According to S&P Global, 2024 will be the year of AI “application builders.” Foundation models, such as large language models (LLMs), have dominated recent discussions. But now investors are increasingly focusing on companies developing AI applications that offer tangible benefits for specific use cases. In fact, according to data from S&P Global Market Intelligence and 451 Research, AI companies without their own foundation models attracted more than twice as much investment in the first quarter of 2024, compared to the same period last year.
One of the most exciting promises of AI is its ability to save workers time. But for AI to have a significant impact, companies need AI tools that are tailored to specific industries or job roles. At the same time, these tools must be reliable and secure. However, while LLM-based AI chatbots can communicate well and offer general advice, they often lack the necessary specialized knowledge or tools. This makes them susceptible to inaccuracies or hallucinations due to their wide range of training data. This is where more specific tools, tuned to their specific use cases, are more likely to provide reliable and accurate results.
Senior Director, New Product Solutions at Dropbox.
To illustrate this point, let’s think about the upcoming Olympic Games. The basic models are like the core traits of a good Olympic athlete, representing physical fitness, dedication, and an unwavering pursuit of excellence. However, the Olympic Games include 32 sports with over 400 different events, each requiring different skills and expertise, just like the various industries and job roles in society. And, while AI provides the basic technology that will power various products and services, each of these individual products must be specialized with the right skills to provide value for its specific use case.
It is rare for an athlete to compete in multiple different sports or disciplines at the Olympics. Each athlete is highly specialized in their specific sport. A sprinter, for example, optimizes their strength and physique to be powerful and fast over short distances. However, this means they are not suited to other disciplines, such as long-distance running. The most prominent AI chatbots today are all-rounders. They are designed to have general knowledge of the world on a wide range of topics. A given chatbot may be able to provide superficial information on a wide range of topics, but it may not excel at more specific tasks.
Take an AI-powered universal search tool as an example. It needs to be able to find and retrieve the right information, quickly. Like a sprinter running the 100-meter dash, it is optimized to shave off crucial seconds each time it is presented. However, there are other tasks that may require an AI designed for sustained performance over a longer period of time, more like the long-distance runner. For example, predictive AI models in business forecasting must learn the activity patterns of each company by analyzing historical data, and develop this knowledge with use over time.
By specializing in the company’s operations, it can provide forecasts about the company’s future trajectory based on past results. Predictive AI models must also constantly adjust forecasts based on ongoing changes in operations and external business factors. But with recent research from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) showing that multiple large language models working together provides a more accurate result, perhaps a new kind of AI ecosystem is emerging.
Is the future of AI a decathlete or a team of specialized athletes?
Looking at the trajectory of the AI ecosystem, we can see two distinct paths the industry can take. The first is a race to create the best general-purpose AI model. This AI system would have high-level performance across a variety of tasks, like a decathlete who can compete in multiple events, from sprinting to long jump to pole vault. The advantage of this path would be a seamless employee experience that streamlines workflow. However, like the decathlete, who may not match the specialist’s performance in a single event, a general-purpose AI model might struggle to reach the same level of excellence as more specialized tools.
The alternative sees the future AI ecosystem as a network of specialized AI products, more akin to a team of specialized athletes. In this model, each AI specializes in a particular domain, much like how individual athletes focus on specific sports. This approach mirrors the way an Olympic team combines the talents of sprinters, swimmers, and gymnasts to maximize their collective medal potential for their country. Specialization ensures that each AI performs optimally within its domain, often outperforming the capabilities of a general-purpose system. However, the success of this networked approach will require sophisticated coordination and interoperability to create a seamless experience for users.
As we try to predict how AI ecosystems will evolve in the future, we can look to the Paris Olympics this summer for a glimpse of two possible paths. Whether we end up with a general-purpose AI tool akin to a decathlete or a network of tools resembling a team of specialized athletes will depend on the goals and decisions of companies in the collective tech industry, much like how each country will have different goals heading into the Olympics.
From strategically focusing on one specialty to optimizing the probability of winning or taking a broader approach to winning as many gold medals in as many disciplines as possible, the type of AI ecosystem each company will implement will largely depend on its own unique goals. For some companies, growth by acquiring market share in a fluid market will require speed and agility, while customer retention in a stagnant market will require a more strategic, long-term plan.
We list the best AI video editors.
This article was produced as part of TechRadarPro's Expert Insights channel, where we showcase the brightest and brightest minds in the tech industry today. The views expressed here are those of the author, and not necessarily those of TechRadarPro or Future plc. If you're interested in contributing, find out more here:
