Best AI Coding Tools? \ stacker news ~devs

pull down to refresh

Best AI Coding Tools?

690 sats \ 35 comments \ @kr 11 Dec devs

Curious to know what AI tools people on here are finding to be helpful as coding assistants/agents.

I know there is a ton of progress being made, and that makes it even harder for me to keep up to date.

What are your favorite AI coding tools? What jobs do they excel at, and where do they fall short?

view all related items

1457 sats \ 1 reply \ @aljaz 12 Dec

v2 of my reply from #348300

i've switched from chatgpt to claude, i still occasionally mix it up with chatgpt, specially for non-coding tasks
i started using perplexity, both for research and occasional data collection for something I'm building
I use bolt.new for building frontends (occasionally I expose backend somewhere publicly so i can have live backend in bolt.new when it develops frontend)
cursor is my main ide now, tho i'll occasionally use copilot instead of their llm integration just to see how it performs

I run a tiny model locally for my terminal llm for quick "i forgot how this bash command should look exactly" lookups, I also have a cloud gpu that runs the same stack but better (slightly) gpu than my local instance so I can have the same setup when i'm away from my battlestation and just use ssh port forwarding to access ollama on my cloud gpu instance from my laptop

I use to have access to A100 so I'd run larger models locally and use open-webui, now i'm experimenting with tiny models like qwen2.5 0.5b or 3b

0 sats \ 0 replies \ @kr OP 12 Dec freebie

interesting, will check out bolt

229 sats \ 13 replies \ @k00b 11 Dec

@bitcoinplebdev cajoled me into using cursor which is excellent. @rblb has been having github copilot give him code reviews. We've also used code rabbit for code reviews. @bitcoinplebdev also recommends v0 which can build UIs.

233 sats \ 0 replies \ @rblb 12 Dec

To add on this, i've tried also cursor (thanks to @k00b), that is nice but I’ve noticed it is somewhat more opinionated about code compared to copilot.

In my experience, copilot feels like it is following you and just completing the code, while cursor feels like it is trying to anticipate what you will want to do several steps ahead. I prefer copilot, but I think it is a matter of personal taste and coding style.

I've been trying Github copilot reviews, that is another beta service, but it doesn't work very well , at least with our code base, but it can catch some oversights, sometimes.

276 sats \ 6 replies \ @ek 11 Dec

Thank you @bitcoinplebdev, I also love cursor now!!

275 sats \ 5 replies \ @satcat 12 Dec

Cursor + Claude is literally all you need. I cancelled all other subscriptions except Cursor (as it includes all top LLMs), and even the infinite slow requests usually only take a couple seconds longer. Use Chat tab for just asking questions or short coding prompts, and use Compose tab for complex prompts when you want agent to automatically edit files, making sure to @tag relevant files, or @codebase, for entire repo.

261 sats \ 4 replies \ @nichro 12 Dec

The way I've been using it:

Inline edit: short edits right where you are in the code and follow up questions about it.

Chat: As you said, to ask more wordy questions and explanations, some coding and follow-up, and use features like @web to search web (and other such tags to add docs into context, etc.) which Composer can't do yet.

Composer: As you said, most complex (and most capable/costly in terms of compute?) for bigger or extensive prompts.

My question to you as you seem experienced with it is for: For Composer, since they released the Agent feature, I've been trying to figure out which option is best for what. Composer with VS without agent. Any tips?

40 sats \ 2 replies \ @satcat 15 Dec

Tried it out over last couple days, seems like it goes of on tangents and little too hard to control code output at the moment. Might need access to reasoning LLM like o1 to be very useful?

222 sats \ 1 reply \ @nichro 15 Dec

I've been trying Normal and Agent back and forth and, while I find it hard to get a definitive read on it so fast, I tend to agree that Agent sometimes tries to think way too far ahead and gets a bit eager to dive in and mess shit up (even in a good way)

On a semi-related note on Cursor: I have a theory, or more of a hunch, that I've been meaning to test:

Using Chat to craft a prompt. Tell it issues, scope, context, documentation, and have it present you with a solution, but without necessarily coding. Maybe pseudocode or steps. Ask follow-up questions, ask why he did X or Y that way, and tweak some stuff ("do it that way, not this way. You forgot to handle X Y").

Iterate until he gives you a game plan and pseudocode that makes sense.

Feed that pseudocode to Composer (Normal or Agent) like "hey this is what we're trying to do and this is the game plan so far". Observe results.

My theory is that because that game plan was generated by AI, the wording and logic is already in "AI speak", with all its quirks and ways of writing, so it will understand what you want to do with more accuracy than typing with all our human-ness.

Note: most of this is bro science coming out of my ass. Would be neat to see if results get better doing things that way.

0 sats \ 0 replies \ @satcat 15 Dec

yeah good call, probably I'm not giving the Agent detailed enough prompts, which it would likely do much better keeping on track and not fucking up the code, which is fine sometimes, like you say, often leading to solutions I would never have thought of. Just have to remember to commit often!

10 sats \ 0 replies \ @satcat 12 Dec

Haven't got the agent feature yet, looking forward to trying it.

0 sats \ 3 replies \ @WeAreAllSatoshi 11 Dec

I’ve heard good things about cursor. One day I’ll give it a shot

111 sats \ 2 replies \ @k00b 11 Dec

If you're using VS Code, there's near zero switching cost. It imports all your extensions and settings.

0 sats \ 1 reply \ @WeAreAllSatoshi 11 Dec

That’s clever. Is it built on top of the vs code open source project, or just has good import support?

100 sats \ 0 replies \ @k00b 11 Dec

It's a fork of VS Code iirc.

0 sats \ 0 replies \ @kr OP 11 Dec freebie

Nice list, will dig into Cursor and v0

98 sats \ 3 replies \ @south_korea_ln 11 Dec

#347574 might be a good starting point

57 sats \ 1 reply \ @kr OP 11 Dec freebie

Awesome, thanks!

It's great that your thread is from exactly one year ago, so I can compare answers and see how the AI landscape has evolved this year.

23 sats \ 0 replies \ @south_korea_ln 11 Dec

Not sure you're aware, but because you're doing the freebie way, i think your posts are collapsed for most users.

132 sats \ 0 replies \ @Bell_curve 12 Dec

One of my favorite posts

I never used Chat GPT or any ai tool before your post

53 sats \ 0 replies \ @TheWildHustle 12 Dec

I'm another cursor fan.

OpenAgents was putting in some work as well.

11 sats \ 1 reply \ @anon 12 Dec

Aider + Claude API + OpenAI API is goated. Only thing you need.

0 sats \ 0 replies \ @kr OP 12 Dec freebie

cool, hadn't heard of Aider before

11 sats \ 0 replies \ @kristapsk 12 Dec

I pay for GitHub Copilot every month, it's useful.

11 sats \ 5 replies \ @Imyourfed 11 Dec

GitHub Copilot, Mistral, ChatGPT (which is obviously the best right now), and llama. If you want to check them (and many more), all out in one spot, give poe.com a shot

0 sats \ 4 replies \ @kr OP 11 Dec freebie

Interesting, I've had a few friends swear by using Claude over ChatGPT lately, what makes you say it's the best?

11 sats \ 0 replies \ @J35TR 25 Dec freebie

switched over to https://zed.dev/ (which introduced me to claude) instead of vscode while trying to get used to lazyvim