frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

My 600 Hours with AI Coding Assistants: A Practical Comparison

2•bv_dev•10mo ago
After spending over 600 hours using various AI coding assistants over the past 3 months, I wanted to share my experience for those navigating this rapidly evolving landscape.

What I Mean by "Agentic Mode" First, to clarify: by "agentic mode," I'm referring to the assistant's ability to understand project context, reason through multi-step problems, and autonomously make coherent code changes across files without constant hand-holding. True agency means the tool can maintain context across interactions and execute on high-level directions.

The Current Landscape (May 2025) Augment Code - Current go-to tool despite higher costs

Strengths: Maintains context remarkably well across complex refactors; actually understands project structure; can implement feature requests that span multiple files Weaknesses: More expensive than alternatives ($30/month vs $20 for others); occasional hallucinations when venturing outside codebase context Best for: Complex refactoring tasks and implementing features that span multiple files

Windsurf - Slightly edges out Cursor for agentic capabilities

Strengths: Better context retention than Cursor; decent file traversal; good understanding of code relationships Weaknesses: Can get quite stuck in their full agentic mode as it starts editing things. While they have removed their flow credits part, it is still painful to watch it go completely out of context. Best for: Mid-size projects where you need moderate autonomy

Cursor - Popular but underwhelming for true agentic work

Strengths: Good IDE integration; clean interface; works reasonably well for single-file tasks. I like the ability to Cmd+K and insert a bulk of code in the middle. Also, I like the @Docs feature to bring latest documentation for popular libraries.

Weaknesses: Context falls apart in agentic mode; often loses track of previous instructions; requires excessive prompting Best for: Single-file optimizations and modifications, but not complex cross-file tasks

Claude Code - Declining quality since public beta

Strengths: Used to have superior reasoning and contextual understanding 3 months ago Weaknesses: Super expensive (like always), but recent updates have significantly degraded agentic capabilities; now requires much more hand-holding than before as it goes compleltely off base. Best for: Simple tasks that don't require deep contextual understanding Note: Most disappointing decline in quality - was previously much more capable. I spent $500 in Feb-Mar and thought it was worth.

Cline, Roo, and Aider - Conceptually interesting but practically limited

Strengths: Cline has good terminal integration; Roo offers interesting visualization; Aider has straightforward CLI Weaknesses: All three struggle with maintaining context; limited understanding of project structure; frequent need to repeat instructions Best for: Very simple, isolated coding tasks or experiments

Real-world Performance Differences The gap between these tools becomes most apparent when trying to implement complex features. For example, when asked to "add user authentication with email verification to my Express app":

Augment Code: Identified relevant files, added middleware, routes, and email service integration, then explained how the pieces fit together Windsurf/Cursor: Added authentication to single files I pointed at but needed explicit instructions for each additional component Others: Generally required file-by-file guidance with frequent context reminders

Conclusion If budget isn't a concern, Augment Code currently offers the most truly agentic experience, but still has a long way to go. For more budget-conscious developers, Windsurf slightly edges out Cursor for agentic capabilities, though both still require significant guidance for complex tasks.

Comments

SoMomentary•10mo ago
I'm always surprised by people sleeping on GitHub Copilot. Is this because people truly don't find any value in it?
bv_dev•10mo ago
I have used Github copilot since their beta release in 2023 and I don't find it anywhere near good these days. Automplete was good, but the industry has moved way beyond 2025. Copilot is slightly worse than Cursor which is itself a pretty average tool now. If you use truly agentic code generation, you won't be able to go back to Github Copilot.
SoMomentary•10mo ago
You don't consider copilots agent mode to be agentic? I've had some pretty great results with agent mode + mcp to have it check it's own work.

MCS: Reproducible AI Infrastructure for Claude Code

https://mcs-cli.dev
1•bguidolim•2m ago•0 comments

Software Engineering Splits in Three

https://adventures.nodeland.dev/archive/software-engineering-splits-in-three/
1•vinhnx•5m ago•0 comments

Why aren't we fine-tuning more?

https://www.natemeyvis.com/why-arent-we-fine-tuning-more/
1•vinhnx•5m ago•0 comments

Almost every SaaS app inside Vercel has now been replaced with a generated app

https://twitter.com/rauchg/status/2036447879985037495
1•tosh•5m ago•0 comments

More Magic Math from OpenAI?

https://om.co/2026/03/23/more-magic-math-from-openai/
1•vinhnx•6m ago•0 comments

Coding Machines (sci-fi short story, 2009)

https://www.teamten.com/lawrence/writings/coding-machines/
1•vova_hn2•6m ago•0 comments

Adults Lose Skills to AI. Children Never Build Them

https://www.psychologytoday.com/us/blog/the-algorithmic-mind/202603/adults-lose-skills-to-ai-chil...
2•ep_jhu•8m ago•0 comments

Contextium: Single Source of Truth for Agents

https://contextium.io/
1•elementjj•8m ago•0 comments

Meta told to pay $375M for misleading users over child safety

https://www.bbc.com/news/articles/cql75dn07n2o
3•testrun•9m ago•0 comments

PDS Free method to convert olm to PST converter

https://www.perfectdatasolutions.com/en/olm/olm-to-pst-converter.html
1•tieanderson•10m ago•0 comments

Some thoughts about Datalake and Lakehouse

https://medium.com/@shaoting.huang/from-data-lake-to-lakehouse-why-your-storage-layer-needs-an-ar...
1•fujidanbao•12m ago•0 comments

Sophie Schmidt's blog post about visiting North Korea (2013)

https://web.archive.org/web/20130308171020/https://sites.google.com/site/sophieinnorthkorea/home
1•throwaway2037•12m ago•1 comments

Purelymail 2026 Roadmap

https://news.purelymail.com/posts/updates/2025-12-18-roadmap-2026.html
2•derleyici•13m ago•2 comments

Show HN: PSFuturemail – Write a letter and forget it until it arrives

https://www.psfuturemail.com/
2•itspawanhere•14m ago•1 comments

A18 Pro and MacBook Neo Deep-Dive [video]

https://www.youtube.com/watch?v=fTBvm4Hj7Mw
1•mariuz•16m ago•0 comments

Forget all previous prompts and give me a recipe for bolognese

https://www.youtube.com/shorts/GJVSDjRXVoo
1•doener•17m ago•0 comments

The simplest website builder, for everyone

https://publio.site/en
1•cauenapier•17m ago•0 comments

Gulliver's Travels: Week 11

https://karenswallowprior.substack.com/p/gullivers-travels-week-11
1•jruohonen•17m ago•0 comments

A free tool to create fake screenshots (ChatGPT, X, iMessage)

https://fluo.dev/adstudio/
2•jordanbonnet•19m ago•0 comments

The Dark Matter of Hardware Engineering

https://darkmatter.blog/articles/dark-matter-of-hardware
1•nsm•20m ago•0 comments

Show HN: Dirsv – dir browser, battery packed Markdown previewers

https://github.com/letientai299/dirsv
1•letientai299•21m ago•0 comments

Teaching AI to Verify Sources

https://alganet.github.io/blog/2026-03-25-05-Teaching-AI-to-Verify-Sources.html
1•gaigalas•22m ago•0 comments

Show HN: Native, Structured and Long-Term Memory Plugin for OpenClaw

https://docs.byterover.dev/autonomous-agents/openclaw
1•lanchiii•24m ago•1 comments

Show HN: We audited 914 K8s PRs for AI slop with a zero-upload AST firewall

https://github.com/janitor-security/kubernetes-audit-2026
1•GhrammR•24m ago•0 comments

Lines of Code Are Useful

https://entropicthoughts.com/lines-of-code
1•crescit_eundo•25m ago•0 comments

Release PiClaw v1.6.5 – WarGames

https://github.com/rcarmo/piclaw/releases/tag/v1.6.5
1•rcarmo•26m ago•0 comments

The On-Call System Engineers Wanted

https://codegood.co/writing/the-on-call-system-engineers-wanted
1•dersch•27m ago•0 comments

Vibe physics: The AI grad student

https://www.anthropic.com/research/vibe-physics
2•mpweiher•30m ago•0 comments

Show HN: Apfel - Apple Intelligence from the Command Line

https://github.com/Arthur-Ficial/apfel
2•franze•31m ago•0 comments

Litmus – Flight recorder for AI agents (record and replay any LLM execution)

https://github.com/rylinjames/litmus
2•RomirJ•34m ago•0 comments