1
Is anyone using LLMs to enforce coding best practices?
Ex: A preference for named enums rather than Integers.
Imagined workflow is commenting on a github PR to add another rul or re-run existing rules on the patch.
I’ve built a lot of software in the past the no one has used. Trying to get feedback before writing a line of code.
Does a status update tool sound useful?
It’d capture a screenshot each hour on the hour while you work. Use OCR to convert the screenshot to text. And then use an LLM to convert the text into a daily summary
You could share it on slack, etc. and later on I’d build a slack intervention
Open to any and all feedback
1. Screenshot captured each minute 2. OCR extracts text 3. LLM converts text to a summary of what you were working on at that exact moment 4. LLM Summarizes the summation of the summaries into a daily update 5. Slack integration so you can decide to share the update or not on Slack
Tech stack: MacOS & Apple Silicon Privacy: Screenshots stay on device, OCR text is sent to OpenAI model for summarization Workflow: Install the app, approve Slack integration, set time for daily update, and wait to get your daily summary (as a DM in Slack), you can edit / approve the message
Open to any / all feedback. Thanks! :)