AI everywhere? Where are we going?

We all know that we have entered into the new era of AI. No matter what field we are in, we are thinking, talking, and working towards AI. I spoke to many founders and co-founders in the past few weeks, and they all are working on something related to AI. The wave is real; it's effective and it's more than just chatbots. The number of applications and research papers are rising in this field, creating really cool possibilities. So YES, we should have a basic knowledge of what is happening around us and what is happening within the products that we are using.

I have been learning a lot about Generative AI, its angle with cloud-native technologies, and how Kubernetes is making things easier. The Kubernetes community, along with other foundations in the cloud-native ecosystem, are working together to solve complex infrastructure challenges. Kubernetes has kind of become the de facto standard for running inference or conducting training via batch jobs that run for several days.

Did you know that OpenAI is the organization that published a blog about scaling Kubernetes nodes to 7,500? Hugging Face, Bloomberg, CERN, and Adobe all run on Kubernetes, which shows how significant it is in the AI ecosystem.

Kubernetes KEPs, like Dynamic resource allocation and Improved multi-NUMA alignment in the Topology Manager, show how invested the Kubernetes community is in order to use GPUs more efficiently.

All this demonstrates progress towards helping innovators innovate. This AI revolution is going to be even bigger, but that does not mean it's sidelining other developments, as there are innovations happening at all levels. For example, Wasm with WASI 0.2 is out, platform engineering is introducing more and more tools, and there are more enhancements in the Kubernetes ecosystem. AI is big and impactful and will continue to rise in terms of how we use it, but there will also be significant improvements and developments in other areas, including operating systems and open-source software projects.

What have I been upto?

I gave my talk at Civo Navigate happened last week in Austin. I was about Generative AI in Kubernetes era with Kubeflow and I showed an end to end demo of stable diffusion model running on Kubeflow. It was fun building the demo and learning along the way. With this I wanted to cover the basics of generative AI and Kubeflow with its cloud native angle.

I have a few sessions selected for KubeCon and WASM/IO so I will be there for both those conferences as well.

WASM IO sessions

• Accelerating ML Inferencing with WebAssembly & Spin 2.0
• Create Production-Grade Wasm Applications on Kubernetes [Workshop]
• Sustainability with WASM? - faster, greener computing [Panel]

KubeCon Sessions

Heating Pools with Cloud Power: A New Wave in Green Computing - Saiyam Pathak, Civo & Mark Bjornsgaard, Deep Green TechnologiesPAVILION 7 | LEVEL 7.3 | S06

Building a Tool to Debug Minimal Container Images in Kubernetes, Docker and ContainerD - Kyle Quest, Slim.AI & Saiyam Pathak, CivoPAVILION 7 | LEVEL 7.3 | S05

Sneakpeak of my demo, I will be giving similar talk again with enhanced demo at CNCF Chandigarh meetup on 10th March 2024.

What are your thoughts on GenAI in general and where it is going? Do comment.

Sponsored Content

Without the sponsors I won’t be able to give you an authentic newsletter with all the cool stuff, so please do check them out

• Cast AI (CAST AI)- How Gaming Companies Reduce Their Cloud Costs Using Automation
• Komodor?(Komodor)- Configuration as Code: Everything Developers Need to Know
• Sysdig (Sysdig)- Exploring Syscall Evasion – Linux Shell Builtins

Awesome Reads

• Monitoring TLS Endpoint Certificate Expiration with Prometheus - If you have HTTPS endpoints in your infrastructure, you will want to make sure that none of their TLS certificates are about to expire without you noticing. This blog walks you though on how you can use Prometheus' Blackbox Exporter to monitor for TLS certificates that are about to expire soon, so you can renew them in time.
• No GPU? No problem. localllm lets you develop gen AI apps on local CPUs - Google Cloud introduces localllm, a tool that enables developers to use large language models (LLMs) locally on CPUs within Cloud Workstations, eliminating the need for GPUs and facilitating AI app development on local devices. This solution leverages quantized models, allowing for efficient, cost-effective development with improved performance and data security, fully integrated within the Google Cloud ecosystem.
• Networking in Kubernetes -Simplified - The article simplifies Kubernetes networking for beginners, covering essential concepts such as network namespaces, pod networking, DNS with CoreDNS, container networking interfaces, and the role of Ingress in managing external access to applications. It highlights the differences between Docker and Kubernetes networking, emphasizing Kubernetes' comprehensive approach to cluster-level communication, service discovery, and security.
• Cloud Native Disaster Recovery for Stateful Workloads - This document introduces Cloud Native Disaster Recovery, emphasizing a new perspective on disaster recovery strategies for stateful applications in cloud-native environments. It outlines the essential concepts of availability and consistency, explores the challenges of distributed stateful workloads, reviews consensus protocols for workload coordination, and concludes with disaster recovery strategies for container-native stateful workloads, providing a comprehensive framework for understanding and implementing effective disaster recovery solutions in modern cloud settings.
• Strimzi joins the CNCF Incubator - Strimzi is focused on deploying and running Apache Kafka clusters on Kubernetes. Apache Kafka is a leading platform for building event-based microservices architectures and real-time data pipelines, and it is horizontally scalable and fault-tolerant by design. Running Apache Kafka on Kubernetes can be complicated, but Strimzi reduces the complexity by using the operator pattern. This includes the initial installation as well as the day-2 operations for upgrades and security.?

Kubesimplify Updates

Wasm course hits 4k hits, keep learning WebAssmebly and share within your network.

We have written two new shiny blogs in past 2 weeks as well by Krish Gupta

• Automated GitHub Releases with GitHub Actions and Conventional Commits
• Reducing Image Size using Multi-stage builds for a Go application

Resources and Repos

• GPTscript - Natural Language Programming
• RubraAI - Develop your AI assistants locally.
• Explainix - Explain Nix syntax visually. Snippet below showcases most of the language features Nix has to offer.
• Kermoo - ?? Kermoo offers resilience testing with Process Delays, Back-end Failures, CPU Simulations, and Memory Leaks. Boost your system reliability effortlessly.
• Kubernetes Gateway API tutorial

Learn from X platform

• https://x.com/natfriedman/status/1754519304471814555?s=20
• https://x.com/samcraigjohnson/status/1755428409059250385?s=20
• https://x.com/Aurimas_Gr/status/1755577998072328638?s=20

So subscribe and share if you like this edition :)