[Dev Catch Up #34] - OpenAI's Realtime API, Deep-learning for ETA, Whisper large-v3 turbo, Spring Boot App with S3, and much more.

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

Welcome to the 34th 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:

• Ramp up with these Cursor resources, Reflection 70B

• Gemini Pro surpasses GPT-4o, Meta AI-Studio

• Meta unveils Llama 3.1, Open AI's GPT-4o mini

• LLM training at Meta, iOS 18 beta release, AI security with PCC

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

Must Read

• OpenAI DevDay recently happened and there’s a bunch of API updates that came on the development front. One such announcement is the introduction of the Realtime API which allows development of applications with advanced voice-mode like that of Chat GPT. Read more about this new API here. 

• Estimated time of arrival, or ETA is a critical component of operational efficiency and customer satisfaction and hence, the challenge of providing accurate ETAs is both complex and essential. Read how Doordash uses deep learning for predicting ETAs smartly from this article published by the engineering team. 

• Key-value databases come with a lot of challenges for datastore misuse and developers in Netflix have experienced them first-hand. To mitigate these problems, the developers of Netflix developed a holistic approach that led to the creation of a Key-Value data abstraction layer which simplifies data access and enhances the reliability of infrastructure. Learn more about this from the official Netflix tech blog regarding the same.

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

• EnVision Research recently launched Lotus, a diffusion-based visual foundation model for dense geometry prediction. Learn more about Lotus from its GitHub page here and try it out from this Huggingface space. 

• OpenAI came out with another model from their Whisper family named Whisper large-v3 turbo. This is a finetuned version of the Whisper large-v3 model and is way faster because of the reduction in the number of decoding layers from 32 to 4. Learn more about this open-source model from its Huggingface page.

• If you are using AWS as a developer, then connecting to S3 services and performing operations on buckets and objects is a core skill that you must possess. Here is a tutorial that shows how to connect a Spring Boot Kotlin application with AWS S3 Object Storage and make use of the S3 client.   

• Diffusion world models have become a dominant approach for image generation with an approach to train reinforcement learning agents in a safe and sample-efficient manner. Diamond or DIffusion As a Model Of eNvironment Dreams is a reinforcement learning agent trained entirely in a diffusion world model. Learn more about it here.

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

Notable FYIs

• OpenAI’s recently concluded DevDay sparked everyone’s interests with a lot of new announcements and releases. But the two things that caught everybody’s attention are the demo of project strawberry along with the discussion on building AGI in real time. Here is a Latent Space podcast which covers every notable thing from DevDay 2024.

• Thinking of backing up your files from any operating system? Then, Restic should be the choice of modern program to use. It can backup your files from any operating system to different storage types. Learn more about Restic from here.  

• There are different types of React components that serve as the building blocks of the UI. Here is an article that provides a complete understanding of modern React components and patterns.    

• OpenAI released Swarm, an agent framework that focuses on making agent coordination and execution lightweight, highly controllable, and easily testable. Check out Swarm from its official GitHub page.

• Quantization lowers the parameters of the model to make it smaller and faster while maintaining accuracy. Since Vector databases are widely used in developing LLMs, three types of quantization techniques are widely used around them and this article provides a detailed picture on it.

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.