kilocode VS claude-code-proxy

Amp Code is also very good, they released about 2 weeks ago their Oracle feature which leverages o3 to do reviews (https://ampcode.com/news/oracle). Amp leans the closer to Claude Code more than other solutions I’ve seen so far, the team there is really leaning into the agentic approach.

I watch the changes on Kilo Code as well (https://github.com/Kilo-Org/kilocode). Their goal is to merge the best from Cline & Roo Code then sprinkle their own improvements on top.

I've been a software developer for over 8 years and recently started using AI to help me code.

The OSS project I'm currently working on (Kilo Code [1]) basically also attracts the same type of people: People with software dev. experience who like to use AI to help them do things they already know, just faster.

They also don't want to be called "vibe coders". A few days ago, I created a question on a subreddit called "ChatGPTCoding" (which you'd expect to be filled with "vibe coders")[2], asking people if they do "vibe coding" or "AI assisted coding".

Surprisingly, almost NOBODY replies that they do "vibe coding". One guy described vibe coding as "using AI while not having a clue what you're doing."

Am I the only one who wants to be using AI to help me code/design/etc., but I don't want to put under the umbrella of a "vibe coder"?

If yes, is there a better phrase to distinguish ourselves from the "vibe" folks out there?

[1] https://github.com/Kilo-Org/kilocode

I learned about your project from your comment on the Windsurf post but wanted to suggest that you plug your URL anytime you plug the project since it can make the difference between someone "adding it to their eventual list" versus "open in new tab, read, enjoy, engage"

https://github.com/Kilo-Org/kilocode#-kilo-code (Apache 2)

It's open source: https://github.com/Kilo-Org/kilocode, so they are giving back.

Well, one option would be to literally use Claude Code with Gemini. YMMV of course. https://github.com/1rgs/claude-code-proxy

But the other big player in this space is Aider: https://aider.chat/

There is a way to proxy it to other models: https://github.com/1rgs/claude-code-proxy

The big step function here seems to be RL on tool calling.

Claude 3.7/3.5 are the only models that seem to be able to handle "pure agent" usecases well (agent in a loop, not in an agentic workflow scaffold[0]).

OpenAI has made a bet on reasoning models as the core to a purely agentic loop, but it hasn't worked particularly well yet (in my own tests, though folks have hacked a Claude Code workaround[1]).

o3-mini has been better at some technical problems than 3.7/3.5 (particularly refactoring, in my experience), but still struggles with long chains of tool calling.

My hunch is that these models were tuned _with_ OpenAI Codex[2], which is presumably what Anthropic was doing internally with Claude Code on 3.5/3.7

tl;dr - GPT-3 launched with completions (predict the next token), then OpenAI fine-tuned that model on "chat completions" which then led GPT-3.5/GPT-4, and ultimately the success of ChatGPT. This new agent paradigm, requires fine-tuning on the LLM interacting with itself (thinking) and with the outside world (tools), sans any human input.

[0]https://www.anthropic.com/engineering/building-effective-age...

[1]https://github.com/1rgs/claude-code-proxy

[2]https://openai.com/index/openai-codex/