In an annual overview interview conducted by investment banks such as The Goldman Sachs, featuring Marco Argenti, the chief investment officer, it was highlighted that the era of “hybrid” AI and the emergence of software operating on extensive language models will take center stage in the realm of artificial intelligence.
Argenti, in reference to foundational models like ChatGPT, explained that “Hybrid AI” involves utilizing these larger models as the guiding force that interprets user input swiftly, acting as an orchestrator that delegates tasks to various specialized employee models tailored for specific functions.
Argenti expressed that only the wealthiest businesses globally would have the financial capability to develop such extensive programs. Consequently, the majority of enterprises are expected to settle for creating smaller neural networks using their proprietary data, either internally or by leveraging cloud computing services externally.
The trend of refining specialized models with corporate data aligns with the current movement towards enhancing functionality, exemplified by projects like LangChain, an open-source framework rooted in generative AI technology.
Moreover, Bill Gates anticipates a significant surge in technological advancement stemming from AI in the near future.
Argenti projected that in 2024, most businesses will focus on proof-of-concept initiatives that promise the highest returns. He also hinted at the introduction of a new category of third-party software built on foundational models in 2024, alongside advancements in infrastructure development.
Argenti envisioned a plethora of untapped applications revolving around these models, likening them to operating systems or platforms that pave the way for further innovations.
Highlighting the importance of investing in the tool and software layers, Argenti suggested that substantial changes could materialize as early as the following year.
Furthermore, Microsoft’s GitHub Copilot is delving into the precise timing and pricing dynamics of AI integration in software development.
Argenti stressed the critical need for seamless security integration across multiple entities as a key challenge moving forward.
Looking ahead, fostering a collaborative environment, promoting open-sourcing of models when appropriate, and establishing principled guidelines to mitigate risks such as bias, discrimination, safety, and privacy will be imperative for advancing technology. Argenti emphasized that these measures are essential for ensuring that the United States maintains its leadership in AI innovation.
