Frontier Watch: monitoring our advancement towards superintelligence

Whidbey Island, WA

An expert-elicited estimate of how far AI has closed the gap to superintelligence, scored by people with no financial stake in the answer, plus plain-language monitoring of what the six major labs actually change.

Project Details

The potentially catastrophic impact of superintelligence upon humanity demands that we develop a trusted instrument, called the Singularity Index, that measures the progress of the world's best frontier labs towards that outcome as a score between 0.0 and 1.0. In addition to myself is my Chief Science Advisor Rachel Grunspan, plus advisors Tony Peng, and Josh Klein. We also monitor how these labs treat the data that they collect from customers in our Privacy Watch dashboard (currently live at https://watch.q16pbc.com). The concrete output of this grant will be a the first full expert elicitation wave + a free public baseline reading.

Theory of Impact

Good decisions about AI risk depend on a calibrated read of how fast capability is actually moving. Today that read is dominated by the actors with the largest financial stake in it: timelines from labs and investors move with funding rounds, not evidence (SoftBank, OpenAI, and xAI have each moved dates or definitions as financing required). Both overstatement and understatement of progress degrade governance.

An independent, incentive-isolated measure improves the epistemic substrate that policymakers, journalists, and the public rely on. It provides a reference reading produced by people with nothing riding on a particular date being believed. Reporting a median and a spread communicates genuine uncertainty rather than false precision, reducing both hype and complacency. And the monitoring side adds a watchdog function: it surfaces when labs weaken safety or data commitments — backsliding that is easy to miss one announcement at a time, and that bears directly on whether safety norms hold under competitive pressure.

I am not an alignment researcher and do not claim to substitute for that work. This improves the shared situational awareness that alignment and governance both depend on — a niche getting little attention from existing funders.

People

Jeffrey Caruso

Team Member

Funding Details

Start Date: May 31, 2026
End Date: -
Expected Duration: -
Funding Raised to Date: -
Annual Budget: -
Monthly Burn Rate: -
Current Runway: -
Funding Goal: -
Funding Stage: -
Fiscal Sponsor: -

Grants Received– no grants recorded

Funding Asks

grantmaking.ai Launch Round

Applied

Minimum

$10,000

Ideal

$50,000

What this ask is

How the money will be spent

At $10k (minimum): Build out the elicitation instrument and recruit the expert pool for the first wave — the two prerequisites for any public reading.At $25k: A complete first wave: at least ten qualified experts, modest honoraria, aggregation, and a free public baseline reading with published methodology.At $50k (ideal): The full version — a larger, deliberately diverse expert pool with competitive honoraria (the main cost driver and the main quality driver), published pool-composition statistics so the sample can be scrutinized, and continued public policy monitoring across all six labs (Anthropic, OpenAI, Google, Meta, xAI, DeepSeek) for the duration.

Funds are ring-fenced to these public deliverables, not the paid product. I commit to public quarterly updates.

Discussion

Private comments in this thread are only shown to approved funders and grant reviewers. Email hi@grantmaking.ai or to request access.

Grant ReviewerPrivate18h?