[Dev Catch Up #37] - LongRAG Framework, OpenAI's Predicted Outputs, Pongo with CRUD, AWS Karpenter, Containerizing Java applications, 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 37th 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 basic retrieval units in the traditional RAG framework are very short which forces the retriever to search over a large corpus to find the “needle” unit. This leads to an imbalanced heavy retriever with sub-optimal performance. Hence, to mitigate the problem, a new framework named LongRAG comes into the picture consisting of a long retriever and long reader. Learn more about it from here.
OpenAI recently added predicted outputs as a new feature in their LLMs with a goal to optimize latency. This feature will reduce significantly latency as it will pass in the existing content as a prediction. Learn more about this new feature from their official documentation.
Live streaming is possible in React with the help of Live Compositor, a media server with its own rendering engine written in Rust. It also provides a JavaScript SDK that allows controlling of the streams with the framework and in return integrates Rust with the JavaScript ecosystem. Know more about live streaming with React from here.
Booting up servers reliably is one of the crucial parts of server management. Better monitoring of the boot-up processes of servers allows better diagnostics in the occurrence of an event during the process. Learn how Cloudflare does the same with their customized BMC firmware from this article.
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.
OSS Highlight of the Week
The OSS project that has raked up quite a bit of stars and getting featured in this issue is Next-forge. It is a Next.js project boilerplate for modern web applications. Check it out from its GitHub page here and leave a star to support it.
Good to know
Amidst all the developments in the space of LLMs, Tencent launched a new LLM from its Hunyuan family named Hunyuan Large. It is an open-source transformer-based MoE model and outperforms Llama 3.1 70B and 405B. Learn more about the model from its Huggingface page.
Pongo is a Node.js tool to provide functionality like MongoDB on top of PostgreSQL. It works well with CRUD implementation style that is suited for content management systems. Here is an article that shows how to bootstrap CRUD with Pongo.
Creating your own authentication system helps understand everything about security for your application and helps prevent potential risks beforehand. This article explains how you can roll your own auth in JavaScript and TypeScript.
Containerization helps pack your application code and dependencies but doing it efficiently with a Java application can pose a lot of challenges. Here is a complete guide on containerizing a Java application with cloud-native buildpacks.
Lastly, we will take a look at some of the trending scoops that hold a special mention for the community.
Notable FYIs
LLM benchmarking is one of the primary ways of comparing the capability of an LLM with its competitors. This podcast discusses how LMSys changed LLM benchmarking forever for good.
State Space Models or SSMs are an alternative to transformers. It processes only the latest input while retaining information from previous inputs. Know more about SSMs from here.
Karpenter is an open-source Kubernetes node management tool developed by AWS. It dynamically provisions the right compute resources to handle cluster resources in real-time. Learn more about this high-performance tool from here.
Control codes in terminals are for everyday use in the day of a developer. This article from Julia Evans discusses all about ASCII control characters.
A pattern is often used to enable React hooks conditionally. There are times when a developer wants to use a react hook based on certain conditions but the rule of hooks doesn’t allow it. This article tells you all about using react hooks conditionally with a pattern.
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.