In Ballot Bench, Anthropic's Claude Opus 5.5 deployed approximately 170 to 180 sub-agents to parallelize state and ballot initiative polling research

Overview Debates Related claims Topics

In Ballot Bench's House and Senate races, frontier artificial intelligence agents outperformed human election modelers while performing slightly worse than prediction markets

Thoughtful developed Ballot Bench, a benchmark evaluating nine frontier artificial intelligence agents forecasting the 2026 United States midterm elections

In Ballot Bench, each artificial intelligence agent received an isolated Modal sandbox, a $500 budget for paid news subscriptions and data, and seven days to complete forecasts

In Ballot Bench's Senate forecasting, Claude Fable 5.1 achieved the highest Brier score among artificial intelligence models while trailing human forecaster Nate Silver

In Ballot Bench, OpenAI's GPT-6.1 SOL aggregated election forecasts by analyzing the historical error shifts of human modelers over the decade prior to 2026

Theory Ventures deploys thousands of autonomous artificial intelligence research agents to link, enrich, and analyze data during investment due diligence

Ballot Bench trading agents executing market orders based on frontier model forecasts lost money on election day due to slow reaction times and fabricated information

Between 5% and 10% of artificial intelligence agent jobs executed at Theory Ventures require immediate processing

METR found that the pace of capability improvement for Anthropic AI models roughly doubled over the course of a year

Enterprise spending on agentic coding applications grew at a 7.7-fold rate in 2026, reaching 34 billion dollars

Annual revenue from the AI lab coding agents Codex and Claude Code grew from 1 billion dollars in 2025 to 24 billion dollars in 2026

CellFlow applies flow-matching generative modeling to latent representations to predict cell state trajectories resulting from biological perturbations

The ARC Institute hosted the Virtual Cell Challenge to benchmark machine learning methods on single-cell perturbation prediction tasks

When an expected reward is omitted, dopamine neurons show a reduction in baseline firing activity that mirrors a negative temporal difference error

Logistic regression baselines match or exceed deep neural networks in single-cell cell type classification accuracy within individual organs

In A. Harry Klopf's neuronal model of reinforcement learning, neurons form coalitions based on mutual reward benefit rather than classical Hebbian cell assemblies

A. Harry Klopf hypothesized that biological intelligence emerges from individual neurons operating as reinforcement learning agents that reward and penalize one another

Traditional multi-objective optimization methods combine multiple objectives into a single value by computing a weighted sum across individual channels

Michael Littman conducted research applying game theoretic principles to multi-agent reinforcement learning systems

Reinforcement learning agents equipped with predictive algorithms can autonomously establish secondary rewards based on environmental patterns