Priority 0 (proposed in 1/21/2024)
- P0 - Improve label and transaction coverage
- discuss with partners to learn their contribution plan and any improvement this year
- reach out to more partners
- brainstorm together what sources (social media, websites, open source, etc) we can scrape more labels
- Metrics - COVERAGE
- P0 - Attestation
- SDK work will unlock many other high-priority goals
- Private labels
- Metrics - Correctness, 100% onboarding
- P1 - Further Label Transparency
- Every partner needs more info. When returning API and saying a label’s source is heuristics, make one more step to provide a link so the user can click on it to learn more about it. The link can be an academic paper about these heuristics, an analysis in Git Hub, a corporate webpage, or anything that describes the heuristics and to what extent users can learn more about it. Such link column also applies to other options such as
ground truth, making it clear how the ground truth work is operated, or 3P to specify which platform and how the label is scraped.
- Metrics - lineage coverage reaches to XX% of new labels
- P1 -Label Governance
- Create a mechanism so partners can contribute
blurry labels
- need further discussion with the partner to agree on the mechanism
- Metrics - adoption
Can define the scopes of labels that a user can see in at least two dimensions
dimension 1: whether you can see fresh labels or 3 monthly old ones
dimension 2: whether can see the CDA or other high risk sensitive label
- Discussion:
- Should a microscope be the custodian for private labels?
- Should we have public service SLA
- DRIs
- Ming to drive 1, 3 and 4 and delegate as needed
- Zhanwu to drive 2
Priority 1