The landscape of generative AI in 2023 resembled a rollercoaster. The year was marked by a significant surge in investment, largely driven by the excitement around cutting-edge AI models such as ChatGPT. This influx of capital boosted the fortunes of leading large language model (LLM) developers, including major players like OpenAI, Anthropic, and the European up-and-comer Mistral AI.
On the flip side, the rapid pace of AI development sparked widespread concern. This culminated in several open letters advocating for a pause in the progression of advanced AI systems. These calls for caution were even endorsed by influential leaders from top tech firms like Elon Musk.
But the year's most dramatic episode unfolded within OpenAI's boardroom. The unexpected ousting of co-founder and CEO Sam Altman sent shockwaves through the AI community. This tumultuous period required Microsoft CEO Satya Nadella to step in to steer the company back on course in a swift and decisive manner.
As we turn our focus to 2024, the generative AI sector appears poised for another year full with excitement and perhaps its fair share of drama. The industry is evolving at a breakneck pace, presenting both challenges and opportunities.
Looking ahead, let's explore some of the emerging trends in generative AI that might not yet be on everyone's radar.
1. Emergence of Small-Language Models (SLMs)
In 2023, LLMs dominated the AI landscape, but 2024 is shaping up to be the year of the small language model (SLM). The greatest potential of generative AI lies in its application to specific business functions such as HR, sales, and finance, as well as its adaptability to distinct industries like healthcare and energy. This is where the role of domain-specific SLMs becomes crucial.
These specialized models are adept at comprehending the unique vocabulary and jargon of specific fields. Their expertise is not just limited to language. They also possess a deep understanding of the nuances of a company's products and services. As a result, SLMs are capable of generating highly relevant and precise content, significantly reducing the likelihood of generating inaccurate information or hallucinations.
2. Data Wars
In 2024, data control and management will become contentious issues. Customers of AI solutions are concerned that generative AI vendors might use their proprietary data to train models and then commercialize the resulting technology.
What's more, the costs associated with data transfer are poised to be a significant challenge. Generative AI relies on extensive datasets, often hosted on major platforms like Snowflake, AWS, Google Cloud, and Azure. These companies may be reluctant to facilitate the movement of data away from their ecosystems. This reluctance could lead to major disputes as customers seek more autonomy and flexibility in how their data is used and stored.
3. Seeking Optionality
The events surrounding OpenAI have underscored the hazards of depending on a single generative AI provider. In response, the trend for 2024 is leaning toward diversification and flexibility in AI licenses. Customers are increasingly seeking options to safeguard against potential risks. This is leading to a strategy where companies are building infrastructures that support the interchangeability of different LLMs and SLMs.
Furthermore, there's a growing interest in open-source AI models. These models offer advantages in terms of greater control and the ability to customize according to specific requirements. Open-source options are becoming attractive as they provide a level of transparency and adaptability that proprietary models may not offer.
4. Reckoning
In 2023, the generative AI sector was largely focused on innovation and exploring new possibilities. However, as we step into the next phase, the emphasis is shifting toward realizing a tangible return on investment (ROI). There's an emerging expectation for clear financial benefits, such as achieving a $3 return for every dollar invested in generative AI technologies.
While this might appear as a steep hurdle, the potential of generative AI to significantly enhance productivity supports these high expectations. One of the key strengths of this technology is its ability to extract valuable insights from unstructured data, a task that has traditionally been challenging and labor-intensive. Nonetheless, it's also likely many corporate ventures into generative AI will not meet these high expectations.
In fact, one factor influencing the difficulty in achieving a satisfactory ROI from generative AI lies in the business models of many large software-as-a-service (SaaS) companies, especially those in the customer relationship management (CRM) and enterprise resource planning (ERP) sectors. These companies traditionally monetize their services based on the number of user licenses or seats.
The incentive will be for them to substantially hike the prices for their generative AI features. Such a pricing strategy could potentially overshadow the cost benefits of implementing these advanced technologies.
5. Where's the Killer App?
Historically, major technological trends have been driven by killer apps. But with generative AI, what's notable is that there are really none. While there are innovative applications, such as ChatGPT, none has yet achieved the status of being a daily necessity for the masses, especially when compared with ubiquitous platforms like Google.
The year 2024 could be pivotal in this regard. If this year doesn't witness the rise of such killer apps in the generative AI space, the growth and wider acceptance of this technology might face significant hurdles.
Conclusion
Heading into 2024, generative AI is at a pivotal point. The sector is on the cusp of significant advancements and breakthroughs, presenting immense potential for growth and innovation. Yet, it's simultaneously navigating a landscape riddled with complex challenges. How successfully these challenges are managed and navigated will be vital in determining the sector's trajectory and its ability to harness its full potential in the years ahead.
About the author:
Muddu Sudhakar is a successful entrepreneur, executive, and investor. Muddu has deep product, technology, and GTM experience and knowledge of enterprise markets such as cloud, SaaS, AI/machine learning, IoT, cybersecurity, big data, storage, and chip/semiconductors. Muddu has strong operating experience with startups as CEO (Caspida, Cetas, Kazeon, Sanera, Rio Design) and in public companies in SVP & GM roles at the likes of ServiceNow, Splunk, VMware, and EMC. Muddu has founded five startups, and all of them have been successfully acquired and provided 10x returns for shareholders and investors. His latest startup, Aisera, has attracted funding from top-tier investors like Webb Investment Network, World Innovation Lab (WiL), True Ventures, and Thoma Bravo.
