Thoughts on the Market: Big Debates - The AI Evolution Hosted by Morgan Stanley | Released on January 10, 2025
Introduction
In the inaugural episode of Morgan Stanley's special series, Big Debates, titled "The AI Evolution", hosts Michelle Weaver, Morgan Stanley's US Thematic and Equity Strategist, and Keith Weiss, Head of US Software Research, delve into the transformative role of artificial intelligence (AI) in shaping global markets in 2025. This episode explores the evolution from AI enablers to more sophisticated applications, the emergence of agentic computing, productivity implications, market perceptions, infrastructure challenges, and the prospects of Artificial General Intelligence (AGI).
1. Transition from AI Enablers to Enterprise AI Applications
Michelle Weaver opens the discussion by reflecting on 2024's focus within Morgan Stanley's research themes, highlighting AI and tech diffusion alongside semiconductors, data centers, and power companies as foundational AI enablers. She notes a paradigm shift as the narrative transitions towards more integrated AI applications in software.
"Throughout last year, the market was largely focused on AI enablers... Now, though, as we're looking ahead, that story is starting to change."
— Michelle Weaver [00:09]
Keith Weiss anticipates 2025 as a pivotal year for software development, emphasizing that the foundational capabilities derived from large language models (LLMs) and significant computational investments are now maturing into tangible software functionalities primed for market adoption.
"2025 is going to be an exciting year for software because... investors can expect to see more of it come into results that the product is there for people to actually buy."
— Keith Weiss [00:58]
2. Emergence of Agentic Computing
A primary focus of the episode is agentic computing, a novel advancement beyond traditional chatbots. Weiss elucidates this concept by distinguishing it from the current state of AI interaction.
"In agentic computing, what you're looking for is to add more agency into that chatbot... It could actually automate the execution of the answer to those problems."
— Keith Weiss [02:02]
He explains that agentic computing integrates reasoning capabilities across multiple models, enabling AI systems to not only respond but also execute solutions to complex issues autonomously. This evolution is poised to revolutionize enterprise software by facilitating more proactive and accurate problem-solving interfaces.
3. Impact on Workplace Productivity
The discussion pivots to the productivity implications of agentic computing. Weiss references Morgan Stanley's 2023 AI Index, projecting significant shifts in the labor market due to AI advancements.
"25% of US occupations are going to be impacted by these technologies... At the high end, $4 trillion of labor that's being augmented or replaced."
— Keith Weiss [03:06]
He anticipates that by 2025, the integration of AI technologies will enhance or replace nearly half of the US labor force, leading to substantial productivity gains. However, the exact magnitude of these gains remains to be quantified.
4. Market Perceptions and Optimism on Adoption
Weiss addresses potential misconceptions within the market regarding agentic computing. He emphasizes that agentic frameworks comprise multiple models and reasoning engines that enhance accuracy and execution capabilities, thereby facilitating smoother adoption in enterprise environments.
"There's a reasoning engine of some sort that's organizing multiple models... It’s going to make this a lot easier to adopt in enterprise environments."
— Keith Weiss [04:01]
This nuanced understanding leads Morgan Stanley to adopt a more optimistic stance on the pace and trajectory of AI adoption compared to the broader market consensus.
5. Overcoming Computational and Infrastructure Constraints
A significant barrier to AI advancement is the computational power required for training and deploying large models. Weiss breaks down the demand into research (core capability development) and development (product creation). He observes that while the emphasis on computational training demand may plateau due to diminishing returns (scaling laws), the demand for inference—the application of trained models to solve real-world problems—is set to surge.
"The demand to utilize these models more fully to solve real business problems... the pace with which data centers can be built out is a fundamental constraint."
— Keith Weiss [05:05]
He highlights the proactive measures companies are taking to secure data center capacities and explore innovative power sources to support the escalating computational needs, ensuring that supply constraints do not hinder immediate advancements in 2025.
6. The Prospect of Artificial General Intelligence (AGI)
The conversation shifts to the highly debated topic of Artificial General Intelligence (AGI). Weiss defines AGI as the ultimate objective of AI research—a model that can reason and operate within the human world with comparable or superior efficacy.
"AGI is basically the holy grail of all of these development efforts... 10 to 20 years before we truly see AGI emerge."
— Keith Weiss [07:22]
He expresses skepticism about the feasibility of achieving AGI by 2025, citing the necessity for significant breakthroughs in algorithms and model sophistication beyond the current trajectory of LLM scaling.
Conclusion
As the episode wraps up, Michelle Weaver and Keith Weiss underscore the transformative trajectory AI is poised to undertake in 2025. While the integration of agentic computing promises substantial productivity enhancements and market optimizations, the realization of AGI remains a long-term aspiration. Morgan Stanley remains vigilant in monitoring these developments, recognizing their profound implications for global markets and the broader economy.
"Looks like 2025 will see some major developments in AI and to our listeners, thanks for listening."
— Michelle Weaver [08:43]
Listeners are encouraged to engage with the series, share insights, and stay informed about the evolving landscape of AI and its market ramifications.
Key Takeaways
- Transition in AI Focus: From AI enablers to integrated enterprise applications.
- Agentic Computing: Advanced AI capable of executing solutions autonomously.
- Productivity Surge: Potential augmentation or replacement of up to 45% of US labor.
- Market Optimism: Morgan Stanley holds a more positive outlook on AI adoption than the consensus.
- Infrastructure Challenges: Critical need for expanded data center capacities and innovative power solutions.
- AGI Outlook: Unlikely to materialize by 2025; remains a decade-plus ambition.
Notable Quotes
- "2025 is going to be an exciting year for software because... investors can expect to see more of it come into results that the product is there for people to actually buy." — Keith Weiss [00:58]
- "25% of US occupations are going to be impacted by these technologies... At the high end, $4 trillion of labor that's being augmented or replaced." — Keith Weiss [03:06]
- "AGI is basically the holy grail of all of these development efforts... 10 to 20 years before we truly see AGI emerge." — Keith Weiss [07:22]
About the Hosts
- Michelle Weaver: US Thematic and Equity Strategist at Morgan Stanley, leading research on market trends and thematic investments.
- Keith Weiss: Head of US Software Research at Morgan Stanley, specializing in enterprise software and AI-driven technologies.
For more insights and updates, listen to the full episode of "Thoughts on the Market: Big Debates - The AI Evolution" available on your preferred podcast platform.