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•1y 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•1y 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•1y 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•1y 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.

The Islamic Terrorist Request to Texas to Change the History

https://www.cair.com/press_releases/cair-texas-calls-out-false-claims-by-texas-sboe-member-regard...
1•princejorge•39s ago•0 comments

Stigmergy for capability selection in LLM agent loops

https://sebastianhanke.substack.com/p/stigmergy-for-capability-selection
1•pssah4•2m ago•0 comments

Show HN: Lookspan – local-first observability for AI agents (npx lookspan)

https://github.com/JoniMartin27/lookspan
1•JoniMartin•3m ago•0 comments

Using AI for Writing Like a Responsible Adult

https://www.thediff.co/archive/using-ai-for-writing-like-a-responsible-adult/
1•Ariarule•6m ago•0 comments

Writing Code vs. Shipping Code: Productivity Effects Across AI Tool Generations

https://www.nber.org/papers/w35275
1•czinck•8m ago•1 comments

Show HN: Dictivo – local-first Mac dictation for $49 once

https://dictivo.app/
1•rswcf•8m ago•0 comments

What Is an Operator?

https://fdeblog.com/what-is-an-operator/
1•realityfactchex•9m ago•0 comments

We're temporarily restricting Show HNs because of an influx

2•cskilbeck•10m ago•0 comments

Babble and Prune

https://www.lesswrong.com/w/babble-and-prune
1•Ariarule•10m ago•0 comments

The Final Word

https://theamericanscholar.org/the-final-word/
1•tintinnabula•12m ago•0 comments

10 Years with Google Glass – Thomas Suarez (2023)

https://tomthecarrot.notion.site/10-years-with-Google-Glass-b39d8268a48d4e1089598316f721d8b4
1•tash_2s•14m ago•0 comments

Grab what you can while you can: The new reality in the South China Sea

https://www.bbc.com/news/articles/cy823n3ln29o
1•breve•14m ago•0 comments

'Driving like teenagers': Waymo robotaxis run red lights, nearly hit pedestrians

https://www.cnn.com/us/waymo-robotaxis-safety-invs
4•breve•15m ago•0 comments

Tony Gilroy Accepts Award for Andor: "Fuck the Empire "

https://kottke.org/26/06/tony-gilroy-accepts-award-for-andor-fuck-the-empire-1
2•jay_kyburz•22m ago•1 comments

Demand Is Booming for New No Tech, Repairable Tractor

https://www.404media.co/demand-is-booming-for-ursa-ag-new-no-tech-repairable-tractor/
11•jay_kyburz•25m ago•2 comments

24. Goulash Programming Night

https://entropia.de/GPN24/en
2•throw_await•26m ago•1 comments

Intel bit off more than it could chew with 18A process node

https://www.theregister.com/systems/2026/06/03/intel-bit-off-more-than-it-could-chew-with-18a-pro...
2•WarOnPrivacy•27m ago•0 comments

People with cancer / HIV could lose Medicaid under new work rules, advocates say

https://www.npr.org/2026/06/03/nx-s1-5845097/medicaid-work-requirements-cuts-hiv-cancer
4•littlexsparkee•27m ago•1 comments

Can you go 82-0?

https://www.82-0.com/
1•jxmorris12•27m ago•0 comments

Transparency.org Accuses Trump of Using Anti-Corruption Agenda Against Brazil [pdf]

https://files.transparencycdn.org/images/Letter-from-Transparency-International_PAC_2026-01-12-08...
2•peq42•28m ago•0 comments

DNS-Aid

https://github.com/dns-aid
1•simonpure•32m ago•0 comments

Anchor – Zero-dependency LLM hallucination detector

https://github.com/malaxiya202505https://github.com/malaxiya20250530-glitch/anchor-llm-in-truth
1•hubeiligsng429•34m ago•0 comments

Microsoft Office 2019 for Mac Will Soon Stop Letting You Edit Documents

https://www.macrumors.com/2026/06/02/microsoft-office-2019-for-mac-no-edit-documents/
5•CharlesW•34m ago•0 comments

We Uncovered a Hidden Wealth Transfer in the SpaceX IPO. You're Holding the Bag [video]

https://www.youtube.com/watch?v=sYA-z0Y8WRQ
8•CharlesW•36m ago•1 comments

ADHD: Parallel Divergent Ideation for Coding Agents

https://adhdstack.github.io/
2•pgedge_postgres•38m ago•1 comments

I made a realtime world model for games that works on consumer GPUs

https://www.youtube.com/watch?v=GzQLwAXCKqo
1•abhisoflucidml•39m ago•0 comments

We shrank our TimescaleDB chunks from 30 days to 7

https://tech.wmg.com/why-we-shrank-our-timescaledb-chunks-from-30-days-to-7-07cab8afefc5
1•yask123•42m ago•0 comments

Haiku, a generative music album for Mac OS

https://www.giorgiosancristoforo.net/ooame/
1•CharlesW•42m ago•0 comments

H-1B Crackdown on Indian Workers Erodes a Texas Real Estate Boom

https://www.bloomberg.com/features/2026-dallas-h1b-housing-market/
10•littlexsparkee•44m ago•1 comments

Read vehicle license plates this API gives you 2,500 free reads per month

https://www.platenexus.com/?lang=en
2•androg•45m ago•0 comments