Q&A
AI from the Trenches, Part 1: Is Generative AI a Leapfrog Technology?
We check in with generative AI expert Cameron Turner to see how transformative and impactful the technology has had on organizations of all sizes.
There is little doubt that generative AI and the inclusion of predictive language models has been a disruptor. Not only has it dominated the public discourse, but it has already made a drastic impact on enterprises and how we get work done.
Redmond sat down for an interview in late January with Cameron Turner in February, an applied AI expert and vice president of data science at consultancy Kin + Carta, to discuss the budding industry and his thoughts on the future.
Now, less than six months since our last talk, and the view a little bit more clearer, we caught up with Turner to see how the industry has changed in that short period of time, and how his company is helping business to transform with generative AI. Here's part one of our interview:
Redmond: Hi Cameron, thanks for talking with us again. What is the biggest development regarding generative AI since we spoke earlier this year?
Turner: An article was published recently by a friend of mine named Morgan Beller. She talks about the concept of leapfrogging and compares generative AI as one of those technologies which will take certain markets and sort of pop over whatever the current approach was, and extend that into interactive and sometimes UI-less context that's afforded by Gen AI. For example, Africa and South America had skipped over the PC revolution directly into mobile, and you can see that Gen AI is affording something very similar.
When you move into a chat front end on top of a large language model, for a lot of scenarios where a probabilistic approach is appropriate, you can enable scenarios immediately so you don't have to sort of go through the whole process of defining a traditional SAS product before you can generate the values. I think that's something that we're going to see as a leapfrog over SAS directly to Gen AI and I credit Morgan with that, that concept.
That's very interesting. So basically you're describing not just a linear progression for technology for a lot of geographies and industries, but almost a hyper linear, exponential growth of technology?
I think that's fair to say. And I'll give you two examples. We're working with a large agriculture company. The scenario that we're focused on is how we can enable agronomists through potentially an earpiece in order for farmers to interact with their historical data for crop, yield and best practices for pesticide and fertilizer application. So while an agronomist is literally working in the field, they can have access to kind of all of the historical knowledge their organization has to better enhance crop yield.
Another example is that we're working with the large construction equipment rental company to improve equipment management, flooring, failures and these kinds of things. What we want to get to is to provide real-time recommendations based on what the operator is seeing out in the yard. So it really just kind of skipping over that time that we would otherwise spend sitting at a PC or a laptop and working with a SaaS product instead.
For these scenarios, what we're trying to answer is, 'how do you enable your workforce and empower them to pull that organizational knowledge in real time in their in their day to day?'
It's interesting because what you're describing are companies in industries that are historically underutilizing tech or are very analog. So it's kind of great to see the effect of this technology for them, not just to catch up, but perhaps also leap ahead, like you said.
Yes, these are industries where paper and clipboards are still around. And there's a lot of green screen applications that are still in use day to day.
"What we're trying to answer is, 'how do you enable your workforce and empower them to pull that organizational knowledge in real time in their in their day to day?'"
Cameron Turner, Vice President of Data Science, Kin + Carta
At the beginning of the generative AI boon, you saw a lot of companies a bit hesitant to start integrating a lot of these new tools into their own work stream because of privacy issues or misinformation. Have you seen the industry, as a whole, take steps to alleviate some of these concerns over the last 10 or so months?
A: I'm old enough to have seen desktop software come in and replace mainframes. What we saw from IT at that time, when I was at Microsoft, was that copies of Excel and Access were coming in through the backdoor because all departments in saw that there was opportunity to use these new tools to help them do their job better. And that led to a reaction by IT to sort of manage that influx of software that people were literally buying at the office supply store and installing on their PCs so they could do their job.
You can say the same thing for mobile. It was undeniable that mobile was going to be an important part of corporate communications, and enterprise IT had to scramble to come up with all the standards for how to manage security and so forth when a lot of information is flying over mobile devices, even though those devices were consumer devices that the employees were using.
I think the same thing is happening now. With generative AI where it's not a matter of something that's going to be approved because OpenAI came through the front door and made a pitch and sale to the IT department. It's that people in marketing departments have a stack of product descriptions that they need to get done by the end of the week. They know that they can do that faster with generative AI, so they're just using it.
So we're seeing the same story play out where it's becoming part of the fabric of what people in all disciplines do in their day to day. So now it's less about an organization making an explicit decision to use it, and more about enterprises racing to catch up with good governance for effective use of the data that belongs to that organization, and finding the pathways that actually ensure compliance and safety with these tools.
How would you rank the biggest players in the generative AI space right now? Is it just concentrated around Microsoft? Should Microsoft be worried about Google? How would you rank this horse race?
Upfront, we have relationships with many of the major players in the space and all have their strengths in the market. And so as a partner to these organizations, one of the things we look for is what their standard is, because we need to work with clients and what the need is. So I wouldn't necessarily make ranks based on one, two, three, but think about what the specific application is needed for every unique situation.
In the context of gen AI, I will say, without going to deep into it, that the source of models have come from Google and how it has organized the world's information, and it's manifested directly into the large language models we use today. So in specific applications, like what we're doing right now for a large food distributor, we're taking millions of product descriptions revitalizing them through generative AI.  Google's large language models work very well for that because they've been scraping the Internet since day one.
Join us in part two, where we continue to discuss how companies are reacting to generative AI, what IT can do to respond and how both private and legislative bodies are reacting for calls for better regulation.