Is 2025 even harder than we expected?

We review 2025 predictions today and dig into why I think this year is going to be both boring and terrifying for a lot of enterprise IT leaders. That, of course, spans Amazon, Reinvent storage, VMware, AI, and Agentic AI – we run the gamut on what is coming and why this is actually going to be a very challenging year.

Transcript: otter.ai/u/H6UvLC-r2zmBO9A5jf…?utm_source=copy_url

Reference: zenoh.io by ZettaScale

Cloud2030predictionsstoragecontainerscloudAWSvmwareai

DeepThink AI and Kubernetes

We springboard from DeepThinking AI and have a robust conversation about what impact DeepThink is having on the industry. We also discuss where we see things going into the dilemma of people building AI infrastructure and working to do that quickly, robustly and with strong governance. This is necessary to ensure that they can quickly update and manage that AI infrastructure that they’re spending so much money to build, and this leads into a broader conversation about virtualization, containers and open shift.

Recorded Jan 30, 2025

Transcript: otter.ai/u/79JxdYOiXUoSS44pYP…?utm_source=copy_url

Reference: www.perplexity.ai/search/provide-a…TQ6LJG_X0SlB5g#8

Why is adding LLM into an App so hard?

We talk about current events, the acquisition of data stacks and the closing of the HashiCorp acquisition by IBM. Later, we dive into the productivity of AI and what’s going on – are companies really getting the benefits that they expect from AI chat bot integrations and what the challenges are?

We touch base on a little bit of something more infrastructure focused, where I give a preview of work I’ve been doing on separating Kubernetes virtualization from Kubernetes development use cases, which is something that we will be talking about more in the future.

References:
www.windowscentral.com/software-apps…ind-a-paywall
www.ibm.com/new/announcements/i…ise-ai-applications
www.youtube.com/watch?v=Ioc3r70HNLM
www.linkedin.com/posts/dhinchclif…9498138624-jR2R/
20250227

Symbolic AI

We discuss Symbolic AI via LLMs for advanced reasoning in manufacturing and real-time analytics. Key points included leveraging symbolic representations and algebraic equations, utilizing knowledge graphs to improve model accuracy, and exploring agentic AI frameworks with specialized agents working together using swarm intelligence principles to tackle complex problems like anomaly detection and process optimization. The group also discussed the challenges of building trust in AI systems and the importance of capturing and storing questions and answers to build a knowledge base.

Training Small LLMs

In this episode, we dive deep into the emerging world of building and training small language models. We’ll discuss the benefits, risks, and challenges companies face as they work to create more targeted and efficient AI models. From managing hardware and power requirements to ensuring data privacy and governance, we’ll cover the key considerations for enterprises looking to leverage the power of small language models. Join us as we unpack this fascinating topic and consider the implications for the future of AI and infrastructure operations.

Transcript otter.ai/u/xJ5T-x70WUFQ55ZAsRQr57q6zwE
Reference: www.composabl.com/

AI Platform Consolidation & Walled Gardens

We discuss the impact AI and data sovereignty data protection will have on platforms, consolidated management of your data like in Office by Microsoft or Google, on premises, and systems. This includes a whole bunch of data that you will want to use to train AI models to improve your day to day operations, but you probably don’t want a lot of vendors pulling that data apart and transiting it. We have a fascinating discussion about how the market is impacting these forces.

Transcript: otter.ai/u/gpyotjoYz5ev-1Q2yv…?utm_source=copy_url