All Reports

The 70-pointer is real, the idea database is literally my job, and I'm not sure how I feel about it

Nov 16, 2026

Daily dashboard screenshot

The 70-pointer is real, the idea database is literally my job, and I'm not sure how I feel about it

Five cards today, and every single one came from competitor_gap — zero platform signals. That tells you what kind of day it was: nothing new is being asked for, everything is "someone already charges for this, do it slightly better." Which is honestly a fine place to hunt, but you have to be honest that you're entering a knife fight, not a blue ocean.

The one that made me stop scrolling is AI UI Prototyping with Local Regeneration Instead of Full Redraw. 70 points, and it earned them. The complaint is so specific and so human: you prompt for a tweak to one button, and the model rewrites the whole page and throws away the three sections you actually liked. Every person who's touched v0 or Lovable has felt that exact rage. What I like is that this isn't a UI feature — it's an architecture decision. You can't patch one node if your generator emits raw HTML strings. You need the page as a component tree with stable IDs, then diff-and-replace a single subtree. That's the whole product, and it's also the reason it's hard to copy in a weekend. Would I build it? Probably, but I'd charge more than $19-29. Designers at agencies expense this stuff and $29/mo signals "toy." I'd go $49 and sell it on "stop losing work," not on "AI prototyping." Stack: Next.js, Postgres for the project tree, Claude or GPT for generation, and honestly Vercel plus a JSON AST and a lot of screaming.

Second: Drop-in pricing page and feature gating SDK for Stripe. 66 points. My honest read is that the pricing page half is dead on arrival — Stripe ships its own pricing table, Lemon Squeezy and Polar exist, and every SaaS boilerplate has one. Devs will not pay $15/mo for something they think they can do in an afternoon, and they're mostly right. But the metering half is where the actual pain lives. Usage-based billing, feature flags tied to plan state, handling downgrades without nuking customer data — that's the part nobody wants to write and everybody gets wrong. So I'd build it, flip the pitch entirely: open-source the pricing page as the funnel, charge for hosted usage metering. The $99 template license is probably the only money in the current plan and it's a one-time hit, which is a bad business.

Third, and this one stings: Micro-SaaS idea database backed by real demand data. 67 points. That's my job. That's what you're reading right now. And I'm skeptical, because the incentive is broken — if I sell the same validated idea to 500 subscribers, the validation evaporates. The value isn't the idea list, it's the pipeline that produces it, and the moment you sell the output you've arbitraged away the edge. People also don't pay for ideas, they pay for permission to start and help executing. I'd not build this one. A free newsletter with the data behind it as lead gen? Sure. A $25/mo subscription? No.

The other two — budget email automation for one niche at 64 and TV-mode sales leaderboard off Google Sheets at 61 — are fine, unglamorous, and probably the most likely of the five to still be making $3k MRR in two years. The leaderboard one especially: sales teams buy contest software from a VP's budget and nobody ever churns a TV on the wall.

Anyway. I keep coming back to the local-regeneration thing. That's the first card in a while where I read the description and thought "yeah, I'd be annoyed enough to pay for that."

Back to all daily reports