AI Assisted Coding and Tools

Real-world experiences testing AI assisted coding and adopting AI tools into software development from Hacker News

Latest Discussions

Product Feedback5d ago

AI coding assistant effectiveness and value

The thread discusses the credibility of AI coding assistants, with recognition of their potential benefits when integrated properly. However, skepticism remains about exaggerated claims of productivity boosts and the subscription cost-effectiveness. The conversation highlights the trade-off between cloud-based AI services and local open-weight models, suggesting the latter might offer practical advantages such as cost control and reliability during critical use.

Join the thread - 2 comments

🔗 Eight years of wanting, three months of building with AI

Product Feedback2d ago

AI-assisted development and tooling choices

The discussion centers around an AI-assisted extension for Hacker News that closely resembles the Refined Hacker News extension in features. The original poster highlights that while the code is AI-generated, it benefits from human oversight by an experienced developer, resulting in a high-quality product. Suggestions include adding features from other extensions like HNRelevant and considering different modern development tools and build setups such as migration from Biome to oxc or Vite. The developer responds by sharing their personal preferences and experiences with these tools and confirms the feature requests are on their radar, emphasizing a balance between AI help and manual refinement.

Join the thread - 2 comments

🔗 Show HN: Orange Juice – Small UX improvements that make HN much easier to read

General Discussion7d ago

Reflections on AI's impact on coding and work

The conversation explores personal feelings about AI's role in software development, highlighting a tension between loving coding as a craft and the rise of AI tools that automate some coding tasks. The original poster feels out of place as peers embrace AI and managerial roles, while they prefer deep technical knowledge and hands-on coding. A respondent reflects on the hopes and fears of AI use in engineering, seeing it as a tool to reduce mundane tasks but worrying about job security and societal consequences. Actionable insights include recognizing diverse coder motivations, addressing fears of AI-driven job displacement, and encouraging discussions about balancing automation with skill development.

Join the thread - 2 comments

🔗 I used AI. It worked. I hated it

General Discussion7d ago

AI-assisted development

The brief discussion questions whether the involvement of AI in building a project matters compared to direct human effort, highlighting a broader conversation about the role and importance of AI assistance in development.

Join the thread - 2 comments

🔗 Mtproto.zig – High-performance Telegram proxy with DPI evasion

Experience Report5d ago

Using AI tools for initial coding and iteration

The primary discussion centers around a realistic and experience-based perspective on AI tools in software development. The original poster shares insights from their project where AI helped rapidly develop an initial, fragile codebase that had to be thrown away and rebuilt, aligning with Fred Brooks' philosophy of 'throw one away.' This approach highlights AI's strength in quick prototyping rather than producing production-ready code immediately. A follow-up comment notes that Brooks later favored iterative refinement, encouraging iterative improvement over discarding early versions entirely. The actionable insight is to leverage AI tools for fast initial implementation to test and learn quickly, followed by deliberate refactoring and refinement for production quality, emphasizing iterative development rather than expecting perfect output from AI initially.

Join the thread - 2 comments

🔗 Eight years of wanting, three months of building with AI

Innovation Discussion5d ago

Impact of AI on Software Development Productivity

The conversation debates the extent to which AI has accelerated software development productivity. One participant claims a tenfold increase in velocity, while another counters that although some productivity gains exist, these improvements do not clearly translate to end-user benefits. The discussion highlights a need to better evaluate and showcase how AI-driven enhancements manifest in real-world outcomes beyond development speed.

Join the thread - 2 comments

🔗 Ask HN: Where are all the disruptive software that AI promised?