3 ms·
Show HN: We Built an API for AI Calling
- janestevenson80 3y agoJust read through and wondering if you could shed a bit more light onto how you’re decreasing latency? Also would like to learn more about the types of prompting you found to be effective.
- slopedintercpts 3y agoSelf hosting is huge - you don't have to wait in queues like everyone else, and can host in a region close to other servers. Similarly, just getting on the enterprise tiers of the services you use can also guarantee lower latency.
- whatsthenews 3y agoNot to be a de-cel, but what are the safety concerns around this? Can see this being great for pranking/fraud?
- slopedintercpts 3y agoYeah, all the safety concerns are very valid. Easiest immediate fixes that we're building into our system are 1) preventing people from scaling before doing KYC, 2) creating a content moderation filter on the AI's prompts Will do more as we grow.
- LLudic1 3y ago[flagged]