![](https://crypto4nerd.com/wp-content/uploads/2023/02/1pO2tpJvySfcrGNXIkD7jUQ.png)
A newsletter about DevOps & Machine Learning & Management
Greetings and welcome to N’s Weekly #26, your source for the latest in DevOps, Machine Learning, and Management. In this edition, you’ll find a diverse range of topics, including the latest tools in DevOps such as Dolt, Directus, and Smithy, an insightful article on the true key to success in platform engineering, and a look into the workings of Duolingo’s AI system. We also explore the latest in data science and machine learning, including a comprehensive examination of the challenges involved in running A/B tests, a detailed article for those starting out in large language models, and the debate over bossless management. Additionally, we’ll delve into the importance of learning and understanding different areas in software development through the lens of bounded rationality. Get ready for a journey into the world of tech and management!
[T] dolt — Dolt is a SQL database that you can fork, clone, branch, merge, push and pull just like a Git repository. Connect to Dolt just like any MySQL database to run queries or update the data using SQL commands. Use the command line interface to import CSV files, commit your changes, push them to a remote, or merge your teammate’s changes.
[T] directus —Directus is an instant REST+GraphQL API and intuitive no-code data collaboration app for any SQL database.
[T] smithy — Smithy is a protocol-agnostic interface definition language and set of tools for generating clients, servers, and documentation for any programming language.
[A] No, Platform Engineering Will NOT Do What You Think It Will Do — An article I wrote recently which emphasizes that cultural change is the key to success, not the tools.
[A] How Duolingo’s AI Learns What You Need to Learn — Duolingo is a language-learning app that has taken the world by storm with its gamelike approach and colorful characters. But what users see on the surface is just the tip of the iceberg. Behind the scenes, a sophisticated AI system called Birdbrain is at work, continuously improving the learner’s experience through algorithms based on educational psychology and machine learning.
[A] Getting to decisions faster in A/B tests — part 1: literature review — In this article, the author delves into the challenges involved in running A/B tests and getting to decisions as quickly as possible. They discuss the importance of choosing the right metrics and the limitations of the traditional statistical approach, null hypothesis testing (NHT). The author highlights that NHT does not allow peeking at results early and its concepts can be unintuitive and confusing in a business setting. The author sets out on a journey to understand how the industry tackles these issues and to find alternative approaches that are more intuitive than NHT. The article provides a summary of their findings so far, including the use of sequential hypothesis tests and “always valid p-values” procedures.
[A] Understanding Large Language Models — A Transformative Reading List — Large language models have taken the field of natural language processing by storm and are now making their presence felt in computer vision and computational biology. With transformers, or large language models, becoming the research agenda of many, it’s important for machine learning researchers and practitioners to understand the basics. To help with this, the author has compiled a short and informative reading list for those starting out. The list focuses on academic research papers, and is meant to be read chronologically, beginning with the first paper that introduced the attention mechanism for recurrent neural networks to improve long-range sequence modeling. From there, the list delves into the original transformer architecture, the BERT paper that introduced masked-language modeling, the original GPT paper that introduced the decoder-style architecture, and finally, the BART paper that dealt with denoising sequence-to-sequence pre-training.
[A] Can bossless management work? — In their book “Why Managers Matter: The Perils of the Bossless Company,” the authors examine the concept of bossless organizations and management. The authors evaluate the various theories and approaches to decentralized management, such as holacracy and agile, and come to the conclusion that bosses do indeed matter in an organization. Through a comprehensive examination of the bossless company narrative, the authors show that hierarchical management structures are necessary, while also acknowledging the need for a redefinition of the traditional management role in today’s fast-paced and rapidly changing environment.
[A] Bounded Rationality in Software Development: The Importance of Learning and Understanding Different Areas — This is a recent article I wrote about the “T-shaped skills, systems thinking, and lifelong learning in navigating complex and dynamic problems”.