Build Python LLM apps in minutes

Curating the ocean of AI news keeping you informed on topics key to the craft of software development

AI tools that help you build software

Command Your Terminal Like Never Before with Warp AI
Warp AI is an AI-powered assistant that's integrated within the Warp terminal. It can help you author, debug, and understand programs by converting natural language descriptions into shell commands, predicting commands, and providing code completion.

A Declarative Way to Work with GPT
I came across the gpt-json library, a Python wrapper for GPT which caught my eye for its declarative approach to defining the expected output format. This allows you to get structured and type-hinted responses from GPT models. This can be helpful for tasks such as generating code, writing documentation, and creating data pipelines.

Another code testing suite with AI
Sapient is another contender in the AI augmented test suite camp (Codium being another example), Recently securing 5M in funding, and currently available as a Jetbrains plugin. The platform uncovers design flaws, generates and maintains test code, and offers comprehensive test coverage.

Code review with a large language model
Lintrule is a tool that uses a large language model to review your code. It can enforce policies your linter can't, find bugs your tests may have missed. It first analyzing your code to understand its structure and intent. It then uses this information to generate a set of rules that can be used to enforce policies such as style guides, coding standards, and security best practices.

A Visual Programming Environment for LLM Response Analysis

ChainForge is a visual programming environment for analyzing and evaluating LLM responses. It is geared towards early-stage, quick-and-dirty exploration of prompts and response quality that goes beyond ad-hoc chatting with individual LLMs.

CodeSandbox releases an AI coding assistant
CodeSandbox (a code in the cloud IDE) has released Boxy an AI coding assistant that can help you write better code by providing contextual explanations, generating code, and refactoring code. It is still under development, but it has the potential to be a valuable tool for developers of all levels.

Utilizing AI in your software

All the Hard Stuff Nobody Talks About when Building Products with LLMs
In this article the development of the Honeycomb natural language querying interface is used to discuss the challenges of building products with LLMs. The author argues that LLMs are not products themselves, but rather engines for features. This means that developers need to be careful about how they use LLMs, and that they need to be aware of the potential risks.

Build Python LLM apps in minutes
Chainlit is an open-source Python package designed to expedite the process of building and sharing LLM (Language Learning Model) applications. It can be effortlessly integrated into existing codebases, allowing users to create a chat interface similar to ChatGPT in a matter of minutes.

Supabase contributes to OpenAI repo with a Postgres and a Supabase implementation to help developers build plugins
ChatGPT plugins now support Postgres and Supabase. This means that developers can use ChatGPT plugins to access data stored in a Postgres database. This integration makes it easier for developers to build ChatGPT applications that can access and process data from a variety of sources.

Latest news

Generative AI: The Future of Coding
This article discusses the rapid pace at which Generative AI is automating mundane coding tasks, with tools like Microsoft's Copilot. The long-term impact of generative AI on coding is still being felt, but it has the potential to revolutionize the way developers work.

Ask HN: Those with success using GPT-4 for programming – what are you doing?
There are some interesting insights and use cases in this thread. There are a lot of accounts of substantial time savings, balanced with folks describing the challenges they are having.

AI-Powered Code Generation: GPT-4 Takes a Crack at Writing a VSCode Extension
In this article, Kevin Lin provides a thorough and balanced discussion of his experience using GPT-4 to automate the creation of a VSCode extension. While GPT-4 produced functioning code, the process wasn't seamless. Even with GPT-4's impressive capabilities, certain issues like package dependencies and testing still required human intervention. Lin does see potential in this process though, he's since improved the prompts and context available to ChatGPT which substantially reduced the error count in producing the VSCode extension.

Low/No code

A No-Code Platform for Creating Backend APIs
Backengine is a no-code platform that allows you to create and deploy backend APIs powered by large language models. With Backengine, you can create APIs without writing any code or managing any infrastructure. Backengine is currently in beta and is available for free.

Tutorials

How to Use LlamaIndex to Connect Your LLMs with External Data
This tutorial will teach you how to use LlamaIndex, a tool that helps you connect your LLMs with external data, so you can ask questions about your documents like you would with ChatGPT.

Building an AWS Well-Architected Chatbot with LangChain
This tutorial will teach you how to build an AWS Well-Architected chatbot leveraging LangChain, the OpenAI GPT model, and Streamlit.

Read/Listen/Watch Learn

Learn while you commute, do chores, exercise…

What is ChatGPT Doing … and Why Does It Work?
I found this article to be very interesting, as it provides a glimpse into the inner workings of large language models. In it, Stephen Wolfram covers all the topics involved in the operation of modern LLMs and tools such as ChatGPT. The discussion is well supported with diagrams and examples.

 Thanks for reading and we will see you next time

Follow me on twitter, DM me links you would like included in a future newsletters.