Question
How far does understanding lag in funding?
Does support for research into understanding match the scale of the capabilities we are creating? The initial case sets private investment in AI against a single consciousness research programme. It shows a difference in magnitude, but does not yet compare the aggregate budgets of the two fields.
0.70 confident · editorial estimate
What we know
- The AI Index 2026 gives 344.66 billion dollars of global private investment in AI for 2025. That is a figure for capital raised, not the total spent on AI research.
- In 2023 Templeton announced a 30 million dollar commitment to a programme of adversarial consciousness research. That is one programme, not the global or annual budget of the discipline.
- The available sources have different coverage and time boundaries. Their ratio is not a valid estimate of global asymmetry.
What we do not know
- Which spending over the same period can be assigned to building capability and which to understanding, without double counting.
- What share of consciousness research is covered by state, university, private and philanthropic budgets.
- Whether funding is the bottleneck in specific lines of research, and how to assess the benefit of changing it.
Our position
We hold that the documented difference in magnitude gives grounds for a systematic audit of how resources are distributed. That limited conclusion is what the stated confidence refers to. A global gap of some particular factor, its causes, and the optimal distribution of resources are not yet established.
What would change this position
The position would be changed by a full comparable accounting showing that the examples chosen do not reflect the wider distribution. Any conclusion about the need for additional resources in a given area will also depend on whether they remove a verified constraint.
Why this matters to synchronism
The proposal to direct resources into critical lags requires data about the lags themselves and about the available ways of working on them. This question lets the resource hypothesis of synchronism be tested — including by finding cases where it fails.