Build or Skip

59 predictions
we have not settled.

What we later learned about these. Each verdict below was produced by an engine that scored an idea's demand and winnability. In September 2026 we tested that engine against 292 launches whose outcomes we already knew, and could not show it predicted which ones survived. Separately, several of the data sources it relied on had gone dead without us noticing, so some of these were scored on far less evidence than the page implied at the time.

We are leaving all of them up. Deleting predictions once they become inconvenient is the exact habit this site exists to argue against, and a record you can quietly edit is not a record. They stay dated, unchanged, and due for resolution twelve months after each was made — the first in 13 Jul 2027.

The thing we can stand behind is on the register: 6,266 launches and what actually became of them. That needs no engine to be true.

Every verdict published, newest first. None resolved yet.
PublishedIdeaSaid
2026-08-30
Demand is genuine and multi-channel: loud HN/Reddit attention, growing keywords, and four serious players. Win
MAYBE
2026-08-29
The category has genuine, growing pull — strong Reddit and HN engagement for local/open-source AI and a clear
MAYBE
2026-08-28
Demand is real but proven only through incumbent revenue, not through any visible pull: no HN, thin Reddit, fl
MAYBE
2026-08-27
Demand for labeling is commercially proven by five paying competitors, but there is zero organic signal in soc
SKIP
2026-08-26
Demand is genuine and multi-channel: real benchmark-driven discussion, growing exact-match keywords, and four
MAYBE
2026-08-25
Storage demand is proven by four commercial platforms, but demand for a paid tooling layer on top of it is com
SKIP
2026-08-25
Demand is proven commercially but entirely at the top of the market — five funded incumbents plus two YC start
MAYBE
2026-08-24
Real, specific demand exists — a 257-point Ask HN thread asking for exactly this, plus paid incumbents chargin
MAYBE
2026-08-24
Demand is proven commercially but invisible organically: four funded and unfunded players charge for lead data
MAYBE
2026-08-23
Real but quiet interest: growing search direction and clear documentation gaps, undermined by zero Reddit sign
SKIP
2026-08-23
Every channel for the actual topic is empty: zero Reddit, zero on-topic HN threads, no keywords. The only real
SKIP
2026-08-22
Demand is real but entirely inferred from incumbent revenue, not from anyone asking for this specific product
MAYBE
2026-08-22
Demand is real but entirely inferred from competitor revenue, not from any observable user pull — no posts, no
MAYBE
2026-08-22
Demand is inferred from adjacent paying products rather than observed for this specific idea: no Reddit chatte
MAYBE
2026-08-22
Evidence shows genuine developer curiosity about agent memory (10 HN threads, one strong) and five funded prod
MAYBE
2026-08-17
This is a marketplace idea with zero demand signal in any social or search channel, a shrinking trend, and no
SKIP
2026-08-17
This is a real but not roaring B2B utility market: four competitors prove people pay, and every one leaves an
MAYBE
2026-08-16
Demand is real and multi-channel — paying competitors, strong HN interest, and growing commercial keywords. Wi
MAYBE
2026-08-16
Demand for the specific local-first health angle is unproven — every social channel is quiet and keywords are
SKIP
2026-08-15
Demand is proven commercially (4 paying products) but silent socially, putting it in the middle of the range.
MAYBE
2026-08-15
Demand is real and proven by five paying competitors plus growing keywords and active HN interest. Winnability
MAYBE
2026-08-14
Demand is genuinely proven—several products already monetize coding agents and the category is growing—but thi
SKIP
2026-08-14
Demand is real but quiet — validated by paying competitors and a commercial keyword rather than community pull
MAYBE
2026-08-13
Demand for coding agents is loud and commercially proven, but the market is dominated by well-funded giants an
MAYBE
2026-08-13
There's a genuine, defensible open-source wedge in SQL-agent evaluation with real gaps competitors leave open,
MAYBE
2026-08-12
Demand is genuine and multi-channel — real Reddit interest, growing search, and paying competitors prove the c
BUILD
2026-08-12
Demand is loud and multi-channel with proven willingness to use these tools, but the market is crowded with en
MAYBE
2026-08-11
There is genuine, growing developer interest in reducing agent token costs and a real structural gap (no cohes
MAYBE
2026-08-11
Demand is real and multi-channel, and a legitimate product gap exists around unified/edge-specific tooling. Bu
MAYBE
2026-08-10
There is genuine, money-backed demand for vertical SaaS and small teams clearly win here, so the category is r
MAYBE
2026-08-10
Demand is genuine and multi-channel: paying competitors, active HN launches, real Reddit interest, and growing
MAYBE
2026-08-09
This is a real B2B market with four paying/active competitors and a plausible open-source wedge, but the deman
MAYBE
2026-08-09
Demand is real but quiet — proven mainly by paying competitors rather than loud user pull — and the market is
SKIP
2026-08-08
Demand is real but modest—proven by four paying competitors and growing search, though social pull is quiet. W
MAYBE
2026-08-08
Proven commercial demand (4 paying competitors + viral enterprise pain + growing keywords) meets a genuine unf
MAYBE
2026-08-05
The evidence shows genuine, growing B2B developer demand with several paying incumbents and a clear underserve
BUILD
2026-08-05
Demand is real and multi-channel but skews toward infra curiosity rather than proven willingness to pay. Winna
MAYBE
2026-08-04
Demand is proven by four paying competitors and a growing commercial keyword, but consumer social pull is thin
MAYBE
2026-08-04
The evidence adds up to a real but middling opportunity: genuine cross-channel interest and proven paying comp
MAYBE
2026-08-03
Demand is genuine and multi-channel — active refactoring pain, a hot COBOL/mainframe Show HN, growing commerci
MAYBE
2026-08-03
Demand is genuine and multi-channel—growing keywords, engaged threads, and four paying competitors prove peopl
MAYBE
2026-08-02
Demand is genuine and multi-channel, but this is a crowded, capital- and trust-intensive fintech market with f
SKIP
2026-08-02
Demand is genuinely proven — paying competitors, growing commercial keywords, and real social interest all poi
BUILD
2026-08-01
AI security is a hot, growing space with paying incumbents, but demand for the specific 'model weights' sub-ni
SKIP
2026-08-01
Demand is real only as a directional search trend; the actual on-topic social signals are dead (1-2 points) an
SKIP
2026-07-31
The positioning gap is real and the space is uncontested, but 'uncontested' here mostly means 'unproven' — the
SKIP
2026-07-31
Demand is proven and multi-channel, but this is a crowded category with funded incumbents and a heavy engineer
MAYBE
2026-07-30
This is a real, proven market with clear demand but heavy, sophisticated competition including a free entrench
MAYBE
2026-07-30
The topic has genuine, growing interest and a real content gap (interactive runtime selector + reproducible be
MAYBE
2026-07-29
Real, growing demand exists across keywords, competitors, and community chatter, and the market is explicitly
BUILD
2026-07-29
Demand is clearly proven across social, HN, keywords, and paying competitors — this is a real market. But it i
MAYBE
2026-07-28
Demand for open-weights deployment is real and growing, proven by multiple paying incumbents, but the market i
SKIP
2026-07-28
Demand for fine-tuning is commercially proven but the 'cheap' angle is a compute-cost war dominated by funded
SKIP
2026-07-27
This is a real but shallow developer-tooling topic: growing search and genuine confusion, but thin social sign
MAYBE
2026-07-27
Demand is genuine and multi-channel, but this is a classic free-utility category where solo founders with no d
MAYBE
2026-07-26
This is a genuinely hot, growing category with proven commercial demand (multiple YC-funded products) and beat
MAYBE
2026-07-26
Evidence shows real but shallow demand for on-device LLM tooling — a few low-engagement launches, some Reddit
MAYBE
2026-07-13
There is a validated commercial market (5 paying competitors) but weak grassroots demand and a fundamentally h
SKIP
2026-07-13
The evidence shows near-zero validated demand (no Reddit, weak HN traction, flat search) against a field of sh
SKIP

How to read these. Every entry is a prediction made on the date shown, about an idea that was trending at the time. None has come due. When one turns twelve months old we will mark it right or wrong here, including the ones we get wrong — that part of the original promise still stands. What changed is our confidence that the verdict was worth much, and saying so is cheaper than pretending otherwise.