Product

RadKits beta is officially live. My co-founder posted about it on LinkedIn yesteday. We’re ready for radiologists beyond our alpha testers to try RadKits—a report assistant that cuts down daily toil with high quality templates, dictation support, and a curated knowledge base. I’m nervous that LinkedIn won’t be enough to reach the right radiologists. My co-founder has the network, but I’m wondering if we need to be doing more. I’m excited to see if beta users adopt it the way our alpha testers did—if it becomes a standard part of their workflow, something they’d feel pain without. That’s the real validation. ...

Update (January 18, 2026): Well, this is embarrassing. After publishing this post and opening a GitHub issue, I discovered the feature DOES exist—I just completely missed it. Where it actually is: When setting up survey targeting with a date property, use the drop down to change the operator to “before” or “after” then click “Select a value” and you’ll see relative date options. PostHog Survey relative datetime It’s right there in the dropdown. I have no excuse except that I was tunnel-visioned on looking for it elsewhere. ...

I spent 4 hours rewriting RadKits’ landing page before my cofounder’s conference. My first draft positioned us as a workflow optimization tool that would “standardize” radiology templates and “increase quality”—language that would have alienated our actual users. My cofounder (who is a radiologist): “This sounds like you’re taking away their autonomy and disparaging their work. They’ll hate that.” He was right. Here’s what I learned about positioning a product for two audiences. ...

My cousin is a radiologist. He built a quick prototype to automate his radiology reports. His colleagues wanted it. That’s when he recruited me: “Can you help turn this into a real product?” This is RadKits. Here’s what I’ve learned building healthcare SaaS with a domain expert co-founder. How RadKits Started Raid had a problem: writing high-quality radiology reports is detailed, time-consuming work. The templates exist, but they take forever to modify for each case. ...

I had multiple frustrations with ChatGPT and Claude: No good way to save and reuse prompts $20/month per service for occasional use Chat history scattered across services No way to compare model responses side-by-side No single pain was unbearable, but together they pushed me to build my own setup. Why I Care About This I experiment with different models—open source and proprietary—to understand their strengths and weaknesses. But this experimentation has a high mental and monetary cost. ...