Skip to content
Talk To Us
March 6, 2023

ChatGPT Series Part 2: Q&A on the Disruptive Power of Conversational AI

This is the second of a two-part series exploring the latest trends in conversational AI. 

It took Netflix nearly ten years to amass 100 million users. Instagram took around two years to get there. ChatGPT, the conversational AI tool from OpenAI, reached the same milestone two months after its launch.

For the first time in years, we have a technology with genuine paradigm-shifting potential. Beyond the media hype and ChatGPT-fueled memes, we are seeing clear evidence of how these artificial intelligence tools, drawing on large language models, will change how millions of people do their work and access information.

So how do we prepare for the AI revolution finally kicking up a gear, and what will it mean for your business? To answer that, Accelerance’s Chief Marketing Officer, Lisa Morrell, sat down with the internationally recognized business strategist, bestselling author and tech industry analyst Dion Hinchcliffe, vice president and principal analyst at IT consulting firm, Constellation Research

Dion Hinchcliffe


A veteran of software development, he is one of the country’s foremost thinkers on technology disruption and how to navigate it. He is also well-placed to assess the impact of conversational AI on the jobs of code cutters everywhere. Here’s his perspective on this trending topic.

Accelerance: It seems like the situation right now is a moving target. Can you provide a quick perspective on the current status, as far as new developments in conversational AI?

Hinchcliffe: It's been a shot across the bow, the first real breakthrough in the democratization of AI, so that very powerful AI is available to the average person. Almost anybody can use ChatGPT to do very powerful things, to answer tough questions in a very applied way to their situation, to retrieve data and format it and structure it in just the way that you need.  

Conversational AI, if it’s done right, and we now have evidence of that in the form of ChatGP, gives a regular human being superhuman capabilities. I use it now for doing all kinds of things that save me so much time. You can say, ‘write a job description for a mid-entry level worker with these types of qualifications in this geography.’ 

It takes all that into account and writes a job description. I asked ChatGPT for my biography, which it then gave me and I said, “Well, can you write an article on immersive technology in my voice?” And it said, “Sure”, and it did, because it read everything I've ever written. It was pretty good. It sounded like me. It’s really astonishing. 

Accelerance: How have ChatGPT and Al-powered large language models impacted tech industry competition?

Hinchcliffe: We've already seen evidence of an artificial intelligence arms race. But it was really kind of high-end for companies who want to bake AI into their products. It takes $100 million in computing power to build a model for something like ChatGPT.

There’s going to be a lot of controversy about the amount of power and cost to create these models. Right now, ChatGPT is often not available. You have to wait in line to use it because it takes so much computing power. It’s very expensive to build the model and to run it. It's going to be like mining for Bitcoin soon.

So only really big companies can do it. But they're all in the race now. Google has probably prematurely released Bard. The model is not fully baked, but they had to respond. Everyone's leaving ChatGPT open all day like we used to do with Google Search because it's so much more value-add. You can ask it to present things in just the way you want.

"Changing tech is always the easy part. It's changing people that is the hard part."


Accelerance: The momentum is certainly picking up. What are the drivers?

Hinchcliffe: It’s going to utterly transform work. That’s why you see billions and billions going into it. Microsoft immediately spent as much as US$20 billion just to lock up ChatGPT. 

It's the next big breakthrough. It can really boil things down for you like nothing else. If you’re trying to understand something complicated, I can ask specific questions to dumb it down for me like I'm five years old. Google can't do anything like that. This is why there's so much excitement. 

You're going to see this Cambrian explosion of people trying to get into this space. It’s an open API, so anyone can add these features to their products.

Accelerance: What timeframe do you foresee for the widespread adoption of this disruptive technology? 

Hinchcliffe: We expect that by 2025, AI will be thoroughly saturating our workplaces and the things that we do. How successful will we be? Well, humans aren't really good at technology adoption. But again, the AI is so smart, it really adapts to your situation. We think the adoption is going to be much faster than normal technology, because the AI will just be willing to do whatever you want, at your level, for whatever you need. You just have to tell it. If you can articulate your need, it can pretty much do it. 

The Future of Generative AI in the Workplace (Landscape)-1

Accelerance: What will this mean for technology executives who are managing and developing software applications for digital transformation or innovation?

Hinchcliffe: For tech executives, it has the chance to be a tremendous boon, because you'll soon be able to ask a future version of ChatGPT to actually write a specific code solution. If somebody has a whole bunch of open tickets in JIRA, you can say, I need this bot to come in and write all the code to close out as many of those as possible and give the rest to humans. That's going to be an astonishing power. You actually do that manually now.

I tested it out. How do I change the code to fix this? It will write the code for you. It's amazing. It all will be systematized. Every step of the software development lifecycle will be transformed by AI, from writing use cases, requirements and test cases, to the code itself, the unit tests, the contracts, the terms of service, and the marketing. AI can assist with the entire lifecycle. 

Accelerance: What should tech executives do if they feel like they're getting left behind, or they’re confused about how to adopt this new wave of technology?

Hinchcliffe: The good news? AI companies will do almost all the work to productize and bring this to your doorstep. I would say, don't panic. This will be laid at your feet. 

It is so easy to use. Using ChatGPT is like falling down the stairs. It’s like the reason that Google was so successful. You ask it something and it gives you an answer that, most of the time, is accurate. What you need to worry about more is how you're going to have confidence in the results and the rules for AI workers using it. You want them to use it because it really amplifies their output. But they need to give citations as to the source.

"The good news is that no tech revolution has destroyed more jobs than it has created, ever."


Accelerance: Should software developers be worried about their jobs?
 
Hinchcliffe: Their jobs are going to change. They're not going to do a lot of the tactical things they do now. They’re going to do the more fun, strategic thinking about what their companies should be doing.

If you have somebody who loves writing low-level code, those days are coming to an end. For people who really like to do the same thing every day and work on the same tactical tasks, AI is going to do that. When ChatGPT can do it for pennies, why would you pay a developer? I’ve seen C++ developer jobs command a $200,000 a year a year salary in the Bay Area. Unless they’re very strategic architects, they're going to have a harder time competing.

The good news is that no tech revolution has destroyed more jobs than it has created, ever. If you look at the industrial revolution, and the information revolution, they threw off more jobs. That's most likely what's going to happen here. 

Accelerance: What specific impacts do you foresee for software development, including low-level coding?

Hinchcliffe: The one that's going to happen first is making small code modifications on existing code. You say, “I want this code to do X.” Then ChatGPT returns the code. You just have to check it for accuracy. The job will be mostly to ask the question and then check that what the AI did was correct and that it works. 

Microsoft already offers open AI services that write much larger blocks of code. You describe the natural language code that does something and it will write all the code. Even ChatGPT will do that.

So writing low-level code is going to go away. The future is learning how to specify what the AI should do. Learning what to ask the AIs will become a really important developer task very soon. 

Even if you're not a coder, it can actually help you write an application or develop a website in a very short amount of time. I saw someone who had never coded before build a fully functional mobile application in two hours. He asked ChatGPT each step - “what's the general process, how do I do step one, how do I do step two, for this application.”

Accelerance: How should technology executives approach preparing their organizations for conversational AI as a way of life?

Hinchcliffe: Changing tech is always the easy part. It's changing people that is the hard part. It’s the cultural changes, the skill changes, the inclination, what people want to work on, what really excites them.

I recommend that tech executives think about the impact this transformation has on people. What's it going to mean? Getting the organization ready, building the skills, establishing the culture around automation, and all the real things that only humans can still do, is going to be crucial.

Accelerance: Are you seeing any examples of companies that have been leading the way and have already adopted ChatGPT to great effect?

Hinchcliffe: I just surveyed a number of people across our CIO audience. Only one has said they've ruled out using anything like that. So almost all say they're already in pilot mode and are trying to figure out where to use it. 

Accelerance: Any closing thoughts?

Hinchcliffe: The takeaway is that everything that can be automated with AI is going to be automated. What will you do with that power? And how will you ensure a smooth and fair transition for the business and your workers?

You can find out more about Dion Hinchcliffe and Constellation Research here

 

Lisa Morrell

Lisa leads the strategy and execution of omnichannel marketing efforts that connect software outsourcing prospects and clients with the right experts, global development partners and services - any way, any where, any time they choose. She brings more than two decades of expertise as both a senior digital marketing...

Recently Published Articles

View All Posts