Good old-fashioned AI remains viable in spite of the rise of LLMs

The article discusses the ongoing relevance of task-based AI models in the enterprise, despite the rise of large language models (LLMs). Task-based models, which are designed to solve specific problems, are still widely used in businesses and are not expected to disappear. They can be smaller, faster, cheaper, and in some cases more performant than LLMs. However, LLMs offer the advantage of reusability and can tackle a variety of use cases with a single model.

The role of data scientists in the era of LLMs is also discussed. Despite the shift towards tools aimed at developers, data scientists still play a crucial role in understanding the relationship between AI and data within large companies. Both task-based models and LLMs are expected to coexist in the enterprise for the foreseeable future, as each approach has its own strengths and weaknesses.

Key takeaways:

Despite the rise of large language models (LLMs), task-based models are still relevant and widely used in the enterprise, solving many real-world problems.
Task models and LLMs are seen as complementary tools in AI, with task models being more efficient for specific tasks, while LLMs offer more flexibility and reusability for a variety of use cases.
Amazon continues to invest in both types of models, with products like SageMaker for data scientists and Bedrock for developers.
Data scientists still play a crucial role in the era of LLMs, helping to understand the relationship between AI and data within large companies.

Good old-fashioned AI remains viable in spite of the rise of LLMs | TechCrunch

Key takeaways:

Comments (0)

Newsletter