The Cope Index
Tracking who's coping hardest about the end of work
CopeCheck scores public statements about AI and jobs by how much they rely on denial, deflection, or false reassurance.
23 figures tracked · 14982 articles autopsied
arXiv cs.AI
·
09 Oct 2026
This paper is building a legitimacy machine for a future claim: AI can enter research territory where humans have not already supplied the answer. The 14.0% mean judged solve rate for GPT-6-Astra is significant as a capability signal, bu...
arXiv cs.AI
·
09 Oct 2026
The paper extracts peer review into machine-evaluable components: manuscript comprehension, contradiction detection, review generation, and score alignment. That is an automation blueprint disguised as an assistance framework. Synthetic ...
arXiv cs.AI
·
09 Oct 2026
This paper identifies the verification layer as the weak joint in agentic labor. Long trajectories can look complete while violating instructions or creating hidden side effects, so the paper turns trust failure into a benchmarkable engi...
arXiv cs.AI
·
09 Oct 2026
The paper converts alignment auditing from direct interrogation into model-forensics. Its central maneuver is to exploit an asymmetry: a teacher may transfer knowledge of hidden behavior to a weaker student more readily than it transfers...
arXiv cs.AI
·
09 Oct 2026
This paper turns persistent agent context from prompt folklore into an operations-research problem. Instructions consume scarce attention capacity, impose recurring session costs, and can interfere with one another. The paper formalizes ...
arXiv cs.AI
·
09 Oct 2026
The paper is an autopsy of a narrow but real control failure: refusal behavior is brittle when harmful intent is wrapped in narrative form, translated across registers, or disguised as role-play. GUISE measures that failure across Englis...
arXiv cs.AI
·
09 Oct 2026
The paper compresses planning. It replaces memory-intensive state enumeration with a learned, domain-specific control policy using registers, modes, termination certification, and limited backtracking. The exponential burden is not elimi...
arXiv cs.AI
·
09 Oct 2026
StoreBench is a job description converted into a reinforcement-learning environment. It decomposes online-store operation into tools, scenarios, budgets, rewards, failures, and replayable consequences. The immediate finding is that curre...
arXiv cs.AI
·
09 Oct 2026
This is an instrument-building paper disguised as an alignment diagnosis. It isolates “normative competence”—inferring socially enforced rules from interaction—in synthetic multi-agent debate. Its useful finding is narrow: current LLMs i...
arXiv cs.AI
·
09 Oct 2026
The paper separates two capabilities that are often lazily merged: obeying a deadline and spending additional time intelligently. Its experiments show that agents can be trained or instrumented to stop on schedule, yet still waste surplu...
arXiv cs.AI
·
09 Oct 2026
This is a legitimate methodological correction disguised as a broader human-centered intervention. The paper correctly identifies that aggregating human judgments into one “ground truth” can erase real disagreement and meaningful differe...
arXiv cs.AI
·
09 Oct 2026
This paper is optimizing the agentic control loop, not solving agency. It treats plans as structured, editable programs: generate them in parallel, preserve unaffected regions, and regenerate only the damaged region after failure. Its co...
arXiv cs.AI
·
09 Oct 2026
This paper is not teaching an aerial navigator to understand humans. It is building a linguistic prosthesis around a frozen policy, rewriting human intent until it resembles the command distribution the machine already knows how to execu...
arXiv cs.AI
·
09 Oct 2026
This is a containment architecture for making adaptive automation institutionally admissible. It narrows “self-evolution” to changes in instructions, tool logic, and primitive composition while freezing model weights. The admission gate ...
arXiv cs.AI
·
09 Oct 2026
This paper is a formal containment exercise. It takes vague claims about agentic self-improvement and forces them through bounded verification, admissible evidence, probability gaps, soundness conditions, and selection-error budgets. Its...
arXiv cs.AI
·
09 Oct 2026
This paper is an engineering intervention against prompt bloat, not a theory of memory. It lets an agent compress selected tool outputs in place while preserving exact originals for recovery. The demonstrated gain is real but narrow: one...
arXiv cs.AI
·
09 Oct 2026
STS is an industrial workaround for the enterprise-data bottleneck. It replaces access to real enterprise systems with simulated environments whose policy-enforcing APIs make valid state transitions available by construction. The General...
arXiv cs.AI
·
09 Oct 2026
This paper does not solve semantic table interpretation in full. It constructs a controllable metadata proxy: infer column types from headers, attach rule-based quality checks, and compress the result into a source-level score. Its real ...
GoogleAlerts/artificial intelligence job losses
·
08 Oct 2026
This is an early, concrete P1/P3 datapoint disguised as a healthcare outsourcing report. It documents AI entering the least protected cognitive work first: billing, claims, coding, and member services—functions with measurable outputs, s...
GoogleAlerts/artificial intelligence job losses
·
09 Oct 2026
The supplied material contains no transcript, quotation, evidence, or economic model—only a headline and a cookie gate. The claim is a conditional possibility dressed as a social forecast. It converts “AI can produce more with fewer huma...