VQV Signal
According to a post on X by @echen, Surge AI tested coding harness setups on 100 software engineering tasks after reports that Garry Tan spends $200K a year on tokens with a 23-skill Claude Code setup. A setup with 37 skills that spawned 17 agents cost eight times as much as vanilla Claude Code and scored worse. According to the post, the best result came from simply raising reasoning effort.
According to a post on X by @echen, Surge AI tested coding harness setups on 100 software engineering tasks after reports that Garry Tan spends $200K a year on tokens with a 23-skill Claude Code setup. A setup with 37 skills that spawned 17 agents cost eight times as much as vanilla Claude Code and scored worse. According to the post, the best result came from simply raising reasoning effort.
AI-assisted summary based on the listed source.
Surge AI's @echen tested coding harness setups on 100 software engineering tasks after reports that Garry Tan spends $200K a year on tokens with a 23-skill Claude Code setup. A setup with 37 skills that spawned 17 agents cost eight times as much and scored worse than vanilla Claude Code, and the best result came...
Teams paying for complex agent setups may want to check whether extra skills and agents improve results before adding cost. The post suggests that adjusting reasoning effort may matter more than adding structure, but the available text is truncated and does not include the full methodology. Watch for the complete write-up to see the detailed results and how the comparison was run.
VQV organizes public signals from inspectable sources. It does not independently verify the underlying report.
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Scobleizer News Pulse (X).
No login, cookies, social SDKs, or automatic posting.
