Everything AI?

Is there any company out there not doing AI? Is there any Keynote that does not involve talking AI revolution? How many raw articles out there since we have the whole LLM ecosystem? The overall advancements on the whole is interesting, helpful but at times I feel its too much and there are other areas too that needs to be highlighted.

Anyways, since AI is changing the way we live and interact, its getting everywhere in our day to day life. With the new OpenAI and Google I/O announcements in the past week it’s clear that these companies are just accelerating in the AI advancements.

I feel we definitely need to keep up with the stuff that is coming up, adapt to the things that are useful for our productivity and ignore the junk(yes tons of junk in AI also out there).

OpenAI announced and showed cool demos → their latest model gpt4-0 which also will be available for free to use, live translator, desktop app and how you can screen share + ask stuff from chatGPT. Some great use-cases I have seen over socials- specially the “help me solve the maths problem” was really dope. I think there will be a lot of use cases specially in the field of study, medical sciences, aircrafts etc.

Google I/O, which was around same time also showcased advanced integrations of Gemini across their wide range of products making it clear that Google is also not stopping to innovate in the AI ecosystem and are trying to make it big. They showcased some cool demo’s and announcements:

• How with Gemini and Gmail integrations you can catchup over email after vacation
• Search across photos to get your child’s growth
• Summary of your google meet meetings as bullet points
• 2million tokens context window
• Most interesting demo was project Astra where they showcased via camera how you can ask and get stuff answered.
• Launched Gemini 1.5 flash which is lightweight, fast and cost efficient model
• Coming to generating videos they gave a demo of Veo where it get the instructions and create 1080p videos
• 6th Generation TPU’s called trillium
• The next iteration of Google search powered by Gemini will be interesting to see and I am most interested to see how the SEO will work in this AI era.
• You can create gems which is similar to creating custom GPT’s
• PaliGemma - Vision language open model and Gemma 2 is on the way
• I am more interested in trying the LearnML: making educational videos more interactive and its integration with YouTube

It’s 2024, can AI book my entire vacation yet? What do you think? I have not seen it happening and have to go through the process individually to get best deals etc. Anyone used AI for their bookings?

While I am excited about what is coming next, for the people in infrastructure space needs to keep innovating to get these models running as efficiently as possible and bridging the gap between infra and ML engineers. If you are in this domain, I think you still need to stay updated in what is coming and what is happening in the generative AI space and how you can deploy local LLM’s to enable your engineers and developers at workplace. I sat with John Mcbridge to record a video on one of their production implementation of running vLLM on Kubernetes to server LLM’s.

What are your thoughts on LLM’s and running them locally? how do you run them? what are use-cases you are using it for?

I have been busy creating more content and also focussing more on editing as well, hope its working out, do let me know other suggestions if there are any.

Two latest ones are on Network policy and WASM on Kubernetes using spinkube.

Now, let’s dive into the awesome stuff by cloud native community.

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 - How Automation Reduces Large Language Model Costs?
• Sysdig - LLMjacking: Stolen Cloud Credentials Used in New AI Attack

Awesome Reads

• How LLMs Work, Explained Without Math - Want to understand LLMs in simple terms without advanced math, this article explains how they predict the next token in a sequence to generate text and the underlying mechanics involving tokenization, predictions, and the training process that makes them seem intelligent.
• Serverless WebAssembly: Security, Speed … and Startup Times - This post highlights the benefits of using WebAssembly for server-side applications, emphasizing its small package sizes, rapid startup times, and enhanced security. He discussed the efficiency of WebAssembly in serverless environments compared to Kubernetes, VMs, and containers, and addressed the current challenges and future potential of porting Rust applications to WebAssembly.
• Wasm vs. Docker: Performant, Secure, and Versatile Containers - Using Docker and Wasm together lets developers create reliable, secure, and fast applications. Docker helps manage containers easily, while Wasm runs code quickly and safely. This combination ensures smooth development, easy deployment, and strong security, making it great for modern software projects.
• Announcing Flux 2.3 GA - This post highlights some of the new features and improvements included in this release.
• Best Practices for Generating High-Quality SBOMs - To generate high-quality SBOM’s use the tools integrated into your dev lifecycle, and ensure the SBOM includes all necessary data fields for your use cases, such as security and license compliance. Keep SBOMs updated with each production deployment and distribute them appropriately, either publicly or privately, based on the nature of the software.

• Awesome Repos/Resources

• multicluster-observability-operator: Operator for Multi-Cluster Monitoring with Thanos.
• WrenAI : WrenAI makes your database RAG-ready. Implement Text-to-SQL more accurately and securely.
• Cloud native AI landscape

Learn from X platform

Do comment your thoughts and share this in your network, also care to subscribe for free.