2024 GenAI Predictions: Part One
Something unexpected happened in 2023: the world learned about generative AI for the first time. Now that the GenAI cat is out of the bag, companies are in a race to monetize it without succumbing to its negative aspects. That makes GenAI extremely interesting to watch 2024, and the subject of numerous 2024 predictions.
According to Will Falcon, the CEO of AI toolmaker LightningAI, language models in 2024 will have the same capability as they do now, but at one-tenth of the parameter count. Language models will also need just one-tenth of the data for the same performance, he says, lowering the bar for practical use.
“Transformer will not be the leading architecture, especially in the lower parameter count models,” Falcon says, adding that systems that allow for multimodal AI will be predominant. “RL/DPO (reinforcement learning and direct preference optimization) will enter the mainstream for open-source models; alignment recipes (current moat) will be unlocked,” he says. “Boundaries between pre-training and alignment will start to blur: next token prediction on large corpora will not be the sole strategy.”
AI will move at “warp speed” in 2024 and will be advance enough by the end of the year for companies to rely on it to make major business decisions, according to SambaNova’s CEO Rodrigo Liang. However, there will also be some wreckage along the way.
“After the boom, there will be an extinction for many AI companies as a direct result of enhanced scrutiny around data privacy, security and safety,” Liang says. “As such, 2024 will be the year of the secure, safe harbor AI company, and the explosion in AI investment and innovation will both consolidate and accelerate. Winners will begin to emerge in all fields.”
Andrew Sellers, head of technology strategy at Confluent, says GenAI will become commoditized and embedded in multiple applications.
“It seems unthinkable that a technology as powerful as generative AI will be commoditized as soon as next year, but in 2024 this will start to happen,” Sellers says. “LLMs and other foundational models are already becoming much easier to train and fine-tune, and next year enterprises will start to embed generative AI into more of their applications.”
There will be a rapid shift from cloud-based GenAI to local Gen AI in 2024, predicts Patrick McFadin, vice president of developer relations at DataStax.
“The average startup doesn’t have thousands of dollars to throw at a cloud provider and it will prove almost impossible to run by yourself, but that is changing quickly with the innovation around local generative AI,” McFadin says. “With it going local, you will have a complete RAG stack under your control with your access controls. That way, you won’t have to expose your proprietary data in any way. When we go from centralized, API-based LLMs to local LLMs, it will happen quickly. The ones that will work will be adopted like wildfire. Just be mindful of the downside as de-centralized LLMs introduce the concept of bad actors in the loop.”
Just adopting GenAI won’t be enough to move the needle as a software as a service (SaaS) company in 2024, predicts Saad Siddiqui, general partner at Telstra Ventures.
“Startups that are looking to be the generative AI version of specific categories will struggle as existing software platforms build out competing products and customers look to consolidate their vendors versus looking for new vendors,” Siddiqui says. “It is critical for startups to find a way to innovate on the business model (Siebel versus Salesforce, for instance) and/or acquire a valuable differentiated dataset. The startups that are able to figure out both have the potential to be generational companies.”
Software companies will find new ways of adoption GenAI in 2024, predicts Ajay Kumar, CEO of SLK Software.
“The technology will go from being merely a cost-saving tool to a fundamental aspect of companies’ operations, with benefits such as revolutionizing supply chain processes and delivering more tailored products to customers,” he says. “Right now, organizations are funding generative AI from other departments’ budgets, most notably from data science and analytics. We will see a shift in how organizations allocate funds, with generative AI getting its own budget and a designated leader to oversee integration. However, as AI does require training and customization to reach its maximum capability, full integration will occur gradually over the course of several years, not in 2024 alone.”
The negatives around GenAI will be hard to ignore in 2024, predicts Sridhar Ramaswamy, senior vice president of AI at Snowflake.
“For a lot of people involved in what we loosely call ‘knowledge work,’ quite a few of their jobs are going to vaporize,” Ramaswamy says. “Deep fakes are also another hurdle, and we can expect increased attacks on what we humans collectively think of as our reality–resulting in a world where no one can, or should, trust a video of you because it may be AI-generated. Finally, advances in AI will exacerbate the digital divide that has been happening over the past 20-30 years between the haves and have-nots, and will further increase inequality across the globe. I can only hope that by making information more accessible, this emerging technology leads to a new generation of young adults who better understand the issues and potential, and can counter that risk.”
After a year of irrational exuberance over GenAI, technology leaders will look to plant their feet more firmly on the ground with respect to GenAI in 2024, predicts CallMiner Chief Marketing Officer Eric Williamson.
“I don’t expect AI adoption to slow in 2024,” he writes. “In fact, I expect it to continue to accelerate, particularly for CX [customer experience] use cases. But more business leaders will come around to the idea that generative AI is not a silver bullet–and it is most powerful when used for specific use cases, often along with other AI techniques, to meet specific business needs. I predict that the organizations who ‘get it right’ will be the ones that effectively balance AI velocity and agility with responsibility and security. Those that do this will find themselves in the position to deliver the most value to their customers and improve the bottom line.”
A similar sentiment was expressed by Zandra Moore, the CEO of UK analytics firm Painintelligence, who sees a lot of potential in empowering individuals through GenAI.
“2024’s the year for pragmatic AI in SaaS, with the focus shifting from a Generative AI spree to more savvy innovation,” Moore says. “Predictive analytics will provide users with crystal ball functionality, ever smarter deep learning will tackle ever more complex problems, and causal AI will take the role of ethical hero, helping to explain decisions and ensuring AI efficacy.”
Another tech leader who’s cautious about jumping onto GenAI’s hype wagon again in 2024 is Ryan Welsh, the founder and CEO of Kyndi, who predicts GenAI and LLM hype will start to fade.
“Without a doubt, GenAI is a major leap forward,” Welsh says. “However, many people have wildly overestimated what is actually possible. Although generated text, images and voices can seem incredibly authentic and appear as if they were created with all the thoughtfulness and the same desire for accuracy as a human, they are really just statistically relevant collections of words or images that fit together well (but in reality, may be completely inaccurate). The good news is the actual outputs of AI can be incredibly useful if all of their benefits and limitations are fully considered by the end user.”
By the end of 2024, 95% of consumers in the U.S. will have fallen victim to a deepfake, according to Stuart Wells, CTO of Jumio, an authentication service provider.
“Deepfakes have become highly sophisticated and practically impossible to detect by the naked eye, and now generative AI makes their creation easier than ever,” Wells says. “Misinformation is already spreading like wildfire, and deepfakes will only get more complicated with the upcoming elections. By the end of 2024, the vast majority of U.S. consumers will have been exposed to a deepfake, whether they knew it to be synthetic media or not.”
2024 will usher in a new job description in biotech: people who are fluent in both AI and bioscience, says Amaro Taylor-Weiner, the chief AI officer for biopharma company Almirall.
“As the fields of AI and bioscience grow more deeply intertwined, there will be a greater influx of hybrid engineer-scientists–workers with a dual Ph.D. in biology and computer science,” Taylor-Weiner says. “This will bridge the gap between the pharma and tech industries, supplying the workforce with an army of specialized workers trained to perform in both specialties simultaneously.”
Smaller companies scrambling for computational oomph needed to train and deploy GenAI applications will learn to live without the latest, greatest GPUs that have already been spoken for by the tech giants, says Greg Osuri, the CEO of Overclock Labs.
“As Big Tech corners the market on powerful GPUs, a growing number of organizations will turn their attention to less powerful chips in 2024. Those seeking alternatives will make progress by using less-intensive data set requirements, using more efficient techniques like Low-Rank Adaptation (LoRA) to train big language models, and ‘parallelizing’ workloads, where they deploy clusters of, say, 100,000 lesser chips to do the job of 10,000 H100s,” Osuri says. “The rise of distributed and permissionless networks will permit organizations to harness the power of these lesser chips and increase the overall utilization of all capable chips, which are abundant and oftentimes sit idle. This will allow smaller companies and startups to innovate and make real contributions to the AI boom, mitigating concerns that Microsoft, Google, and Meta will dominate the tech transformation.”
That’s our first batch of GenAI predictions for 2024. Stay tuned for more predictions in the coming days.