Is Artificial Intelligence going to put us all out of a job? Will it replace software engineers? Will we all have to bow down before our new AI overlords, or will AI just provide an extra pair of hands when they’re needed? Dootrix’s Director of Technology, Kevin Smith, thinks we have less to fear than to look forward to.
‘One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations; to put it on is a peculiar business, to whiten the pins is another; it is even a trade by itself to put them into the paper; and the important business of making a pin is, in this manner, divided into about eighteen distinct operations, which, in some manufactories, are all performed by distinct hands.’
Adam Smith published his famous description of pin manufacturing in his book, The Wealth of Nations, in 1776, the year of the American Revolution. There were other revolutions underway. The world of work was changing. Like many other contemporary walks of life, the complex – by the standards of the day – process of making pins was being made that much more efficient by dividing the task into separate functions.
In the two hundred and fifty years since Smith wrote his seminal book, we’ve seen three major periods of automation-driven change in the workplace: the age of steam, the age of mass production and the age of the computer. Now we’re about to see a fourth.
And in every instance, the changes have provoked fear and have been resisted. Even now. For instance, research published four years ago suggests there was a strong correlation between areas that voted for Brexit and those most likely to be affected by automation. People get worried. And yet each period of change has resulted from, and encouraged, innovation. They’ve led to new products, new jobs and new services and have raised our standards of living.
We don’t yet know exactly how AI will change the world we live in but many of the sectors most likely to be affected involve human knowledge, ingenuity and creativity; amongst them law, the arts and entertainment, media, design and software engineering.
So, will the developer of the future bear more resemblance to The Terminator than to your average Jane or Joe who spends a few too many hours staring at a screen? Probably not, at least not for a while yet. But what we are likely to see is the increasing availability and sophistication of tools that will allow engineers to do more.
AI rests on the data it’s trained on. The more specific and targeted the data set the better tuned the AI is for a given application. However, step outside that field of training and it becomes much less useful than a general AI.
For instance, if you’re a competent developer, you could expect to achieve significant productivity gains with an innate coding assistant like GitHub Copilot. Embedded into existing development workflows it can automate a lot of the boilerplate code a developer would normally have to write by hand. Auto-complete is hardly new, but Copilot takes it to a whole new level and can generate a lot of code with minimal prompting.
Copilot, and other innate coding assistants, are fine-tuned for development; they are not a general-purpose AI, nor do they necessarily provide a chat interface like ChatGPT.
ChatGPT is more akin to Google. It ‘knows’ lots of stuff about lots of things, so it too can be used to generate code, via its chat interface. However, if you choose to use ChatGPT you need to know the right questions to ask, and you’ll need to check that you’ve got the right answers. It can be a better solution if you want to 'explore' certain ideas and have them expressed as code, or if you want to elicit specific algorithms to do certain things.
From baby steps to giant leaps
Plenty of people have been caught by surprise at how fast AI is developing. As existing AI speeds the development of future AI, we can expect the rate of development to accelerate.
We will see AI become more pervasive. Just as every word processor ended up with spelling and grammar checking, every development tool will end up making use of AI to improve developer throughput.
It will become more advanced. We are still at the 'baby-AI' stage and this technology is its still in its infancy.
It may also become more tailored. Already experts are suggesting that large language models may be capping out and future gains will be made from more specific models, trained on specific data sets/problem domains.
I would also anticipate convergence. Just as we all gravitated towards the search engine and started seeing little search boxes in all our apps, maybe the chat interface will end up dominating and will be the primary way we interact with AI.
And yet the path to an AI-rich world may not necessarily be without its problems. Recently we’ve seen governments becoming increasingly concerned about the lack of protocols and guardrails to stop AI from doing harm either by being abused by humans or by becoming a dominant super intelligence (where’s Isaac Asimov when you need him).
And while these big potential problems are out of the reach of the likes of you and me there are issues we will almost certainly encounter.
The first is data security: there are already huge concerns about security and data privacy. If you’re typing sensitive data into ChatGTP you may be in violation of the law, or in breach of non-disclosure agreements. You need to be clear on the terms under which you are using the tools and invest in alternative plans or enterprise agreements where necessary.
The second is accountability: for the moment, it is still a tool to help experts be more productive, not a tool to replace the experts. As inferred earlier, in terms of software development you should treat ChatGPT like you’d treat a junior developer; you still need to check its work. The results may still need some significant re-working, and in some cases it might just be wrong. "The AI messed up" is not going to be an answer that inspires confidence if everything goes belly-up.
I want to say that because AI is still in its infancy we have time to get used to it and adapt around it. But I’m also reminded that at one moment a baby can only crawl and all of a sudden it’s got up on two feet and is running around, the next moment it’s talking and before you know it wants its own car. AI is developing fast. It might be comforting to think we can keep ahead of it. The reality may be that we find ourselves running to keep up. Time to get in shape.
Want to find out how to prepare your organisation for the AI revolution. Join our Azure Integration Services Workshop.