What is technical debt, and how does it apply to large language models? We dive into a really interesting conversation that goes from technical debt into system and code maintenance, which is probably a much better way to think about the challenges we have in maintaining the infrastructure systems, code, data and data lakes that we have to deal with on an everyday basis.
How do we maintain, store, track and update the LLMs themselves? How do we know and manage which model is being used when we retire a model?
References:
www.linkedin.com/pulse/how-google…h-debt-abi-noda/
devops.com/are-llms-leading-de…o-a-tech-debt-trap/
Transcript: otter.ai/u/ngUClgtMmLLKXCFalD…?utm_source=copy_url
Image: www.pexels.com/photo/anonymous-w…y-gloves-3846440/
Cloud2030LLMsChatGPTTechnical DebtMaintenanceCoding
Rob’s Hot Take:
In the Cloud 2030 podcast episode on August 17th, Rob Hirschfeld discusses the distinction between technical debt and system maintenance costs, emphasizing the importance of understanding the ongoing effort needed to maintain and improve systems. He points out that overlooking the maintenance costs while building a system leads to technical debt. Hirschfeld raises the question of whether large language models and AI can change the equation of system maintenance, a topic yet to be explored fully in the podcast.
