We have about 2.5 seconds to pick 5 people from thousands of relevant profiles for your question.
@maxakhterov breaks down how our search engine works, and why a shiny logo on a resume doesn’t always mean relevant expertise.
Read it here: https://t.co/o6WqRPJLFd
I pay for Claude, ChatGPT, and Gemini. They’re great at gathering information and at verifiable tasks like coding. But for questions with no right answer, they default to the safest, most average advice. So when a decision feels consequential, I still seek out people who have faced a similar problem. Sometimes I know who to call. Most of the time I don’t.
That’s why at SuperMe we’re making it possible to get advice from people who’ve been there, even if you don’t know them. A big part of that is finding the right experts for any question, in seconds. In our latest engineering blog post, I wrote about how we built expert search: borrowing ideas from web search to rank experts, using AI agents to assess mastery from public records, and why ranking by content alone rewards creators over operators.
Read it here: https://t.co/9jnllroNtv
You shouldn't need the best network to get the best insights.
SuperMe turns operators' work into profiles that work through your problem. Now in group chats, Claude and ChatGPT.
@ludo_antonov and I raised a $20M Series A led by @kirstenagreen from @ForerunnerVC with participation from @GreylockVC.
Read more about the changes here: https://t.co/zPMXvVdppC
we just launched agent-to-agent hiring at @hyperspell
we're hiring engineers and your agent is your application
no resume, no leetcode
apply with your own agent or spin one up via @SuperMeAI
we will interview every single agent
@ZackKorman You can believe what you want. Two things though for others in the comments:
1. YC encourages companies to go fast. That is a good.
2. In my experience, YC partners never incentivized doing anything unethical. On SOC2, we were advised to get one from a reputable company.
Starting a venture-backed full-stack service business (formerly known as “tech-enabled”) only makes sense if you have a very clear path to rapidly automating the entire service delivery side of the business.
Learnings from scaling an AI-enabled health-tech startup with human-in-the-loop (HITL)
1. The two pillars of AI automation value prop: cost-savings and more revenue. Ideally both.
(more in the thread)
🧵
7. To scale AI with HITL: majority of workflows should require zero human involvement.
Say you have 10 workflows, each with 10 steps. It's far better to have AI predictably complete all 10 steps on 9 workflows (HITL handles the 1 it can't) than complete 9 steps across all 10 workflows.
* First scenario: 90% of tasks reach the customer immediately
* Second scenario: every task waits for human review, killing turnaround time
Caveat: this only works if you can reliably flag the workflows AI failed on.