[Dev Catch Up #28] - Gemini Pro surpasses GPT-4o, OpenAI Structured Outputs, Meta AI-Studio, and much more.

Bringing devs up to speed on the latest dev news from the trends including, a bunch of exciting developments and articles.

and

Aug 12, 2024

Welcome to the 28th edition of DevShorts, Dev Catch Up!

I write about developer stories and open source, partly from my work and experience interacting with people all over the globe.

Some recent issues from Dev Catch up:

Join 1000+ developers to hear stories from Open source and technology.

Must Read

The race to build the top and most powerful LLM has just started and the battle among big techs regarding the same is getting heated up. There have been attempts to dethrone OpenAI after their introduction of GPT-4o, which was dominating the leaderboard. But Google has finally left them behind with the testing of its new model in the Gemini Pro family of models. Recently, Google tested their Gemini 1.5 Pro (Experimental 0801) in the arena and it surpassed both GPT-4o and Claude 3.5 in the capability of chatbots. Look at the benchmark results from this short twitter thread posted by the Thread Reader and learn how Gemini dominated the leaderboard in different areas.
If you are a developer and want to build AI-powered applications with the help of OpenAI LLMs, JSON mode as a building block is the best bet. Although this way of developing improves the model reliability for generating valid JSON outputs, the model’s response may not conform to a particular scheme with this model. To tackle this problem, OpenAI came up with the concept of structured outputs. This is a new feature that will ensure model-generated output will exactly match the developer-provided JSON schemas. Learn all about structured outputs from this detailed article published by the OpenAI team, where more details on the feature are available.
Meta is firing shots in the AI race left and right. Alongside introducing the powerful Llama 3.1, they came up with AI Studio that allows users to create, share, and discover AI chatbots. No technical skills will be required to build these chatbots and creators can build them as an extension to their activities. The feature has started rolling out to the Instagram Business accounts in the US and will be available to all Meta users in the coming weeks. This platform can be accessed from the web and is available in all Meta products. Learn more about this platform from this article published by Zdnet, where more detail on the platform has been discussed extensively.

Now, we will head over to some of the news and articles that will be at a place of interest for developers and the tech community out there.

Good to know

Managing cloud costs is highly essential for any business and doing it essentially with the help of monitoring and observability is expensive as the tools available for monitoring charge a good amount of money as subscription fees. To deal with this, organizations must focus on keeping telemetry data that are important to them and discard the rest of the data. This will reduce the cost and will allow the org to stay within the budget. There are several ways of reducing telemetry data such as, identifying the unnecessary labels and metrics, reducing the number of traces, and customizing the histogram views. Learn about them from this article published by Bright Inventions, where the ways to reduce telemetry data are discussed in detail.
Storage is an important aspect of cloud engineering and in the previous issues, we have talked massively on it. Today as well we are focusing on Persistent Storage which is the type of storage commonly used in Kubernetes. As the de facto orchestration platform for modern day applications, Kubernetes is in the forefront of cloud-native engineering. The building block of Kubernetes storage architecture is volume and there are multiple options for the consumption and requisition of the storage resources. Learn about how Kubernetes handles the provisioning of highly available storage from this article published by Dzone, where various tools and strategies are explored through which facilitation of persistent data storage happens in Kubernetes.
Open-Source projects, tools, and technologies have always been a major contributor in the field of software engineering and celebrating them in our weekly is a joy in itself. This week, we are focusing on DiceDB which is a drop-in replacement for Redis with SQL-based real-time activity. It is still a work in progress and supports a subset of redis commands. With that being said, it follows a shared-nothing architecture and comes with a multi-threaded core. With the support of new commands, it enables users to listen to a SQL-Query and get notified in real-time whenever any changes take place. Learn more about DiceDB from its official GitHub repository here and leave a star to support them.

Lastly, we will take a look at some of the trending scoops that hold a special mention for the community.

Notable FYIs

AI is awesome and since there are a lot of developments in this field with the introductions of new LLMs, it is hard to keep a close look how each LLM differs from each other along with the other developments that are taking place. This podcast from Latent Space discusses the shift between Claude 3.5, Llama 3.1, Apple intelligence, and the RAG/Ops expansion into the new battle for LLM OS.
Nowadays, Kubernetes is widely considered as the default container orchestration platform and managing applications in Kubernetes is a task that many like to avoid. Hence, troubleshooting becomes a must-know item, if your application faces any type of problem. Here is a youtube video from Henrik Rexed that discusses various tips and tricks to troubleshoot your Kubernetes application.
If you are into cloud or network engineering, and have loitered around site reliability engineering, then you must have heard about load balancing. It is one of the most important aspects in system design for modern applications and is considered as a device or application that distributes network or traffic among multiple servers. Learn about the different types of load balancing from this short article from ByteByteGo where basics of the concept are discussed in detail.
File permissions in Linux is an important thing to master and what better way to explain it in a single video. This YouTube video from ByteByteGo explains in layman terms the difference in different file permissions in Linux and will be helpful for developers wanting to get started with Linux.
Hackathon is something that every developer enjoys as it brings out the creativity in them and allows them to hone on their skills. Nylas AI is hosting a hackathon that exactly fits this criteria and the developers can show off their skills developing AI solutions. Check it out from the official Dev.to announcement page and learn more about the exciting prizes that you can get.

That’s it from us with this edition. We hope you are going away with a ton of new information. Lastly, share this newsletter with your colleagues and pals if you find it valuable and a subscription to the newsletter will be awesome if you are reading for the first time.

Thank you for reading Dev Shorts. This post is public so feel free to share it.