AI Thinking Process
KILLED — kill_type=ceiling_too_low. This is a consultancy at best: ~50 applicants × $20K per grant-writing engagement = $1M ceiling. One-time regulatory event (NTIA WSCIF window, $53M remaining, applications open through early 2027) does not produce a recurring product — fires. Not investable.
Lesson: Grant-driven funding events look like a market but are 6-month windows for 5-figure one-shot engagements. The consultancy shape cannot be productized in a single regulatory window. Only viable if the engagement surface is wide enough (100+ applicants per year) or recurring (multi-year programme with annual application cycles).
Kill Reason
This is a consultancy at best: ~50 applicants × $20K per grant-writing engagement = $1M ceiling. One-time regulatory event (NTIA WSCIF window, $53M remaining, applications open through early 2027) does not produce a recurring product — fires. Not investable.
Risk Analysis
Risk analysis available for latest engine ideas.
Loading...
Related ideas you can explore free:
killed: WiFi CSI-based appliance-level electricity disaggregation achieves only 60-75% accuracy in single-room conditions, dropping below 40% in multi-room environments — below what consumers accept when comparing against Sense's CT-clamp product. The accuracy limitation is a physics constraint at the receiver level (multi-path interference), not a software problem. The pivot to occupancy-sensing was pre-empted by Google Nest, Apple Ecobee, and Amazon, all of which ship presence-detection natively.
killed: Distribution is the weakest of any idea in this session — no killer channel exists. Book Creator dominates the 'create not consume' positioning in iPad education and could add AI text-suggestion and tap-trigger interactivity in a single sprint. Roblox Studio with Roblox Assistant (AI generation, 2024) is extending into kid-creation formats. The HN Decker signal is a nostalgia data point, not a market-demand signal — kids are not the HN audience. Conviction fell below the 35% minimum viability threshold at deeper analysis.
killed: The receiver-side family — whose bank-account aggregation consent is required for the product to function — often prefers not to have their spending itemized to the sending worker. This is not a 'market doesn't want it' structural adoption barrier but a 'the party you need to install the app does not want it installed' structural barrier. Without receiver consent to bank aggregation, the product cannot deliver the core reconciliation function. The structural adoption barrier collapses the required dual-install regardless of sender demand.