{"ontology":"https://layer.re/machine/schema.json","epistemic_model":"evidence-backed-temporal-layer-graph","canonical":true,"type":"civilization_snapshot","snapshot":{"id":"civ-2023q2-reconstructed","as_of":"2023-06-30 23:59:59.000","summary":"By the end of Q2 2023, generative AI had moved beyond becoming a software platform toward forming reusable ecosystem conventions. Model-to-tool interaction broadened into wider plugin deployment, cross-product plugin interoperability and structured API function calling, while execution remained application-controlled rather than persistently autonomous. Open foundation-model development became materially more complete through open training data and multiple commercially reusable model families. Assistant context expanded toward document scale and public assistant access broadened. Governance became more operational and multilateral through accountability-policy formation, compliance-driven software changes, active EU AI Act negotiation and G7 coordination. Semiconductor industrial policy broadened through a provisional EU Chips Act agreement, while robotics research demonstrated iterative multi-embodiment training using newly generated experience. Persistent autonomous agents, generalized permission infrastructure, reliable tool execution, cheap frontier compute, realized new semiconductor capacity and binding international AI governance were not yet established.","cross_layer_changes":["tool integration became reusable across products/APIs","open foundation-model development matured","long context expanded","governance became more negotiated and multilateral","EU semiconductor policy matured","robotics added iterative generated-experience training"],"dependencies":["models + software -> tool-enabled systems","open data + weights + compute -> open models","governance + software -> compliance adaptation","models + robot data -> iterative robotics training"],"constraints":["authorization","tool correctness","long-context reliability","compute cost","unfinished regulation","semiconductor implementation lag"],"emerging_capabilities":["common plugin interfaces","structured function calling","100K-class context","commercially reusable open models","open training datasets","iterative robotics retraining"],"possible_bottlenecks":["permissions","tool safety","open-model compute","governance interoperability","semiconductor timelines","robotics data"],"uncertainties":["plugins/function calling are not persistent agents","open models do not remove compute constraints","long context does not guarantee reliability","EU/G7 processes were not final binding law","RoboCat was bounded research"],"mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.614Z","updated_at":"2026-10-07T14:04:48.614Z"},"claims":[{"id":"claim-q2-eu-semiconductor-policy","statement":"The provisional EU Chips Act agreement broadened semiconductor capacity/resilience policy beyond the U.S. implementation path.","epistemic_status":"supported_inference","confidence":0.95,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-manufacturing-2023q2","competing_interpretation":"This may be a legislative milestone without immediate manufacturing-state effect.","uncertainty":"No physical capacity is inferred.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.617Z","updated_at":"2026-10-07T14:04:48.617Z"},{"id":"claim-q2-governance-accountability","statement":"Governance moved toward accountability mechanisms, compliance feedback, active legislative negotiation and multilateral coordination.","epistemic_status":"supported_inference","confidence":0.97,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-governance-2023q2","competing_interpretation":"These remained heterogeneous governance mechanisms.","uncertainty":"No final EU law or binding G7 framework existed.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.617Z","updated_at":"2026-10-07T14:04:48.617Z"},{"id":"claim-q2-long-context","statement":"Document-scale context became operationally available in assistant models without establishing uniformly reliable long-context reasoning.","epistemic_status":"supported_inference","confidence":0.95,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-models-2023q2","competing_interpretation":"This may be incremental specification growth.","uncertainty":"Recall and reasoning remained uneven.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.616Z","updated_at":"2026-10-07T14:04:48.616Z"},{"id":"claim-q2-open-foundation-stack","statement":"Open training data and multiple commercially reusable foundation models formed a more complete open alternative development stack.","epistemic_status":"supported_inference","confidence":0.97,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-open-source-2023q2","competing_interpretation":"These may be independent releases rather than one stack transition.","uncertainty":"Frontier reproduction still required substantial compute.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.616Z","updated_at":"2026-10-07T14:04:48.616Z"},{"id":"claim-q2-permissions-still-emergent","statement":"Q2 strengthened evidence that tool-enabled assistants depend on authorization boundaries, but permissions remained application-specific and below independent historical-state threshold.","epistemic_status":"supported_inference","confidence":0.92,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-software-2023q2","competing_interpretation":"Permissions can remain modeled within software/governance.","uncertainty":"No interoperable agent authorization infrastructure is evidenced.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.618Z","updated_at":"2026-10-07T14:04:48.618Z"},{"id":"claim-q2-robotics-training-loop","statement":"RoboCat supports an iterative robotics training path using newly generated experience across tasks and embodiments.","epistemic_status":"supported_inference","confidence":0.94,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-robotics-2023q2","competing_interpretation":"This may extend imitation learning rather than constitute a new state.","uncertainty":"No autonomous recursive self-improvement is inferred.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.618Z","updated_at":"2026-10-07T14:04:48.618Z"},{"id":"claim-q2-tool-interface-standardization","statement":"Q2 made model-to-tool interaction a reusable integration pattern through wider plugins, cross-product plugin interoperability and structured function calling.","epistemic_status":"supported_inference","confidence":0.98,"civilization_snapshot_id":"civ-2023q2-reconstructed","layer_state_snapshot_id":"state-agents-2023q2","competing_interpretation":"These mechanisms may be ordinary API integration rather than a distinct agent transition.","uncertainty":"Execution remained app-controlled.","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.615Z","updated_at":"2026-10-07T14:04:48.615Z"}],"layer_states":[{"id":"state-ai-2023q2","layer_id":"layer-ai","as_of":"2023-06-30 23:59:59.000","summary":"Generative AI broadened into an ecosystem of public assistants, integrated provider products, open model families and tool-enabled experiences, while reliability and autonomy remained constrained.","dependencies":["layer-models","layer-software","layer-human","layer-governance"],"constraints":["hallucination","tool safety","uneven access","privacy/compliance"],"emerging_capabilities":["tool-enabled assistants","open model alternatives","long-context assistants"],"underdeveloped_components":["persistent autonomous agents","general permission infrastructure"],"uncertainties":["surface expansion does not establish usage scale"],"confidence":0.97,"previous_snapshot_id":"state-ai-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.608Z","updated_at":"2026-10-07T14:04:48.608Z","layer_slug":"ai","layer_name":"AI"},{"id":"state-agents-2023q2","layer_id":"layer-agents","as_of":"2023-06-30 23:59:59.000","summary":"Agent-like systems gained wider deployed tool access and reusable integration contracts, while execution remained application-controlled and persistence/authority immature.","dependencies":["layer-models","layer-software","layer-human","layer-governance"],"constraints":["tool safety","authorization","persistent identity","durable memory"],"emerging_capabilities":["broader plugins","common plugin interfaces","structured function calls"],"underdeveloped_components":["persistent agents","portable permissions","reliable autonomy"],"uncertainties":["integration primitives are not persistent autonomy"],"confidence":0.96,"previous_snapshot_id":"state-agents-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.610Z","updated_at":"2026-10-07T14:04:48.610Z","layer_slug":"agents","layer_name":"Agents"},{"id":"state-geopolitics-2023q2","layer_id":"layer-geopolitics","as_of":"2023-06-30 23:59:59.000","summary":"Technology geopolitics broadened toward EU semiconductor resilience policy and multilateral AI-governance coordination alongside continuing technology controls.","dependencies":["layer-governance","layer-manufacturing","layer-chips","layer-ai"],"constraints":["policy fragmentation","supply concentration","implementation lag"],"emerging_capabilities":["EU chip resilience policy","G7 AI coordination"],"underdeveloped_components":["stable international AI rules","diversified advanced-node supply"],"uncertainties":["coordination does not guarantee harmonized implementation"],"confidence":0.94,"previous_snapshot_id":"state-geopolitics-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.613Z","updated_at":"2026-10-07T14:04:48.613Z","layer_slug":"geopolitics","layer_name":"Geopolitics"},{"id":"state-governance-2023q2","layer_id":"layer-governance","as_of":"2023-06-30 23:59:59.000","summary":"AI governance became more operational, negotiated and multilateral through accountability work, compliance adaptation, EU AI Act negotiation and G7 coordination.","dependencies":["layer-ai","layer-software","layer-human","layer-geopolitics"],"constraints":["unfinished legislation","jurisdictional divergence","assurance maturity"],"emerging_capabilities":["accountability policy","compliance feedback","EU negotiation","G7 coordination"],"underdeveloped_components":["final AI law","standardized audits","mature agent governance"],"uncertainties":["communiqués and negotiating positions are not binding law"],"confidence":0.98,"previous_snapshot_id":"state-governance-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.613Z","updated_at":"2026-10-07T14:04:48.613Z","layer_slug":"governance","layer_name":"Governance"},{"id":"state-human-2023q2","layer_id":"layer-human","as_of":"2023-06-30 23:59:59.000","summary":"Human access broadened geographically and functionally through wider assistant availability, long-document interaction, tool-enabled chat and visible data controls.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","privacy","tool risk","hallucination"],"emerging_capabilities":["broader Bard access","long-document interaction","wider plugins","data controls"],"underdeveloped_components":["reliable factual interaction","safe delegated action"],"uncertainties":["availability differs from impact"],"confidence":0.96,"previous_snapshot_id":"state-human-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.612Z","updated_at":"2026-10-07T14:04:48.612Z","layer_slug":"human","layer_name":"Human"},{"id":"state-manufacturing-2023q2","layer_id":"layer-manufacturing","as_of":"2023-06-30 23:59:59.000","summary":"Semiconductor industrial policy broadened geographically as EU institutions reached provisional agreement on capacity, resilience and crisis response.","dependencies":["layer-chips","layer-governance","layer-geopolitics"],"constraints":["fab lead times","capital intensity","policy implementation lag"],"emerging_capabilities":["EU semiconductor resilience framework"],"underdeveloped_components":["realized policy-funded capacity"],"uncertainties":["political agreement does not create fabs"],"confidence":0.95,"previous_snapshot_id":"state-manufacturing-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.611Z","updated_at":"2026-10-07T14:04:48.611Z","layer_slug":"manufacturing","layer_name":"Manufacturing"},{"id":"state-models-2023q2","layer_id":"layer-models","as_of":"2023-06-30 23:59:59.000","summary":"Models broadened through new provider families, document-scale context, open foundation models and structured function-call interfaces for larger systems.","dependencies":["layer-research","layer-software","layer-compute","layer-open-source"],"constraints":["context reliability","tool-call correctness","inference cost"],"emerging_capabilities":["100K-class context","structured function calling","open foundation models"],"underdeveloped_components":["reliable long-context reasoning","autonomous execution"],"uncertainties":["vendor benchmark claims remain noncanonical"],"confidence":0.97,"previous_snapshot_id":"state-models-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.609Z","updated_at":"2026-10-07T14:04:48.609Z","layer_slug":"models","layer_name":"Models"},{"id":"state-open-source-2023q2","layer_id":"layer-open-source","as_of":"2023-06-30 23:59:59.000","summary":"Open foundation-model development became materially more complete through open training data, research resources and multiple commercially reusable language-model families.","dependencies":["layer-research","layer-models","layer-software","layer-compute"],"constraints":["training compute","licensing heterogeneity","frontier gap"],"emerging_capabilities":["RedPajama","MPT-7B","Falcon 40B","SAM resources"],"underdeveloped_components":["frontier-scale reproduction without major compute"],"uncertainties":["open access does not mean low-cost reproduction"],"confidence":0.97,"previous_snapshot_id":"state-open-source-2022q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.611Z","updated_at":"2026-10-07T14:04:48.611Z","layer_slug":"open-source","layer_name":"Open source"},{"id":"state-research-2023q2","layer_id":"layer-research","as_of":"2023-06-30 23:59:59.000","summary":"Research increasingly supported reusable model ecosystems through open data/models, multimodal resources and iterative robotics training.","dependencies":["layer-models","layer-open-source","layer-robotics","layer-compute"],"constraints":["compute access","replication cost","robot-data cost"],"emerging_capabilities":["open LLM datasets","open vision resources","iterative robot experience generation"],"underdeveloped_components":["frontier-scale open replication","standardized agent evaluation"],"uncertainties":["open resources did not remove compute bottlenecks"],"confidence":0.96,"previous_snapshot_id":"state-research-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.609Z","updated_at":"2026-10-07T14:04:48.609Z","layer_slug":"research","layer_name":"Research"},{"id":"state-robotics-2023q2","layer_id":"layer-robotics","as_of":"2023-06-30 23:59:59.000","summary":"Robotics research added iterative multi-embodiment learning in which RoboCat generated additional trajectories for later retraining.","dependencies":["layer-models","layer-research","layer-agents"],"constraints":["data collection","safe exploration","embodiment transfer"],"emerging_capabilities":["multi-embodiment adaptation","generated training trajectories"],"underdeveloped_components":["recursive autonomous improvement","open-ended autonomy"],"uncertainties":["bounded research training loop"],"confidence":0.94,"previous_snapshot_id":"state-robotics-2022q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.612Z","updated_at":"2026-10-07T14:04:48.612Z","layer_slug":"robotics","layer_name":"Robotics"},{"id":"state-software-2023q2","layer_id":"layer-software","as_of":"2023-06-30 23:59:59.000","summary":"Software began converging on reusable conventions connecting models to APIs, data and actions, while governance requirements produced observable product changes.","dependencies":["layer-models","layer-agents","layer-human","layer-governance"],"constraints":["authorization","tool reliability","privacy/compliance","vendor fragmentation"],"emerging_capabilities":["common plugin interfaces","structured function calling","product data controls"],"underdeveloped_components":["cross-provider permission standards","portable trust policies"],"uncertainties":["interface convergence is not universal standardization"],"confidence":0.98,"previous_snapshot_id":"state-software-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.610Z","updated_at":"2026-10-07T14:04:48.610Z","layer_slug":"software","layer_name":"Software"}]}