Phodal | 全栈工程师

Original link: http://www.phodal.com/blog/llm-at-core-for-developer/ In the past few months, our series of explorations on large language models, such as ChatGPT end-to-end practice and application development, LLaMA and ChatGLM fine-tuning experiments, GitHub Copilot reverse engineering analysis, and dynamic context engineering (ie LangChain), have driven Let me write a summary, which is also a programmer’s guide to LLM. …

With LLM as the core LLM@Core: Programmer’s guide to large language model technology Read More »

Original link: http://www.phodal.com/blog/llm-friendly-api/ In the past few months, we have been exploring: how to use large language models to improve R&D efficiency? With the gradual advancement of our research, we slowly entered some deep water areas, such as: how to better use AI to assist architecture design? Based on our many years of experience in …

Big Language Model Friendly API: Building Better Software Architectures with Collective Intelligence Read More »

Original link: http://www.phodal.com/blog/unit-runtime/ After we built the Unit Mesh architecture and the corresponding demo, we started to implement the Unit Mesh architecture. So, we continued to start Unit Runtime, which is used to directly run the code generated by AI. This article is transferred from: http://www.phodal.com/blog/unit-runtime/ This site is only for collection, and the copyright …

Unit Runtime: Run the code generated by AI with one click, maybe it will become your copy + paste artifact Read More »

Original link: http://www.phodal.com/blog/prompt-as-code/ > Prompt as code is a programming paradigm based on multiple input modes. It provides richer contextual information by combining text, image, voice and other input methods to help programmers better express their intentions and Generate the corresponding code implementation. Prompt as Code uses prompts as part of the code and as …

Prompt as Code: Best Practices for Designing and Managing AI Programming Read More »

Original link: http://www.phodal.com/blog/unit-mesh/ > Unit Mesh is a distributed architecture based on artificial intelligence. Different from the traditional distributed architecture, the service unit (Unit) in Unit Mesh is generated by AI. The services and data in the application are abstracted into independent units, and are managed and deployed through a unified control plane. This article …

Asymptotic AI programming mode: design ideas and exploration of Unit Mesh architecture Read More »

Original link: http://www.phodal.com/blog/chatflow-opensource-ai-workflow/ **Too long to read version**: This article is transferred from: http://www.phodal.com/blog/chatflow-opensource-ai-workflow/ This site is only for collection, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/thinking-in-ai-programming/ For the past month, I have been researching ChatGPT and talking to many people (regular people, programmers, startups, VCs, etc.). Whether it is in the industry (programming) or outside the industry, people’s views on AI programming are: the future is promising-**GPT, he seems to be able to do everything, and he seems …

AI programming in the future: Is it the ultimate liberation of programmers or the beginning of unemployment? Read More »

Original link: http://www.phodal.com/blog/dsl-ide-support-with-intellij-idea/ In the past few months, when designing and developing [Fklang](https://ift.tt/hy0JYom), corresponding IDE support has also been added. Due to the complexity of the development of the Intellij IDEA plug-in, it cannot be described in detail in the document, and can only be tried through code. Writing an article to record this painful …

IDE Support for Domain Specific Languages: The Basics Read More »

Original link: http://www.phodal.com/blog/explore-low-latency-architecture/ A few years ago, my technical consulting experience in the communication field gave me a preliminary understanding of how obvious the performance improvement of **pre-allocated memory management** is. Recently, I have also seen the low-latency architecture technology required for high-frequency transactions (limited by the unique national conditions in China) from the bits …

A Preliminary Study of Low Latency Architecture Read More »

Original link: http://www.phodal.com/blog/ddd-for-ddd/ > “All models are wrong, but some are useful.” – George Box This article is reprinted from: http://www.phodal.com/blog/ddd-for-ddd/ This site is for inclusion only, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/collaboration-with-crdt/ Compared with conventional online visual collaboration, for graph-as-code drawing tools like Feakin, online collaboration can be directly simplified into three elements: This article is reprinted from: http://www.phodal.com/blog/collaboration-with-crdt/ This site is for inclusion only, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/design-online-diagram-tools/ In high school, I read several books related to 3D graphics programming. Anyway, since then, anything related to graphics has been out of my sphere of interest. Until recently, I’ve rekindled a little interest: This article is reproduced from: http://www.phodal.com/blog/design-online-diagram-tools/ This site is for inclusion only, and the copyright belongs to the …

How to Build an Online Drawing Tool: How was Feakin designed and built? Read More »

Original link: http://www.phodal.com/blog/android-architecture-governance-start-from-dependencies-analysis/ The recent project is relatively busy, and there is not much spare time to spare. During the week, under the “coincidence”, I talked with the architects of a well-known domestic mobile phone manufacturer about how to manage the Android architecture, and the starting point was: how to implement the Android architecture from …

A Preliminary Exploration of Mobile Application Architecture Governance: Starting from Dependency Analysis and the Life Cycle of Android Applications Read More »