Shengkun Ye's people search runs at $0.0024 a lead
A second open source people search on Jev landed, this one quoting $0.0024 per lead by waterfalling six data providers and paying for enrichment only on the leads Jev keeps. Mike Taylor, separately, says short-lived skills are an eval failure, not a skill problem.
Built with it
Shengkun Ye's open source people search quotes $0.0024 a lead
Shengkun Ye released a Jev-powered people search that waterfalls across Apollo, Clay, People Data Labs, Hunter, ContactOut and Ploid, runs 1,000 leads through Jev, and only pays for enrichment on the leads Jev kept. He puts it at $0.0024 per lead, open source with no markup, against the $2,000 subscriptions it is meant to replace.
Pay for enrichment only on the leads Jev kept
Writing criteria
Allie's line-by-line search pattern: nouls for one answer, scores for ranking
Allie the Icon described a pattern she calls line-by-line search: put the whole document in state and ask the question once per line or paragraph. Use nouls when you expect one dominant answer, use scores when you want to highlight lines by overall relevance. Diogo Almeida has been making visual explainers of the parallelism it relies on.
Use nouls when you expect one dominate answer, use scores when you want to highlight lines by overall relevance.
The case against
Mike Taylor says a short skill half-life is an eval failure
Responding to a summary of Felix from Cowork warning that skills have a short half-life and need updating with each model release, Mike Taylor called that a failure to have proper evals.
If you have good evals you don't even need to look at your skills and can recompile them on your tasks every time a new model comes out