[Dev Catch Up #18] - Gemini 1.5, Gemma LLMs, JSR Package Registry, 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 17th 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
AI is bringing a wave of innovations and big techs are advancing more on their quest to refine their large language models with better optimization for everyday use. We have already heard of Gemini, its outstanding capabilities when compared to GPT from OpenAI, and how it took the tech market by storm. Recently, Google launched a new version of the model, Gemini 1.5 and made it available to developers and enterprise users ahead of a full consumer roll-out. Apparently, this new model is on par with the highest end model i.e. Gemini Ultra. With a context window of 1.5 million tokens which is equivalent to tens of thousands of lines of code, it was created using the Mixture of Experts technique which made the model efficient for Google to run and faster for users. With the same model, Google researchers are currently testing a 10 million context window. Learn more about it from this article from The Verge where they have done a detailed analysis on the model.
Making AI accessible to everyone is one of the visions that big techs are seeing nowadays and steps are taken in that direction. Stack Overflow and Google Cloud made a strategic partnership to bring generative AI to millions of developers using the platform. With this partnership, the developer knowledge sharing platform will add new AI-powered features to its platform which will empower developers to work more efficiently. Google Cloud will integrate Gemini for Google Cloud with Stack Overflow which will help surface important knowledge base information and coding assistance capabilities to developers. This article from Stack Overflow talks about this partnership in detail.
Google has been in the news a lot in the recent days with all the products and developments they are introducing in the world of AI. The company has stepped up and created a new family of AI language models. Gemma, the open-source LLM family of models free open weights models built on technology similar to Gemini. There are two models in the Gemma family, one with 2 billion parameters and the other with 7 billion parameters and both of them can be run locally on a desktop or a laptop computer. According to Google, its 7 billion parameter Gemma model outperforms the 7 billion and 13 billion parameter models of Meta’s Llama 2 family on several benchmarks like maths, python code generation, general knowledge, etc. Know more about this model from this article by ArsTechnica, where they had a detailed discussion on the model and its performance.
Apart from all the AI hype fueling the market, there is also interesting news in the field of JavaScript. The community is buzzing with the arrival of JSR which is a new package registry. It is created by the same team that created Deno, a runtime for JavaScript, TypeScript, and WebAssembly based on the V8 JavaScript engine and Rust. This new package registry is created as an effort to redefine practices within the JavaScript package registry domain. JSR contains a TypeScript-first environment which enables you to write code and deploy directly. It has several significant technical divergences from npm, the standard package registry of the ecosystem. With integrated workspaces alongside seamless npm integration, JSR has usability on the forefront with secure and accessible modules. Learn more about JSR from this article published by Socket, where detailed impression on this new package registry has been discussed with a demo and implementation advantages.
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
While running an ecommerce business that provides a vast collection of products, there is a common pain point that emerges during peak hours. Users experience a slow website while browsing through the categories and as a result, experience the frustration of latency. But this problem is not because of a high user increment or lack of server resources. It is mainly due to the way your application interacts with your postgres database and causes N+1 query problems. With an increase of users navigating the site, there will be a proportional increment in the amount of requests made to the database to fetch information. Instead getting retrieved in grouped queries, each request individually pulls associated data and thus turns a streamlined operation into a burdensome load on the database. For developers and database administrators, understanding the nature of these queries, their impact on the database, and ways of optimizing them is crucial and this article from Readyset talks about that in detail.
OpenTelemetry has turned a new leaf in the world of monitoring and this leads to otelsql, the OpenTelemetry instrumentation for GoSQL. It is an instrumentation library of the database/sql library of the Go programming language. This library generates traces and metrics from the application while interacting with databases. It allows you to identify errors or slowdowns in your SQL queries that can potentially impact the performance of your application. Learn more about otelsql from this article published by the team of OpenTelemetry which explains running the otel-collector from the otelsql repository.
While moving from on-prem to the cloud environment a lot of major shifts appear in an organization. One of them is to monitor the overall health of the system. Typically, the availability, reliability, and performance of a system is measured with the help of Service Level Indicators or SLIs. They are valuable tools for both on-prem and the cloud. So, choosing good SLIs are critical for evaluating the performance, availability, and reliability of systems and organizations face a lot of problems in selecting those specific indicators. Oftentimes with monitoring and SLIs, it is evident that indicators that are performing good for on-prem might not translate well for cloud environments. This article from Bravenewgeek talks about the different ways with which you will be able to choose the right SLIs for your system infrastructure.
New open source projects are always sprouting and in each one of our editions, we celebrate a famed one from the bunch. In this edition, we will talk about Justpath. It is a utility for exploring the path environment variable in both Windows and Linux. It along with its child processes does not have the ability to modify any shell PATH directly but can view the PATH variable. It provides a modified version of the PATH that can be used later in a shell startup script or with an environment manager.With Justpath, you can filter directory names, purge incorrect paths, create new content strings for PATH, dump PATH as JSON, etc. Check out Justpath from its GitHub repository here and leave a star if you like it.
Lastly, we will take a look at some of the trending scoops that hold a special mention for the community.
Notable FYIs
GitOpsCon NA is around the corner and hundreds of DevOps, GitOps, and Cloud-native enthusiasts will join in virtually or flock in for the in-person conference at Seattle. Grab your pass from the official Linux Foundation website and join the biggest GitOps conference of the year.
GitOps has brought a huge technological shift with automating the entire lifecycle of application and infrastructure. It uses Git as the central hub for managing software and infrastructure. Learn more about GitOps as a layman from this short article published by ByteByteGo.
Nowadays, using git is like an unspoken rule for version control. Here is a short article from Martin which discusses the modern git commands and features that a developer should use.
Performance analysis involves monitoring and evaluating the performance of your system or application so that it can deliver desired performance and optimal user experience. Here is a video walkthrough from Kemal Akkoyun’s FOSDEM session in which he talked about python profiling with eBPF for performance analysis.
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.