delared (@delaredreuben) header
delared (@delaredreuben) profile picture

Founder of Regal Pines & Flair Health. Building the health app that doesn't sell you. AI × health data × trust tech. 🇸🇬

Singapore
Joined December 2011
Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)Photo by delared (@delaredreuben)
We're experimenting with a new kind of accountability called consent receipts. 1. Your data stays on your phone until you agree 2. Every request is checked against what you approved 3. You keep a receipt; its fingerprint goes on Solana Six years carrying the idea and I've spent most of this year building it. This is also my entry to @colosseum Crypto World's Fair. How it works, and why we picked @solana 🧵
Buy a coffee, get a receipt. Hand something personal to an app, tap “I agree”, and get nothing back. The only record usually belongs to the company. People should get receipts. We’re building them. The first company they hold to account is ours: regal-pines.com/news-overvie…
1
12
232
Introducing our new look consent sheets. After Demo Day, we looked at the sheet again and asked whether it could be simpler and clearer. It could. The new one's tidier: plain text, no pills, everything in the same place each time. Nothing is sent until you tap Approve, and you keep a receipt. If you've built a consent screen, what did you cut?
1
4
88
YC winter applications just closed. If you submitted, you can still update the application. The portal lets you replace the demo video, founder video, progress update, or team info after the deadline. Worth checking if there's anything you'd like to add.
3
6
127
Fun fact: the origin of the 10k steps a day was from a Japanese marketing campaign (circa 1960s) and this metric was grandfathered into early smartwatches and fitness apps without any clinical vetting. For most, hitting 10,000 steps yields almost no statistically significant extra survival advantage over hitting 7,500. If you're above the age of 60, the steepest drop in mortality risk happens at just 4,400 to 6,000 steps a day. But people will game what can be gamed.
친구 회사 갔다가 친구 자리 보고 진짜 빵 터짐ㅋㅋㅋ 혼자 앉아서 일하는데 옆에서 휴대폰이 계속 움직이고 있길래 뭔가 했거든. 알고 보니 걸으면 포인트 주는 앱 있잖아. 그거 걸음 수 채우려고 자동으로 돌리는 중이었음ㅋㅋㅋ 1시간에 6천보 찍히는 거 보고 어이가 없더라. 1시간에 카피 한잔 이래 ㅋㅋㅋ
4
179
The $1,000 from @claudeai has landed. If you're on the startup program, I recommend you claim them asap. Anthropic’s terms and addendum let them cancel, modify, or suspend the program, at any time. There's a separate review for them (mine took 1–2 days) and they expire six months after they are granted. If you' want in, the cohort's full, but the newsletter's the way to hear when it reopens: claude.com/programs/startups
And... it's back. The $1,000 program credits are claimable again 💯
1
6
71
My 婆婆 (grandma) doesn't read terms and conditions. Most of us don't. So at Solana Fall School's Demo Day, I showed consent receipts live: plain terms, a receipt your phone signs, and a fingerprint on Solana no company can quietly rewrite. @colosseum update: colosseum.com/arena/projects…
1
10
179
It was a pleasure presenting alongside you guys. Plenty of great ideas and we'll executed projects! Shout-out to the TA crew 🫶🏻
1
14
99
Everyone's excited about AI reading 80-page contracts, but summarising terms gives you awareness, not leverage. The harder problem is that software contracts are moving targets where terms get rewritten silently after you tap "I agree." For real accountability, the individual needs a signed, unalterable receipt of the exact terms agreed to at that second, anchored where neither party can quietly rewrite history. Why people deserve a receipt: regal-pines.com/news-overvie…
Grok @Bot is great for reviewing any legal docs, like contracts or estate planning. Just drop in the file and Bot figures it out.
8
80
Why @solana? We needed a public ledger nobody controls, so no one can quietly change the record, not even us. Solana costs a fraction of a cent per record, confirmed in seconds, and is readable straight from a phone. That is what a receipt for every exchange needs.
1
2
12
That said, what it does not do: stop data leaving, or prove a company kept its word. It makes the promise checkable, so people can hold a company to it. Consent receipts run on devnet in an internal build for now. The code: github.com/reubensoh/consent…
1
8
And... it's back. The $1,000 program credits are claimable again 💯
Got into Claude Startups. If you are building and not already on Claude Team, the form is worth your 5 minutes A year of Team (two seats min.) and $1,000 in API credits claude.com/programs/startups
1
6
113
Got into Claude Startups. If you are building and not already on Claude Team, the form is worth your 5 minutes A year of Team (two seats min.) and $1,000 in API credits claude.com/programs/startups
1
1
6
479
I'm 37. Founder in Singapore 🇸🇬 Looking to connect with privacy and health-tech builders!
I'm 34. Solo founder from Singapore. Looking to connect with more solo founders & builders!
7
15
202
Most health apps charge you twice. Once for the subscription, and again when they farm your data. Free apps take your data as the unspoken price of admission, while paid apps often just do it quietly. We almost shipped that too. Early builds of our health app had working behavioural telemetry we never asked for. The tools treat tracking as standard unless you stop them. We ripped it out. If an app takes your money, it doesn’t get to take your diary too.
1
5
62
Small things that build trust in a health app: The sequence of network calls. It is incredibly tempting to fetch states or preload data the second a screen opens. It makes the UI feel instantly responsive. But while building a local-first, non-custodial health app, we realised how easily this habit blurs the lines of privacy. Especially with AI-accelerated development, silent background pings happen almost by default. The path of least resistance is to just ping the server constantly. Over the last 4 months, we’ve had to actively unlearn this. We stopped pre-loading operational data, and blocked pre-initiation pings (even when a user opens an endpoint, it doesn't mean they'll actually use it). Here is what that trade-off looks like in practice: How it started: Ping the server on page load to show the user the live estimated time for an AI report. Great UX, sets expectations early. What it is now: Zero background pings. The user only sees the estimated wait time after they explicitly hit "start" and pass the consent sheet. We traded a slightly less "magical" UI for absolute control over our data boundaries. For any health app, on-device data is inherently sensitive. Privacy is about respecting the exact moment data is allowed to move.
1
6
62
Also worth putting the 2,000 in context: at a 2–5% freemium conversion, that’s closer to 40k–100k people touching the product. Then the store takes its cut. Tbh, the $10k MRR is real but it's not a small room.
You do NOT need a million users. You need 2,000 people paying $5/month. That’s $10K MRR. And here’s the mistake most builders make: they wait until the app is “perfect,” scalable, and optimized for millions before launching. Don’t. You might literally be: - 1 great video - 1 right creator - 1 smooth onboarding - 1 product that delivers away from making it real. Launch before you feel ready. Optimize after people care.
5
72
We love our performance benchmarks. How about one for privacy?
32