🔥 ☁️Stuff for FinOps, DevOps & CloudOps - FinKube newsletter edition #1

🔥 ☁️Stuff for FinOps, DevOps & CloudOps - FinKube newsletter edition #1
The evolution of Computing with Generative AI and ChatGPT like LLMs

Month in-review

  • ChatGPT - No newsletter or post this month is going to be complete without mention of ChatGPT by the Open AI Foundation
  • What is ChatGPT?

  • ChatGPT for Developers and Operators - Great Stuff!!

Great for writing code (not inventing novel methods and algorithm)

Also, great for tuning K8s or Terraform YAML syntax

  • ChatGPT for FinOps - Not so great ;)

The answer is hand-wavvy at best but ChatGPT does point you in the right direction!

  • ChatGPT for predicting the future - Completely falls apart!

I am so glad ChatGPT falls apart at predicting the future 🙂

  • Takeaways evolution of computing and DevOps teams

The above graph was inspired by a recent tweet from Martin Casado, where he laid out the case of how AI will not optimize compute but rather capitalize on using “more” compute. For example, we can consider the last 20+ years of programming models used by DevOps teams as follows:

  1. Iterative programming - In this modality, we tell the computer what to do, step by step, and most of us grew doing that over the last few decades. This modality has served us well with the limited CPU cycles that most folks had access to.
  2. Declarative programming - This modality is a recent emergence where we tell the computer “what we want” and let the computer “figure” out what to do to get us what we want. Terraform and K8s YAML come to mind among some of the more popular incarnations of this computing model. I was exposed to this model as in early 2010’s we wanted to abstract away the complexity of iterative programming infrastructure.
  3. Generative AI programming - Generative AI has been the buzz word of the month, from ChatGPT to Lensa to MidJourney to Soundful. Now, for generative AI, we just provide them with text prompts or examples of what we want and the computer figures out and delivers us what we want.

The emergence of generative AI will increase the focus oncost optimization more than ever. - Somik Behera

  • Even SBF from the FTX Fiasco hasn’t escaped the “Cloud cost hell”

Source: FTX bankruptcy filings.

This note from FTX Bankruptcy filing showed that Alameda was the wholly owned subsidiary of the now defunct FTX, the crypto version of “enron” of our times.

  • AWS RE:invent 2022 - Quick summary

Summary of ALL strategies that are becoming prevalent within AWS

What AWS wants to provide

Main conflicts that AWS works to resolve

What is Backstage by Spotify - The new UI for all things Developer Infrastructure

It has been close to 20 years since the launch of AWS by a book company caught all of us by surprise. I think this project of courting Developers by a “music company” and monetizing it as a business could be as significant as the launch of AWS. Happy to hear arguments for/against the next developer-led cloud hypothesis in comments.

News you can use

  • ChatGPT - Many use cases to help DevOps on your Day2Day tasks can be accomplished with ChatGPT.
  • OpenBB - Tools you  can use for FinOps teams.
  • Spotify to commercialize Backstage platform - Is it the next “AWS '' moment fueled via developer-led revolution?
  • Skypilot - Optimize your ML Algorithms for cost-effective runtime
  • What do StableDiffusion, GenerativeAI, LargeLanguageModels (LLM) have in common? They are all varying and COSTLY to run models. With the rise of costly to run models, we see an increase in the toolchain and OSS projects that aim to make these models cost efficient. Skypilot does cost optimization by abstracting the models across clouds and running it cost-effectively.
  • Apache TVM - Apache TVM is another tool in the “cost-efficiency” toolchain for folks who want to run ML models cost effectively. Apache TVM takes the approach of leveraging compile time techniques to optimize your ML model for Hardware accelerators  (popularly known as AI chips or FPGAs that can increase cost and perf efficiency by orders of magnitude)

🙏🏼 for reading and being part of the FinKube community,  please share with your friends and colleagues!

Exclusive *Free early access* invitation to CloudNatix Community edition for you! CloudNatix Community Edition is a great free tool that can connect to your Cloud or Kubernetes clusters and quickly provide cost visibility and efficiency recommendations using Autopilot tech pioneered at Google.

If you want this content delivered to your email, please sign-up at www.finkube.io and get learning on pricing, packaging, business cases for DevOps team and how to level up your career.