HackQuest Interviews Phil, the Builder of CreatorBid and ForeverMoney (SN98)

Meet Phil, a Hong Kong–based builder behind CreatorBid and ForeverMoney, who’s taking a refreshingly practical approach to Bittensor. In this interview, he breaks down how his subnet works, why simplicity matters, and the key mistakes builders should avoid when launching on Bittensor.

Q: Can you do a quick intro of yourself and the subnet you’re working on?

Phil: Sure. I’m Phil. This year, we’re building a new version of our launchpad using V3 liquidity. We realized we needed something to manage liquidity and maximize fees.

But we didn’t have internal resources to develop strong strategies. So we thought: why not use a subnet? We can incentivize miners to come up with strategies, then deploy the best outputs directly into our liquidity system.

Q: When did you first learn about Bittensor, and what drew you to it?

Phil: Our subnet is number 98. We like to say “98 is money.” What we do is use decentralized intelligence for liquidity management.

I first heard about Bittensor around two years ago. I actually came into the ecosystem as a subnet user. Over time, I got in contact with subnet owners, and then attended the conference in Austin.

Q: If you were introducing your subnet to someone who hasn’t heard about Bittensor, how would you explain it?

Phil: It’s a simple organizational framework, and honestly, it’s not really crypto.

You have the subnet owner, who is like an entrepreneur bringing a challenge. Validators act like managers reviewing KPIs. Miners are like employees solving tasks.

In our case, miners produce liquidity strategies and positions. Validators review them, backtest them against live data, score them based on performance, and then deploy the ranges on-chain.

Q: For miners, what does a “strategy” actually mean? Is it like quant trading?

Phil: It’s related. We focus on concentrated liquidity, like Uniswap V3 pools. Concentrated liquidity is more capital-efficient because you can specify how much liquidity sits within specific price ranges.

For example, you can provide liquidity only between $1 and $1.50. If the price moves too far up or down, liquidity providers go “out of range.” That creates an issue because you end up holding more of one asset than the other, which can lead to impermanent loss.

So miners look at transaction flow, volatility, and market behavior, then structure liquidity positions to reduce impermanent loss and maximize time spent in range to capture fees.

Q: Many teams struggle to balance quality vs quantity when designing a subnet. What tradeoffs do you think are acceptable?

Phil: When designing a subnet, there are two key things to focus on:

1. Don’t overcomplicate it. Build something manageable, launch it, test it, and improve from there.
2. Try to build something that generates revenue early.

Q: Based on your experience, what problems is the Bittensor ecosystem naturally good at solving?

Phil: The Bittensor framework works really well when you want to outsource a challenge.

The cool part is that it creates a platform where market participants can compete to solve that challenge.

Q: When you went live, what worked exactly as you expected — and what turned out more challenging?

Phil: For us, the ideation phase was the most important, and it worked perfectly during the hackathon.

In general, the framework is straightforward. The real challenge is going live and seeing how miners behave — including how they might try to cheat. That’s still a chapter ahead of us.

Q: What advice would you give to someone building their first subnet?

Phil: Start by choosing a real problem that makes sense to solve through market competition. Something where individuals or teams can compete to produce the best output.

Also, make sure the output is easy to monetize. Ideally, the miners’ output directly generates revenue. That way, you can tie emissions to real performance and revenue flows, and even buy back your subnet token to support the ecosystem.

Q: What subnet ideas sound good on paper but might fail once they hit testnet?

Phil: First, you need to find something truly worth being a subnet.

If it’s a problem you can easily solve in-house, then you don’t need market competition; it’s probably not a good subnet use case.

And if you don’t have any clear idea how to monetize the output, it’s also not a good idea.

That said, with AI agents evolving quickly, there’s massive opportunity. If you position your subnet as infrastructure for agents and future AI players, it’s definitely worth exploring.

Q: Thank you so much for your time.

Phil: Awesome, you’re very welcome.