block platforms being able to replace programmers in any meaningful sense. Or from demand for custom-built software tapering off. Over the next 2-3 decades we'll see more and more software engineers employed.

But in the long run, yes, AI will replace almost all of us (assuming we don't do something dumb like submerge the planet in nuclear winter before then). How long? Wish I knew. Probably less than two centuries.

Any time frame in between then? It's difficult to impossible to know how long the process of developing AI that powerful will take, how long it will be until it's cheap enough to handle most demand for software, or when the inflection point will be in terms of employment.

I guess coding will be more common over time; but people coding will not be calling themselves "programmers": their main job function will be something different (like marketing, sales, or customer service).

There is some work for developers integrating those "small" developments into something bigger enterprise-wide. This is happenning today, but I guess this trend will get more volume and speed.

Great developers will always be scarce, like any other professionals (doctors, lawyers, etc.)

As artificial intelligence increases, the number of jobs will start to decrease. Once the job started decrease, the policy makers will step in and AI will be in handcuffed. This will result in many monotonous jobs without a learning curve for lots of overqualified programmers

I think the more we deploy AI to tackle trivial jobs, the more creative we will get. Because we would have more free time to think about bigger ideas. True, it'll be a slow transition, but I think our transition from manual programming to a world with only AI programming will also be slow. We, as programmers are creative people and wouldn't want to feel like victims of our own invention. In fact the whole human civilization will benefit by having powerful tools at their disposal to build great things.

That is, if there will be very few or no programmers in the future.

This assumes that AI will be centralized. Far from it (brains are small, and in terms of pure computing power, artificial computers will reach and surpass brainpower very soon; the rest is software).

Once each programmer can have his own artificial intelligence doing his job, how will you prevent programmers to replace themselves by their personal AIs, and go to the beach instead of slaving behind their computer?

cf. http://www.bbc.com/news/technology-21043693 the best employee of the world!

We need both engineer and mechanic. A mechanic can never replace an engineer.

When AI are everywhere, the ability to tinker with them, repurpose and reshape them to your specific needs and desires, will then be the point of difference. So coding will always be important.

If AI-in-everything becomes as ubiquitous as the written word, which it will, our lives will be vastly transformed. The ability to code will become like the ability to write is now: the absolute bottom rung of the ladder to enter society as a participant.

If AI and energy become essentially zero cost then the differentiator will be how clearly you can tell them what you want them to build, and how useful are the things you choose to build. Oh, wait, so no different to coding today. So coding will always be important, and increasing saturation of advanced technology will only make it more so.

If you are assuming, as counter to these points, AI where you can essentially say: make me a FB clone that will make me 1 trillion dollars per year, finish it in 2 hours. As in, you have a Magic Lamp, then, either everyone will have a Magic Lamp, and the differentiator will be your mind, or only those who can afford it will have a Magic Lamp, so the differentiator will be money, equivalently again, your mind. So coding, in the purest sense of giving instructions, will be important.

To put all this another way, when the number of things to which people can give instructions to produce results increases in number then will also increase the number of people who are capable of giving instructions to such things.

Reply on news.ycombinator.com