6 ms·
This has also been my biggest gripe with Gemini 2.5 Pro. While it is fantastic at one-shotting major new features, when wanting to make smaller iterative change
by mgw 1y ago
This has also been my biggest gripe with Gemini 2.5 Pro. While it is fantastic at one-shotting major new features, when wanting to make smaller iterative changes, it always does big refactors at the same time. I haven't found a way to change that behavior through changes in my prompts.
Claude 3.7 Sonnet is much more restrained and does smaller changes.
- nolist_policy 1y agoCan't you just commit the relevant parts? The git index is made for this sort of thing.
- tasuki 1y agoIt's not always trivial to find the relevant 5 line change in a diff of 200 lines...
- cryptoz 1y agoThis exact problem is something I’m hoping to fix with a tool that parses the source to AST and then has the LLM write code to modify the AST (which you then run to get your changes) rather than output code directly. I’ve started in a narrow niche of python/flask webapps and constrained to that stack for now, but if you’re interested I’ve just opened it for signups: https://codeplusequalsai.com https://codeplusequalsai.com Would love feedback! Especially if you see promising results in not getting huge refactors out of small change requests! (Edit: I also blogged about how the AST idea works in case you're just that curious: https://codeplusequalsai.com/static/blog/prompting_llms_to_modify_existing_code_using_asts.html https://codeplusequalsai.com/static/blog/prompting_llms_to_m...)
- jtwaleson 1y agoHaving the LLM modify the AST seems like a great idea. Constraining an LLM to only generate valid code would be super interesting too. Hope this works out!
- HenriNext 1y agoInteresting idea. But LLMs are trained on vast amount of "code as text" and tiny fraction of "code as AST"; wouldn't that significantly hurt the result quality?
- cryptoz 1y agoThanks and yeah that is a concern; however I have been getting quite good results from this AST approach, at least for building medium-complexity webapps. On the other hand though, this wasn't always true...the only OpenAI model that really works well is o3 series. Older models do write AST code but fail to do a good job because of the exact issue you mention, I suspect!
- tough 1y agoInteresting, i started playing with ts-morph and neo4j to parse TypeScript codebases. simonw has symbex which could be useful for you for python
- fwip 1y agoReally? I haven't tried Gemini 2.5 yet, but my main complaint with Claude 3.7 is this exact behavior - creating 200+ line diffs when I asked it to fix one function.
- polyaniline 1y agoAsking it explicitly once (not necessarily every new prompt in context) to keep output minimal and strive to do nothing more than it is told works for me.