Enablers for Decision Making, Prompt Engineering, Cheap LLM Training, Even Driven Architectures, How generative AI is changing the way developers work
Hi colleagues,
in this edition of the engineering ecosystem newsletter, we are switching to a curated format that features a selection of different topics. Our plan is to alternate between this format, which provides an overview of several subjects within the engineering ecosystem, and a format that focuses on a single, in-depth topic. In this curated edition, we will be covering a range of subjects:
Enablers for Decision Making: The article by McKinsey explains that decision making can be improved by setting up enablers, team empowerment, tailored decision processes for different decision types, coaching and like more less or more effective meetings. It also suggests eliminating “system noise” and addressing other sources of bias that color your decisions. Empowering employees can also lead to smarter decisions.
Prompt Engineering: In one of the last newsletter on How to learn Prompt Engineering to get better results from ChatGPT, Bing and Dall E I have included some pragmatic resources with examples on how to improve results from ChatGPT or Bing with prompt engineering. Lilian Weng, who leads applied research at OpenAI, has written a more systemic view on the many different forms of prompt engineering (e.g. chain of thought, few shot, self-consistency sampling), which can be of interest for those who want to fine tune models to dedicated purposes. Especially interesting sounds self-consistency sampling allowing to verify the correctness of generated code with Unit Tests, which triggers the question of how this could be used for TDD.
Cheap LLM model by Stanford: According to New Atlas, researchers at Stanford’s Center for Research on Foundation Models (CRFM) have unveiled an artificial intelligence (AI) model that works much like the famous ChatGPT but cost them only $600 to train up. The AI model, called Alpaca, behaves similarly to OpenAI’s text-davinci-003, while being surprisingly small and easy/cheap to reproduce. According to the latest information from the Stanford CRFM website, the Alpaca model is still available for academic research only and any commercial use is prohibited. However, the live demo of Alpaca is suspended until further notice due to feedback from the community and safety concerns.
Event Driven Architectures: Serverless Land is a website that provides information on event-driven architectures, which can help to build resilient, available and scalable applications. There are many patterns for building event-driven architectures. Often you see a mixture of point-to-point messaging, pub/sub, choreography, orchestration, event sourcing and more. The website depicts those patterns in visuals. The site has a strong focus on AWS, but still contains some valuable learning resource.
How generative AI is changing the way developers work: The article How generative AI is changing the way developers work explores how generative AI coding tools, such as GitHub Copilot, are transforming the way developers work by offering code suggestions and functions based on natural language prompts and existing code. The article covers:
The unique value generative AI brings to the developer workflow, such as enabling more creativity and productivity, reducing context switching, and saving mental energy.
How generative AI coding tools are designed and built, using large language models trained on massive amounts of code and natural language data from GitHub and other sources.
Why developers should care about large language models, such as their potential benefits and challenges and best practices for using them responsibly and effectively.
How developers are using generative AI coding tools, such as for brainstorming new ideas, breaking down big tasks, writing tests and documentation, fixing bugs, and learning new skills.
Clean Code
If you want to participate in shaping the future of the Clean Code guide for ABAP. There are 5 Changes (Pull Requests) ready for decision. If there are no valid reasons against the changes, those will be merged tomorrow.
You can find a general introduction, motivation and overview about clean code in the blog Clean Code: Writing maintainable, readable and testable code.
The guide for Git-Based Code Reviews for ABAP is now available as plain Markdown guide leveraging Mermaid for diagrams.
Thanks and Regards,
Klaus
Subscription: If you want to get updates, you can subscribe to the free newsletter:
Mark as not spam: : When you subscribe to the newsletter please do not forget to check your spam / junk folder. Make sure to "mark as not spam" in your email client and move it to your Inbox. Add the publication's Substack email address to your contact list. All posts will be sent from this address: ecosystem4engineering@substack.com.
✉️ Subscribe to the newsletter — if you aren’t already.
❤️ Share it — The engineering ecosystem newsletter lives thanks to word of mouth. Share the article with someone to whom it might be useful! By forwarding the email or sharing it on social media.