scott_waddell

4 ms·

1. 1.▲
   by scott_waddell 8mo ago

   That timeline feels aggressive but not impossible. The tooling has gotten scary good - I've seen so many people (including myself) prototype complex UIs in hours that would've taken weeks before. Browser engines are a real challen
2. 2.▲
   by scott_waddell 8mo ago

   Academic comparisons rarely capture the messy reality of production use. The "hipper language" advantage tends to fade when you're debugging at 2am with actual users.
3. 3.▲
   by scott_waddell 8mo ago

   The best engineering often comes from that "let's mash things together and see what happens" approach. Most breakthroughs started as someone's weird experiment that made the serious folks uncomfortable.
4. 4.▲
   by scott_waddell 10mo ago

   Hey HN. I built DriftOS because I kept running into the same problem with multi-topic conversations: only a small part of the chat is relevant, but the whole history gets fed into the model anyway. DriftOS routes each message into a semanti
5. 5.▲
   Show HN: DriftOS – Stop dumping chat history into LLM context windows(driftos.dev)

   4 points by scott_waddell 10mo ago|1 comments
6. 6.▲
   Show HN: Driftos-core – Conversation routing for AI apps(github.com)

   1 points by scott_waddell 10mo ago|0 comments
7. 7.▲
   I got Claude and ChatGPT to stop being sycophantic cheerleaders(medium.com)

   4 points by scott_waddell 10mo ago|0 comments