Technical notes on applied ML, physics-based modeling, and engineering decisions for physical systems.https://john-hodge.com/en-ushttps://john-hodge.com/blog/slopscore-0-9-0-performative-candor/https://john-hodge.com/blog/slopscore-0-9-0-performative-candor/Candor-heavy prose scored 0.0 on every slop dimension and earned credit as human writing. Version 0.9.0 adds performative_candor and ten rules to score it.Mon, 03 Aug 2026 00:00:00 GMTperformative candormanufactured sincerityAI writing tellsprose linterslopscoreAI slopwriting qualitynon-native English biaspre-commitstatic analysisPythonopen sourcehttps://john-hodge.com/blog/auditing-tool-call-traces-mcp-agents/https://john-hodge.com/blog/auditing-tool-call-traces-mcp-agents/An outcome grader tells you the refund never happened. Only the trace tells you which of the four layers between the model and the payment API broke.Sat, 01 Aug 2026 00:00:00 GMTMCP tracingModel Context Protocoltool call auditagent observabilityagent debuggingMCP InspectorOpenTelemetry GenAItrace redactionagent evalstool call errorsisErroragent telemetryhttps://john-hodge.com/blog/how-to-measure-a-grader/https://john-hodge.com/blog/how-to-measure-a-grader/A grader is a role in an eval rather than a technology. Deterministic checks and LLM judges both fill it, and each carries a different measurement cost.Wed, 29 Jul 2026 00:00:00 GMTgraderLLM-as-judgeagent evalseval harnessjudge calibrationinter-annotator agreementCohen's kappaposition biasverbosity biaseval validitydeterministic gradingreward hackinghttps://john-hodge.com/blog/failure-taxonomies-ml-decision-systems/https://john-hodge.com/blog/failure-taxonomies-ml-decision-systems/Aggregate scores detect that behavior changed. Bucketing failures by stage, causal role, and opportunity rate is what routes a fix to the right owner.Tue, 28 Jul 2026 00:00:00 GMTfailure taxonomyerror analysisML evaluationfailure modesagent evalsroot cause analysismodel debuggingslice discoverybenchmark validitydecision systemsfailure bucketingeval metricshttps://john-hodge.com/blog/phased-array-systems-0-9-0/https://john-hodge.com/blog/phased-array-systems-0-9-0/v0.9.0 brings NSGA-II multi-objective optimization and Sobol sensitivity to phased-array trade studies, on top of ITU-R propagation, real clutter models, and exact detection statistics.Wed, 22 Jul 2026 00:00:00 GMTphased array trade studyMDAONSGA-IIPareto frontradar detection rangelink budgetITU-R P.676Sobol sensitivity analysisdigital beamforming architectureRF system engineeringrequirements verificationPython radar modelinghttps://john-hodge.com/blog/phased-array-modeling-1-4-0/https://john-hodge.com/blog/phased-array-modeling-1-4-0/v1.4.0 adds a vector pattern engine: polarized element models, co/cross-pol cuts, axial ratio and XPD maps, measured pattern import, and polarization-correct conformal arrays.Tue, 21 Jul 2026 00:00:00 GMTphased array polarizationco-pol cross-pol patternaxial ratiocross-polarization discriminationLudwig-3polarized element patternconformal arrayantenna array Pythonphased array modelingmeasured element patternPython antenna libraryXPDhttps://john-hodge.com/blog/opentelemetry-genai-semantic-conventions/https://john-hodge.com/blog/opentelemetry-genai-semantic-conventions/The OTel GenAI conventions moved to a dedicated repository with no versioned release yet, nothing GenAI-specific is stable, and frameworks emit several generations at once. The current map.Fri, 17 Jul 2026 00:00:00 GMTOpenTelemetry GenAI semantic conventionsgen_ai attributesgen_ai.provider.namegen_ai.systemLLM observabilityagent tracingOTEL_SEMCONV_STABILITY_OPT_INgen_ai.evaluation.resultinvoke_agent spanLLM telemetrysemantic conventions updateGenAI observabilityhttps://john-hodge.com/blog/metasurface-py-python-metasurface-design/https://john-hodge.com/blog/metasurface-py-python-metasurface-design/An open-source Python toolkit for RIS, reflectarray, and metasurface antenna design: beam steering, phase quantization, hardware constraints, link budgets, sensing, and optimization.Wed, 15 Jul 2026 00:00:00 GMTPython metasurface designRIS simulation Pythonreflectarray Pythonmetasurface optimizationreconfigurable intelligent surface softwarebeam steering Pythonphase quantizationRIS link budgetISAC sensingmetasurface antennaopen source RF engineeringmetasurface-pyhttps://john-hodge.com/blog/openem-wave-lab-interactive-em-physics/https://john-hodge.com/blog/openem-wave-lab-interactive-em-physics/An open-source browser lab for EM wave physics: plane waves, Fresnel coefficients, phase vs group velocity, and standing waves, computed live from validated closed-form models.Wed, 15 Jul 2026 00:00:00 GMTelectromagnetic wave simulationinteractive electromagneticsEM wave visualizationFresnel equationsBrewster angletotal internal reflectionphase velocity vs group velocitystanding wave ratioRF engineering educationplane wave polarizationPoynting vectorphysics education toolshttps://john-hodge.com/blog/reconfigurable-intelligent-metasurfaces-phd-research/https://john-hodge.com/blog/reconfigurable-intelligent-metasurfaces-phd-research/A retrospective on my Virginia Tech PhD: reconfigurable intelligent metasurfaces (RIS) for beam steering and index modulation, a fabricated two-bit prototype, and generative deep learning for inverse design.Wed, 15 Jul 2026 00:00:00 GMTreconfigurable intelligent metasurfacesreconfigurable intelligent surfaceRISprogrammable metasurfacereflectarrayindex modulationmetasurface beamforminginverse designgenerative deep learning electromagnetics6G wirelessmetasurface hardware prototypePhD researchhttps://john-hodge.com/blog/llm-agent-eval-harness/https://john-hodge.com/blog/llm-agent-eval-harness/A working recipe for evaluating agentic workflows: isolated trials, outcome-first grading, invariant gates, reliability across repeated runs, and a benchmark you can trust.Sun, 12 Jul 2026 00:00:00 GMTLLM agent evalseval harnessagent evaluationagent benchmarkspass^kLLM-as-judgeagent testingreward hackingtrace gradingagent reliabilityAI agent testingevaluating AI agentshttps://john-hodge.com/blog/what-engineers-review-in-agentic-systems/https://john-hodge.com/blog/what-engineers-review-in-agentic-systems/When a model chooses tools and actions at runtime, requirements and code review do not disappear. Engineers review the decision environment: tools, permissions, traces, and evals.Sat, 11 Jul 2026 00:00:00 GMTagentic systemsAI agentsagentic workflowscode review for AI agentsagent observabilityagent evaluationLLM tool useagent harnessAI requirementshuman oversight of AIsoftware engineering with AIagent traceshttps://john-hodge.com/blog/strands-ollama-opentelemetry-local-agent-tracing/https://john-hodge.com/blog/strands-ollama-opentelemetry-local-agent-tracing/Instrument a fully local tool-calling agent with OpenTelemetry: Strands Agents plus Ollama, traces in Jaeger or Grafana, no cloud APIs and no observability SaaS.Sat, 11 Jul 2026 00:00:00 GMTStrands AgentsOllamaOpenTelemetryLLM observabilitylocal LLM agentagent tracingJaegerGrafanaGenAI semantic conventionsagent evalstool callingLLM agent tutorialhttps://john-hodge.com/blog/auditable-financial-planning-agent/https://john-hodge.com/blog/auditable-financial-planning-agent/planner-lab turns a typed household case file into a retirement-readiness memo. Deterministic code does the math; a critic gate blocks any memo with untraceable numbers.Fri, 10 Jul 2026 00:00:00 GMTauditable LLM agentsagent verificationLLM critic gatefinancial planning agentretirement planning softwarefunded ratioMonte Carlo retirement simulationStrands AgentsModel Context Protocollocal LLMOllama agentopen source Pythonhttps://john-hodge.com/blog/apab-0-3-0/https://john-hodge.com/blog/apab-0-3-0/APAB 0.3.0 makes agentic phased-array design auditable: OpenTelemetry traces of every MCP tool call, runtime provenance manifests, real LLM providers with uniform cost tracking, and a golden-task eval harness.Thu, 09 Jul 2026 00:00:00 GMTagentic phased-array designAI antenna designLLM agent for RF engineeringModel Context ProtocolMCP engineering toolsphased array design automationagentic workflowAPABOpenTelemetry LLM tracingagent observabilityrun provenancereproducible AI workflowphased array antenna designhttps://john-hodge.com/blog/gumline-dental-habit-tracker/https://john-hodge.com/blog/gumline-dental-habit-tracker/A private, on-device iPhone app for logging your dental routine and remembering the exact spots your dentist flagged. Why I built it and how it works.Thu, 09 Jul 2026 00:00:00 GMTdental habit trackerbrushing trackerfloss reminderoral hygiene appdentist notes apptooth mapprivate health appon-deviceiOS apphabit streakhttps://john-hodge.com/blog/model-based-engineering-phased-arrays/https://john-hodge.com/blog/model-based-engineering-phased-arrays/A graduate radar project led me to MBSE and MDAO, then to model-based engineering of AESA phased-array radar systems, and finally to a set of open-source phased-array design tools.Thu, 09 Jul 2026 00:00:00 GMTmodel-based engineeringMBSEMDAOmultidisciplinary design optimizationphased array designAESA radarmodel-based systems engineeringSWaP-Cdigital engineeringradar systemsopen-source phased arrayantenna designtrade studyModelCentermetasurfacereconfigurable intelligent surfacehttps://john-hodge.com/blog/pitchphys-baseball-pitch-simulator/https://john-hodge.com/blog/pitchphys-baseball-pitch-simulator/A browser simulator for the physics of a thrown baseball: gravity, drag, and the Magnus effect. Built to understand why spin moves a pitch, and where the model stops.Thu, 09 Jul 2026 00:00:00 GMTMagnus effectbaseball physicspitch trajectoryspin rateinduced vertical breakactive spingyro spindrag coefficientaerodynamicsStreamlithttps://john-hodge.com/blog/connect-four-ai-minimax/https://john-hodge.com/blog/connect-four-ai-minimax/How the Connect Four AI in Neon Drop works: minimax with alpha-beta pruning, a window-scoring heuristic, tunable difficulty, and what it reveals about four-in-a-row strategy.Mon, 06 Jul 2026 00:00:00 GMTconnect four AIconnect four algorithmconnect four minimaxalpha-beta pruningconnect four strategyfour in a row AIconnect four evaluation functionis connect four solvedhow to build a connect four AIgame tree searchhttps://john-hodge.com/blog/helicopter-ml-simulator/https://john-hodge.com/blog/helicopter-ml-simulator/An interactive browser simulator of the Abbeel-Coates-Ng autonomous helicopter: flight physics, apprenticeship trajectory learning, system identification, and LQR control.Mon, 06 Jul 2026 00:00:00 GMTapprenticeship learninglearning from demonstrationhelicopter dynamicssystem identificationLQR controlimitation learningphysics simulationmachine learning for controlroboticsAbbeel Coates Nghttps://john-hodge.com/blog/metasurface-index-modulation/https://john-hodge.com/blog/metasurface-index-modulation/A plain-language explainer of our IEEE JSTSP paper: reconfigurable metasurfaces that perform index modulation for 6G, encoding bits in which beam or frequency harmonic is active, grounded in full-wave EM.Mon, 06 Jul 2026 00:00:00 GMTreconfigurable intelligent surfaceRISindex modulationmetasurface6G wirelessbeam steeringfrequency harmonicsmmWavereflectarraywireless communicationshttps://john-hodge.com/blog/modeling-roasted-coffee-as-a-perishable-asset/https://john-hodge.com/blog/modeling-roasted-coffee-as-a-perishable-asset/Roasted coffee has a hump-shaped quality curve: it improves while resting, peaks, then stales. That makes it a clean toy problem for perishable-inventory pricing, storage economics, and a small client-side model.Sun, 28 Jun 2026 00:00:00 GMTperishable inventorydynamic pricingrevenue managementnewsvendorcoffee freshnessclient-side modelingTypeScriptopen sourcehttps://john-hodge.com/blog/opensatcom/https://john-hodge.com/blog/opensatcom/An open-source Python toolkit for satellite communications: snapshot link budgets, ITU-R propagation, a DVB-S2 modem, multibeam payloads, and Pareto trade studies.Thu, 25 Jun 2026 00:00:00 GMTsatellite communicationslink budgetITU-R propagationDVB-S2rain fadephased arraytrade studyPareto optimizationRF engineeringPythonhttps://john-hodge.com/blog/slopscore/https://john-hodge.com/blog/slopscore/A rule-based prose linter that scores formulaic, generic, low-specificity writing patterns with visible evidence spans. It scores patterns rather than authorship.Thu, 25 Jun 2026 00:00:00 GMTAI slopprose linterwriting qualityAI detectionnon-native English biasSARIFpre-commitcontinuous integrationopen sourcePythonhttps://john-hodge.com/blog/mdo-for-axial-flux-actuators/https://john-hodge.com/blog/mdo-for-axial-flux-actuators/axfluxmdo: fast analytical models, 2.5D annular analysis, Pareto trade studies, and FEA hooks for early-stage axial-flux motor design, aimed at robot actuators.Wed, 24 Jun 2026 00:00:00 GMTaxial-flux motorrobot actuatormultidisciplinary design optimizationPareto optimizationtorque densitypole pairselectric machine designOpenMDAOpymoodirect-drive actuatorhttps://john-hodge.com/blog/agentic-workflow-for-phased-arrays/https://john-hodge.com/blog/agentic-workflow-for-phased-arrays/I built APAB, an LLM that orchestrates a chain of phased-array physics tools through MCP and emits an auditable run bundle. The agent drives; the solvers compute.Tue, 23 Jun 2026 00:00:00 GMTphased arrayLLM agentMCPModel Context Protocolagentic workflowEdgeFEMmutual couplingprovenanceRF engineeringOllamahttps://john-hodge.com/blog/predicting-the-next-pitch/https://john-hodge.com/blog/predicting-the-next-pitch/A baseball pitch-prediction package as a lab: separate pitcher pitch-mix, count, fatigue, and sequence memory, and test which models actually use each.Tue, 23 Jun 2026 00:00:00 GMTpitch sequencingbaseball analyticssequence modelingbaseline laddermodel calibrationdata leakagecross-validationmachine learning evaluationlog lossStatcasthttps://john-hodge.com/blog/modeling-phased-arrays-in-python/https://john-hodge.com/blog/modeling-phased-arrays-in-python/Two open Python tools for phased arrays: one computes the beam pattern, the other asks whether a design meets its link, radar, cost, and reliability margins.Mon, 22 Jun 2026 00:00:00 GMTphased arrayantenna patternarray factorbeam steeringgrating lobesradar equationlink budgettrade studyPareto frontierRF systems engineeringhttps://john-hodge.com/blog/sizing-ai-training-by-bottleneck/https://john-hodge.com/blog/sizing-ai-training-by-bottleneck/A small open model for AI hardware: find whether a workload is compute-, memory-, or network-bound, then price the resource that actually binds it.Mon, 22 Jun 2026 00:00:00 GMTAI training hardwareGPU sizingmemory bandwidthroofline modelarithmetic intensityHBMcost per TB/sKV cacheLLM inferencetokens per secondhttps://john-hodge.com/blog/orbital-data-centers-by-the-numbers/https://john-hodge.com/blog/orbital-data-centers-by-the-numbers/An open, physics-based model of space-based AI data centers: why inter-satellite links are easy, the downlink is hard, and orbit runs about 19x Earth's cost today.Sun, 21 Jun 2026 00:00:00 GMTspace-based AI data centersorbital data centersProject SuncatcherStarcloudlevelized cost of computeoptical downlink bandwidththermal management in spacedelivered computeAI infrastructure