In the latest episode of Talking Tech with Travis, I’m joined by my good friend, Matt Baker, who you may know as Dell Technologies Chief of AI Enablement. I couldn’t think of a better guest to join me for a lively discussion as we reflect on what we got right and what we got wrong about generative AI (GenAI) as an industry, followed by an exploration of where we currently stand and where our customers are in their journey with GenAI. Finally, we’ll conclude with a bold prediction for the future of GenAI based on our evolving understanding and technological advancements.
Contrary to initial expectations of widespread customization and fine-tuning of models, the reality has led to a pragmatic adaptation approach. Enterprises are realizing that existing foundation models often lack the context necessary for effective business use cases, prompting a shift toward infusing AI applications with their own data for relevance and efficacy. The focus has narrowed down to key applications such as enhancing coder efficiency, support and sales assistance and marketing content generation, where understanding enterprise-specific knowledge is paramount.
We’ve shifted from the idea of everyone building their own models to a pragmatic approach of adapting existing models for business use cases by infusing them with our own data. The focus now is on understanding enterprise-specific knowledge and modernizing processes to optimize the effectiveness of AI applications.”
Matt and I emphasized the importance of modernizing processes before integrating AI technology. As businesses transition from experimentation to building core AI capabilities, we recognize the value in smaller models combined with well-designed pipelines, challenging the notion that bigger models always yield better results. Looking ahead, I challenged Matt to give us a bold prediction, and he predicts a future where the combination of smaller models and techniques like retrieval-augmented generation (RAG) will drive more efficient and cost-effective AI applications, paving the way for a new era of innovation and productivity in enterprises. For a deeper dive, be sure to tune into the episode below or find my series Talking Tech with Travis on YouTube. If you learned something new, please like and subscribe!
We’re realizing that smaller models combined with well-built pipelines can deliver excellent results at a fraction of the cost and computing complexity compared to larger models. It’s about leveraging the right techniques like retrieval-augmented generation to drive efficiency and effectiveness in AI applications.”
See how we’re simplifying AI in the enterprise with the Dell AI factory with NVIDIA. Learn about Dell’s Gen AI solutions, server solutions for AI and GenAI and Dell validated designs for AI and GenAI.
