Reddit thread: judging agents by deliverables, hunting a Manus alternative
Nu_Exia299 · reddit · 2026-11-05
A Reddit user argues "task completed" is a generous description of what AI agents deliver: an agent researching 10 competitors and building a report, deck and landing page will declare DONE while half the citations need checking, the deck drops competitors, and the page advertises unapproved features.
They survey Manus (broad agent execution), Genspark (research + slides), Claude (reasoning), Gamma (presentation polish) and Runable (business deliverables), and propose a boring acceptance checklist — files exist, sources check out, outputs are editable, nothing missing, consistent facts — as a better lens than demo videos.
More from coding & agent
- Build your own AI agent harness: memory, permission gates, and subagents — PawelHuryn · 2026-11-05
- Six agent harnesses compared: 120 lines vs 1.4M lines, feature by feature — PawelHuryn · 2026-11-05
- How agent harness context windows work: append-only for caching and reasoning — PawelHuryn · 2026-11-05
- How agent harness tools work: five local tools plus server-side web search — PawelHuryn · 2026-11-05
- Dev builds 1,200-line Mini Harness that beats Claude Code on Terminal-Bench — PawelHuryn · 2026-11-05
- MCP vs API: the integration-math answer that survives AI engineering interviews — Ok-News-7546 · 2026-11-05