{"ontology":"https://layer.re/machine/schema.json","epistemic_model":"evidence-backed-temporal-layer-graph","canonical":true,"type":"layer_history","layer":{"id":"layer-research","slug":"research","name":"Research","description":"Formal inquiry, papers, datasets and benchmarks.","machine_description":"Formal inquiry, papers, datasets and benchmarks.","active":1},"snapshots":[{"id":"state-research-2022q1","layer_id":"layer-research","as_of":"2022-03-31 23:59:59.000","summary":"Research was a primary upstream driver of model capability, with Q1 work on RLHF deployment, prompting, code generation and compute-optimal scaling quickly connecting to deployable systems.","dependencies":["layer-compute","layer-models","academic and industrial research organizations"],"constraints":["experimental compute requirements","benchmark representativeness","rapidly changing baselines"],"emerging_capabilities":["RLHF methods","chain-of-thought prompting","compute-optimal scaling analysis"],"underdeveloped_components":["standardized independent evaluation"],"uncertainties":["some preprints were revised later; only Q1-supported claims should be attributed to the quarter"],"confidence":0.94,"previous_snapshot_id":null,"mapping_mode":"retrospective","reconstructed_at":"2026-10-01T17:34:32.000Z","created_at":"2026-10-01T18:05:36.388Z","updated_at":"2026-10-01T18:05:36.388Z"},{"id":"state-research-2022q2","layer_id":"layer-research","as_of":"2022-06-30 23:59:59.000","summary":"Research became more tightly coupled to system engineering and deployment surfaces: papers increasingly described complete model+compute+interface or model+environment systems rather than isolated algorithmic improvements.","dependencies":["layer-models","layer-compute","layer-software","layer-robotics"],"constraints":["compute concentration","reproducibility","benchmark validity","controlled access to frontier models"],"emerging_capabilities":["generalist policy research","multimodal prompting","quantitative reasoning systems"],"underdeveloped_components":["independent replication of frontier-scale results","standard evaluation of agent-like behavior"],"uncertainties":["frontier results remained concentrated in a small number of labs"],"confidence":0.95,"previous_snapshot_id":"state-research-2022q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-01T18:14:00.000Z","created_at":"2026-10-01T19:22:17.762Z","updated_at":"2026-10-01T19:22:17.762Z"},{"id":"state-research-2022q3","layer_id":"layer-research","as_of":"2022-09-30 23:59:59.000","summary":"Research in Q3 increasingly produced reusable public model artifacts and persistent scientific infrastructure while embodied-agent work closed feedback loops between language models and environments.","dependencies":["layer-models","layer-compute","layer-software","layer-open-source","layer-robotics"],"constraints":["frontier compute concentration","replication cost","evaluation validity","difference between predicted and experimentally verified scientific outputs"],"emerging_capabilities":["large open collaborative model releases","AI-generated scientific databases at near-catalog scale","closed-loop embodied planning research"],"underdeveloped_components":["independent frontier replication","standardized agent evaluation"],"uncertainties":["downstream scientific impact of model-generated databases was not yet fully measurable"],"confidence":0.95,"previous_snapshot_id":"state-research-2022q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-02T15:12:29.000Z","created_at":"2026-10-02T16:08:16.256Z","updated_at":"2026-10-02T16:08:16.256Z"},{"id":"state-research-2022q4","layer_id":"layer-research","as_of":"2022-12-31 23:59:59.000","summary":"Q4 research increasingly focused on composing model capabilities into interactive systems and scalable behavior shaping: ReAct coupled reasoning/actions, CICERO coupled planning/dialogue, RT-1 scaled real-world robot policies, and Constitutional AI explored AI-assisted supervision.","dependencies":["layer-models","layer-software","layer-robotics","layer-human","layer-compute"],"constraints":["benchmark-to-deployment gap","frontier compute concentration","evaluation validity","safety measurement","robot data cost"],"emerging_capabilities":["reasoning-action loops","strategic language agents","large-data transformer robot policies","RLAIF and constitutional self-revision"],"underdeveloped_components":["standard agent evaluation","general-purpose physical autonomy","validated scalable alignment"],"uncertainties":["each system remained bounded by its research setting and evaluation design"],"confidence":0.97,"previous_snapshot_id":"state-research-2022q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-03T14:26:11.000Z","created_at":"2026-10-03T14:45:10.924Z","updated_at":"2026-10-03T14:45:10.924Z"},{"id":"state-research-2023q1","layer_id":"layer-research","as_of":"2023-03-31 23:59:59.000","summary":"Research broadened around tool use, frontier-model evaluation and alternative model access, while deployment increasingly fed back into governance and safety questions.","dependencies":["layer-models","layer-compute","layer-software","layer-human"],"constraints":["frontier reproducibility","evaluation opacity","compute access","deployment/research gap"],"emerging_capabilities":["Toolformer-style tool use","GPT-4 capability/safety evaluation","LLaMA research experimentation"],"underdeveloped_components":["standardized agent/tool evaluation","independent frontier replication","transparent frontier training detail"],"uncertainties":["Research results do not establish broad product reliability."],"confidence":0.95,"previous_snapshot_id":"state-research-2022q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-05T15:23:51.000Z","created_at":"2026-10-05T18:17:53.800Z","updated_at":"2026-10-05T18:17:53.800Z"},{"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"},{"id":"state-research-2023q3","layer_id":"layer-research","as_of":"2023-09-30 23:59:59.000","summary":"Q3 research connected general model representations more directly to physical action and complex software orchestration through RT-2 and AutoGen while open model research ecosystems continued expanding.","dependencies":["layer-models","layer-robotics","layer-agents","layer-open-source","layer-compute"],"constraints":["evaluation validity","robot data cost","multi-agent reliability","frontier replication cost"],"emerging_capabilities":["vision-language-action transfer","multi-agent workflow orchestration","commercial/open model experimentation"],"underdeveloped_components":["standard agent evaluation","general physical autonomy","reliable emergent multi-agent coordination"],"uncertainties":["research frameworks and robotics demonstrations should not be conflated with production autonomy"],"confidence":0.96,"previous_snapshot_id":"state-research-2023q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T16:04:51.000Z","created_at":"2026-10-07T16:56:16.866Z","updated_at":"2026-10-07T16:56:16.866Z"},{"id":"state-research-2023q4","layer_id":"layer-research","as_of":"2023-12-31 23:59:59.000","summary":"Research expanded cross-embodiment robotics, open safety evaluation tooling and frontier-model governance/evaluation requirements.","dependencies":["layer-models","layer-robotics","layer-open-source","layer-governance","layer-compute"],"constraints":["evaluation standardization","compute access","robot data heterogeneity"],"emerging_capabilities":["Open X-Embodiment","RT-X","CyberSec Eval","Llama Guard"],"underdeveloped_components":["general cross-embodiment autonomy","validated universal frontier-safety tests"],"uncertainties":["Research releases and policy-directed evaluation do not prove broad deployment effectiveness"],"confidence":0.96,"previous_snapshot_id":"state-research-2023q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-08T15:50:00.000Z","created_at":"2026-10-08T17:00:00.130Z","updated_at":"2026-10-08T17:00:00.130Z"},{"id":"state-research-2024q1","layer_id":"layer-research","as_of":"2024-03-31 23:59:59.000","summary":"Research combined inspectable language-model training, generated interactive environments and instruction following across game interfaces; production autonomy and semantic state were not established.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["OLMo publishes an open training and research stack","Sora research demonstrates text-conditioned video generation","Genie paper demonstrates action-controllable generated environments","SIMA research follows instructions through game interfaces"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["Research reproducibility does not imply frontier parity or independently reproduced vendor performance.","The research report does not establish general availability, accurate physics or an executable game-state engine.","Generated frames and latent actions do not establish persistent semantic game state or a production game engine.","Approximately ten-second tasks do not establish long-horizon autonomy, public deployment or semantic runtime-state access."],"confidence":0.93,"previous_snapshot_id":"state-research-2023q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-09T21:04:17.186Z","updated_at":"2026-10-09T21:04:17.186Z"},{"id":"state-research-2024q2","layer_id":"layer-research","as_of":"2024-06-30 23:59:59.000","summary":"Research linked responsive video/speech assistant prototypes and open robot-action policies while transfer and deployment limitations remained explicit.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["Project Astra demonstrates responsive multimodal assistant prototypes","OpenVLA publishes a vision-language-action research model"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["A demonstrated responsive prototype is not a shipped persistent autonomous agent.","Selected robot evaluations do not establish universal embodiment transfer or commercial autonomous deployment."],"confidence":0.93,"previous_snapshot_id":"state-research-2024q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T07:29:06.788Z","updated_at":"2026-10-10T07:29:06.788Z"},{"id":"state-research-2024q3","layer_id":"layer-research","as_of":"2024-09-30 23:59:59.000","summary":"Inference-time reasoning and action-conditioned simulation became explicit research trajectories without proving general autonomy.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["NVIDIA introduces humanoid simulation and orchestration tooling","GameNGen publishes action-conditioned Doom simulation research","OpenAI releases o1 reasoning previews"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["Training infrastructure and teleoperation do not establish autonomous humanoid production deployment.","Reported real-time simulation is not a deterministic semantic engine, editable authoritative rules, or a shipped general game runtime.","Next-update benchmark results are not measured capabilities of the released preview; launch lacked browsing and image upload."],"confidence":0.93,"previous_snapshot_id":"state-research-2024q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T07:49:15.921Z","updated_at":"2026-10-10T07:49:15.921Z"},{"id":"state-research-2024q4","layer_id":"layer-research","as_of":"2024-12-31 23:59:59.000","summary":"Experimental multimodal and browser/coding prototypes explored more agentic interaction, with access and reliability limits intact.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["Pi-zero publishes general robot-control research","Genie 2 demonstrates action-controlled generative 3D worlds","Gemini 2.0 Flash releases experimentally"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["Experimental cross-embodiment results are not general commercial autonomy; later weight releases and revised results are excluded.","Generated visual continuity is not an authoritative editable semantic game engine or a released universal runtime.","Experimental model access is not GA; Project Mariner and Jules are research prototypes or limited trials."],"confidence":0.93,"previous_snapshot_id":"state-research-2024q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T08:02:14.506Z","updated_at":"2026-10-10T08:02:14.506Z"},{"id":"state-research-2025q1","layer_id":"layer-research","as_of":"2025-03-31 23:59:59.000","summary":"Reasoning and embodied-action research expanded through open resources and partner demonstrations.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["DeepSeek-R1 releases reasoning weights","Gemini Robotics introduces VLA and embodied reasoning research","GR00T N1 releases customizable humanoid foundation model"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["Vendor o1-parity and low-cost narratives are not independent total-cost evidence; distilled models retain upstream-license considerations.","Demonstrations and trusted access are not broad commercial robot autonomy.","Customizable research resources do not establish general-purpose humanoid deployment."],"confidence":0.93,"previous_snapshot_id":"state-research-2024q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T08:11:02.815Z","updated_at":"2026-10-10T08:11:02.815Z"},{"id":"state-research-2025q2","layer_id":"layer-research","as_of":"2025-06-30 23:59:59.000","summary":"This period adds specific research evidence through Gemini Robotics On-Device is introduced. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["Selected-access demonstrations do not prove general commercial deployment or cross-embodiment reliability."],"emerging_capabilities":["Gemini Robotics On-Device is introduced"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-research-2025q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T08:25:19.367Z","updated_at":"2026-10-10T08:25:19.367Z"},{"id":"state-research-2025q3","layer_id":"layer-research","as_of":"2025-09-30 23:59:59.000","summary":"This period adds specific research evidence through METR publishes experienced-developer productivity experiment; Genie 3 interactive world-model research is introduced. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["The measured nineteen-percent slowdown is specific to this sample and setting, not every developer or later model.","Generated video worlds do not establish semantically addressable persistent game state or a shipped game engine."],"emerging_capabilities":["METR publishes experienced-developer productivity experiment","Genie 3 interactive world-model research is introduced"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-research-2025q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T08:31:22.795Z","updated_at":"2026-10-10T08:31:22.795Z"},{"id":"state-research-2025q4","layer_id":"layer-research","as_of":"2025-12-31 23:59:59.000","summary":"This period adds specific research evidence through SIMA 2 research extends virtual-world agent interaction. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["Research interaction through game interfaces does not prove native semantic game-state export or production deployment."],"emerging_capabilities":["SIMA 2 research extends virtual-world agent interaction"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-research-2025q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T08:37:10.930Z","updated_at":"2026-10-10T08:37:10.930Z"},{"id":"state-research-2026q1","layer_id":"layer-research","as_of":"2026-03-31 23:59:59.000","summary":"This period adds specific research evidence through Project Genie opens a limited interactive-world experiment. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["Bounded experimental access is not a general game platform or persistent semantic game state."],"emerging_capabilities":["Project Genie opens a limited interactive-world experiment"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-research-2025q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T08:45:32.713Z","updated_at":"2026-10-10T08:45:32.713Z"},{"id":"state-research-2026q2","layer_id":"layer-research","as_of":"2026-06-30 23:59:59.000","summary":"This period adds specific research evidence through Project Glasswing launches with restricted security-model access; Project Genie expands access and adds location-based worlds; Glasswing expands selected defensive access. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["Partner access is restricted; October updates on the same page are excluded from April reconstruction.","Experimental location reconstruction is distinct from native game runtime state or production simulation fidelity.","Expanded eligibility remains controlled access and does not establish general availability of Mythos."],"emerging_capabilities":["Project Glasswing launches with restricted security-model access","Project Genie expands access and adds location-based worlds","Glasswing expands selected defensive access"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-research-2026q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T09:06:48.256Z","updated_at":"2026-10-10T09:06:48.256Z"},{"id":"state-research-20260930-late","layer_id":"layer-research","as_of":"2026-09-30 21:20:00.000","summary":"AI and cybersecurity research collaboration is increasingly intersecting with national-security screening, as shown by MI5's alert over alleged covert state-linked funding of UK academic work.","dependencies":["academic institutions","layer-geopolitics","layer-governance"],"constraints":["research security","attribution uncertainty","cross-border trust"],"emerging_capabilities":["more explicit research-security screening"],"underdeveloped_components":["transparent standards that preserve legitimate collaboration while mitigating covert influence"],"uncertainties":["extent and impact of the alleged activity"],"confidence":0.87,"previous_snapshot_id":null,"mapping_mode":"contemporaneous","reconstructed_at":null,"created_at":"2026-09-30T21:57:03.551Z","updated_at":"2026-09-30T21:57:03.551Z"},{"id":"state-research-2026q3","layer_id":"layer-research","as_of":"2026-09-30 23:59:59.000","summary":"This period adds specific research evidence through Gemini Robotics 2 extends whole-body robot control; OpenAI and Synopsys agree to develop a chip-design model. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["Demonstrated embodiments and bounded access do not establish general commercial humanoid autonomy.","The specialized model and bundled service are prospective; the agreement is not a shipped autonomous chip-design system."],"emerging_capabilities":["Gemini Robotics 2 extends whole-body robot control","OpenAI and Synopsys agree to develop a chip-design model"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-research-2026q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T09:18:22.071Z","updated_at":"2026-10-10T09:18:22.071Z"}]}