AI Assisted Coding and Tools

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

Latest Discussions

Experience Report9d ago

Using AI tools for coding workflows

The thread discusses different AI code generation tools (Codex, Opus, Clawdbot, GLM) and how to combine them effectively in a workflow. Users share their experiences leveraging faster but less precise models like Opus to generate code drafts or plans, then using Codex or other models for review and refinement. An actionable insight is to structure AI coding sessions as multi-step processes, delegating tasks like planning, drafting, and reviewing to distinct models to balance speed and code quality.

Join the thread - 2 comments

🔗 Opus 4.5 is going to change everything

Experience Report2d ago

Sharing AI coding workflows

The thread discusses the challenge of making AI coding efforts reproducible and shareable among programmers. One participant highlights the difficulty in capturing AI-driven coding sessions beyond CLI command outputs, suggesting the need for a structured way to share 'AI instructions' as code. Another participant responds by emphasizing process improvements such as logging sessions, leveraging Git commit histories, and generating documentation through AI summaries to better preserve and communicate what was accomplished. Actionable insights include adopting systematic logging and documentation practices to aid in sharing AI-assisted coding workflows.

Join the thread - 2 comments

🔗 Ask HN: Why are AI coding agents not working for me?

Experience Report8d ago

Using AI for Learning and Productivity

The conversation revolves around using AI as a tool for learning new technologies efficiently during personal time. One participant expresses initial hesitation about coding in downtime, while the other highlights how AI enables hands-on learning by quickly generating prototypes and offering explanations. This method helps integrate new concepts relevant to their work, improving their ability to contribute professionally without traditional time-consuming research. The actionable insight is that leveraging AI for practical learning tasks can transform personal time into productive skill development aligned with career goals.

Join the thread - 2 comments

🔗 Stop Doom Scrolling, Start Doom Coding: Build via the terminal from your phone

Experience Report8d ago

Practical use of AI coding tools

The thread discusses practical experiences with AI coding tools, highlighting that while they may not handle complex or greenfield tasks entirely, they are still valuable for routine, repetitive work. One user emphasizes using AI like 'lego blocks' to assemble common code components, accepting and correcting occasional errors, thus enhancing productivity. Another user counters by stating the frustration engineers feel when tools are overhyped as fully capable, despite their experience showing limitations. The actionable insight is to adopt AI tools for manageable tasks to save time and effort, while maintaining realistic expectations about current capabilities.

Join the thread - 2 comments

🔗 Opus 4.5 is going to change everything

Experience Report1d ago

Gradual AI integration in coding workflows

The thread explores a gradual acceptance and integration of AI tools like Gemini and Opus into coding workflows. One user reports substantial time savings, reducing tasks from hours to minutes with AI assistance, though not yet fully delegating tasks like pull requests. Another user supports this perspective by referencing Andrej Karpathy's podcast, which emphasizes a slow rollout of AI automating manual work incrementally until it becomes indispensable. The insights suggest a practical approach to adopting AI gradually, highlighting the evolving role of engineers alongside AI, rather than replacement.

Join the thread - 2 comments

🔗 Let's be honest, Generative AI isn't going all that well

Experience Report5d ago

AI-assisted coding productivity

Users share their positive experiences using AI (Claude Code) to maintain coding momentum during limited time slots. They highlight how AI helps tackle non-critical, time-consuming tasks, allowing them to contribute more effectively without burdening development teams. This suggests actionable insights for integrating AI coding assistants to improve workflow and productivity.

Join the thread - 2 comments

🔗 Ask HN: Senior engineering mngrs: how has AI changed your day-to-day work?