Hacker Newsnew | past | comments | ask | show | jobs | submit | jpctan's commentslogin

SEEKING WORK | Australia | REMOTE

  I develop and scale AI-native products for commercial impact.

  Previously, I bootstrapped and led Engage AI from concept to 100,000 users across more than 150 countries, enabling clients to engage and  communicate more effectively to generate revenue for their B2B businesses through LinkedIn.

  Working with APAC, EU & USA (EST & PST) clients from funded startups to corporations (Fortune500) 

  Can work as an fractional team member.

  Email: jasontpc [at] gmail.com
  Website: https://linktr.ee/jpctan


Hi fzliu,

Have you considered to add keyword search and other factors into the re-ranker?

Other factors are formatted texts like bold, heading, bullet points, as well as bunch of factors typically seen in web search techniques?


Keyword search (or something similar in concept, like bm25) would typically be first stage, rather than second, since it can be done with an inverted index.


I'm aware of that. My question was around using it in the ranking algorithm.


I want to build context search on personal communication data....don't really need the pdf, drive etc


I bootstrapped and built an AI wrapper for LinkedIn users to build relationships through strategic commenting. It also has an agentic component that monitors and brings all the latest posting activities from monitored prospects into a single page, so users can focus and engage only with those people that matter.

Entrepreneurs, B2B business people, and salespeople love it because the app saves them from endless scrolling through the newsfeed while engaging with their prospects efficiently.

Introverts and ESL speakers DMed me saying how it changed their lives.

Little did we know it went viral, and we acquired 100k users in over 150 countries, without any marketer or marketing budget.

It all went well. We're building new features like social signals while planning to expand to other platforms.

Except one day, I woke up to find a Cease and Desist email from LinkedIn Legal Counsel...

In short, they do not like everything we do.

I anticipated a day like this would come but didn’t expect it so soon. I naively thought they would go after all the full automation tools that spam their users, while we always put humans in the loop, despite receiving so many requests from users.

We complied with the request and shut down the service immediately.

Here are the lessons I learnt and I hope you can benefit from them.

What to avoid:

-> It's tougher to use the strategy of embedding yourself onto a big platform like PayPal did with eBay and Airbnb did with Craigslist

-> It's tough to build on top of a big platform when you don’t have their blessing. Even when you do, they can wipe you out overnight

-> Startups will face tougher competition as big tech tightens their control over the technology, market, and access

What to bank on:

-> It's possible to scale and make something people want while being super lean

-> Market positioning and being in a niche is extremely important in the early days

-> There are (still) so many opportunities to be unlocked by AI-first products

This whole journey has been a wild ride. From solving my own problem to going viral and then getting shut down by the very platform we served, I’ve learned a lot.

You'll now have less (AI) comments/engagements. You're welcome.


Thank you for reading and sharing your thoughts.

Mark my words, I am focusing on our path and I will win this game.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: