Video unavailable. Open the source below.
NewPersonalized X Feed Optimizer
A bookmark-informed policy sorts an X feed into Hide, Skim and Keep decisions.
Added to Jevfast
How it uses Jev
Inferred from the creator's description: the optimizer uses 16 saved posts as examples to personalize a decision policy, then evaluates each feed post in one Jev call with nine typed probability answers. Code maps those judgments to HIDE, SKIM, or KEEP. The post names example dimensions such as ragebait, ads, low effort, hostility, usefulness, and whether the user would save a post, but does not show exact question keys or types.
Why it is interesting
The creator describes scoring 150 real feed posts in under ten seconds without another LLM or model training, then testing against 300 previously unseen bookmarks. The post reports text cost as $0.015 while the demo end screen shows $0.007; retain this discrepancy rather than reconciling the figures. These results are creator-reported.