Coalescepedia
What it means, and whether it concerns you · 711 terms
The entries on data, AI, cloud, security and regulation I run into on projects. Each one carries the definition, why it matters to whoever decides, a company example and the related terms.
Most read
Glossary map
Data
AI
Getting AI into production
Trusting the output
Deciding
Getting found
711 terms found
GenAI & LLM · 60
Back to filtersLanguage models: how they work, how they are adapted and where they go wrong.
- AGI (Artificial General Intelligence)
- AI co-scientist
- AI Hallucinations
- AI output homogenization
- AI slop
- AI wrapper
- ANI, AGI and ASI
- Benchmark contamination
- Causal AI
- Composite AI
- Constitutional AI
- Decision trees
- Diffusion language model
- Discriminative vs Generative models
- Domain-Specific Language Model (DSLM)
- Enterprise text-to-video
- Evals (model evaluation)
- Fine-tuning
- Foundation model
- Generative UI
- Knowledge cutoff
- LLM
- LLM leaderboards
- LLM sampling parameters: temperature and top-p
- LLM-as-a-judge
- Looped transformer
- LoRA
- Machine Learning vs Deep Learning
- Multimodal AI
- Neural Network
- Neuro-symbolic AI
- Neuromorphic computing
- Open weights
- Pattern recognition
- Perceptron
- Physical AI
- Predictive vs Generative vs Agentic AI
- Prompting, ICL and Chain of Thought
- Q-learning
- Quantum AI (Quantum Machine Learning)
- Reasoning models
- Reasoning trace
- Recursive Language Models (RLM)
- RLHF vs DPO vs GRPO
- RLVR
- Self-play
- Smart robots
- Spatial Computing
- Structured outputs
- SVM (support vector machine)
- Sycophancy
- System prompt
- Tabular foundation model
- Technological singularity
- Token
- Turing test
- Typed decision models
- VLA (Vision-Language-Action)
- World models
- Zero-shot classification
RAG, search and context · 21
Back to filtersHow a model finds and uses company data: RAG, embeddings, search and context.
- CAG vs RAG
- Context engineering
- Context Rot
- Context window
- Corrective RAG
- Embeddings
- Faceted search
- GraphRAG
- Grounding
- Hybrid search and reranking
- HyDE
- Just-in-time retrieval
- License plate recognition
- LLM-Wiki
- Location biasing
- Multimodal document parsing
- Multimodal RAG
- OCR (Optical Character Recognition)
- Personal Knowledge Management (PKM)
- RAG
- Text-to-SQL
Inference and model efficiency · 21
Back to filtersHow to run a model fast and cheaply: serving, caching and compression.
- AI gateway
- BitNet b1.58 (1-bit LLM)
- Continuous batching
- Distillation
- Flash models
- FlashAttention vs PagedAttention
- Inference
- KV cache
- Local LLM
- Mixture of experts (MoE)
- Model routing
- Multi-token prediction
- Prefill, decode, TTFT and ITL
- Prompt caching
- Quantization (quantizzazione)
- Semantic caching
- SLM
- Speculative decoding
- Test-time compute
- TurboQuant
- vLLM, SGLang and llama.cpp
AI agents and protocols · 50
Back to filtersAgents that carry out tasks: memory, tools, protocols and coordination.
- A2A Protocol (Agent2Agent)
- Agent handoff pattern
- Agent harness
- Agent memory consolidation
- Agent Memory Layer
- Agent planning tool
- Agent Plugins
- Agent Swarm
- Agent virtual filesystem
- Agentic commerce
- Agentic CX
- Agentic RAG
- Agentic Resource Discovery (ARD)
- AgentOps
- AI Agents (Agentic AI)
- AI Employees / Digital Workers
- AI Orchestration
- AI runtime layer
- AI workflows vs AI agents
- APA vs RPA vs IPA
- Argumentation framework
- Computer Use (AI agents that use a computer)
- Conversational IVR
- Durable execution
- FIPA-ACL (Agent Communication Language)
- Graph Engineering
- Guardian agents
- Intent engineering
- Interaction Models
- LangChain, LangGraph and LangSmith
- Loop engineering
- MCP (Model Context Protocol)
- Multiplayer AI
- NLWeb
- OpenClaw
- Persistent agents
- Personal AI agent
- Proactive AI
- Ralph Wiggum Loop
- ReAct pattern
- Reflection pattern
- Skill evolution
- Skill vs MCP vs Hooks
- Subagents
- Task decomposition
- Tool calling (function calling)
- Types of AI agent memory
- UCP vs ACP
- WebMCP
- x402
Tools and models · 41
Back to filtersThe models and products you meet on projects, with their strengths and limits.
- Anthropic / Claude
- ChatGPT
- Claude Code
- CLAUDE.md / AGENTS.md
- Command Code
- CRM (Customer Relationship Management)
- Cursor
- Deep Research
- DeepSeek
- Docker
- dots vs Muse vs Grok Bot
- E-commerce platform
- ERP (Enterprise Resource Planning)
- ERP vs HCM vs CRM vs SCM
- GitHub Copilot
- Google Gemini
- Grok vs Groq
- Hermes Agent
- IEC 61131-3
- Instinct
- Jev
- JSON vs TOON vs CSV
- Lean
- Microsoft 365 vs Google Workspace
- Microsoft Copilot
- Mistral AI
- MLX
- n8n vs Make vs Zapier vs Power Automate
- No-Code vs Low-Code vs Vibe Coding
- Ollama
- OpenAI
- OpenAI Codex
- OpenAI dots
- opencode
- OpenRouter
- Perplexity
- Pi and oh-my-pi
- PolyBuzz vs Character.AI
- Qwen
- ROS 2
- T3 Code
Data platform & architecture · 78
Back to filtersThe architectures and concepts that organize and connect data.
- AI Wearables
- AI-ready data
- Apache Airflow
- Apache Flink
- Apache Kafka
- Apache Ossie (Open Semantic Interchange)
- Backpressure
- Batch vs streaming
- BOM data reconciliation
- Business Intelligence (BI)
- Catchment area
- CDC (Change Data Capture)
- Cohort analysis
- Composable Customer Data Platform (CDP)
- DAG (directed acyclic graph)
- Data App
- Data branching
- Data engineering
- Data fabric
- Data ingestion
- Data lakehouse
- Data mesh
- Data product
- Data Storytelling
- Data Virtualization
- Database-per-service
- Databricks vs Snowflake vs BigQuery vs Fabric
- dbt vs SQLMesh
- Decision Intelligence
- Digital twin
- DuckDB
- EDI
- Embedded Analytics
- ETL vs ELT
- Event-driven architecture (EDA)
- Gaussian splatting
- GigE Vision
- Graph Neural Network (GNN)
- Idempotency and backfill
- IIoT (Industrial IoT)
- iPaaS
- Jinja
- Lambda vs Kappa architecture
- Location Intelligence
- LTAP (Lake Transactional/Analytical Processing)
- MES
- MongoDB vs DynamoDB
- Multi-model database
- Multimodal lakehouse
- OPC UA
- Open table formats (Iceberg, Delta, Hudi)
- Palantir Foundry
- Parquet, Avro and ORC
- Partitioning and clustering
- Player tracking data
- Polars vs Pandas vs Spark
- Predictive analytics
- Predictive maintenance
- Prescriptive Analytics
- Process mining
- Product Information Management (PIM)
- Query execution plan
- Read replica vs CDC
- Replication
- Reverse ETL
- RFQ automation
- SaaS (software as a service)
- SCADA
- Self-Service Analytics
- Semantic layer
- Sports data platform
- Synthetic data
- Tableau vs Power BI vs Looker vs Qlik
- Trino and Presto
- Unified Namespace (UNS)
- Vector database
- Write-Audit-Publish
- Zero-Copy Data Sharing
Data modeling · 37
Back to filtersHow data is shaped: schemas, history, graphs and ontologies.
- ACID
- Bitemporal modeling
- Bounded context
- Bridge table
- Cumulative table design
- Cypher
- Data mart
- Data Vault
- Data warehouse vs data lake
- Entity resolution
- Event-based reification
- Fuzzy matching
- Galaxy schema
- Graph database
- JSON-LD
- JSONL (JSON Lines)
- Junk dimension
- Kimball vs Inmon
- Knowledge graph
- Materialized view
- Medallion architecture
- Named graph
- Normalization vs denormalization
- OLAP vs OLTP
- One Big Table (OBT) vs Dimensional Modeling
- Ontology (data & AI)
- OWL (Web Ontology Language)
- POLE+O model
- RDF (Resource Description Framework)
- RDF-star (RDF*)
- Relational vs NoSQL vs vector vs graph database
- Schema evolution
- SHACL
- Slowly Changing Dimensions (SCD)
- SPARQL
- Star schema
- Taxonomy vs ontology
Data governance & quality · 44
Back to filtersRules, quality and accountability over data, and who answers for what.
- AI Center of Excellence
- AI code contribution policy
- AI consciousness (enterprise risk)
- AI Consultant
- AI crawler access control
- AI governance
- AI Product Manager
- Business glossary
- Catalog as Control Plane
- Chief AI Officer (CAIO)
- Code ownership
- Critical data element
- DAMA-DMBOK
- Data as a balance sheet asset
- Data catalog
- Data Clean Room
- Data contract
- Data democratization
- Data governance
- Data lineage
- Data literacy
- Data observability
- Data portability
- Data quality
- Data silos
- Data steward, data owner and data classification
- eIDAS 2 and the EUDI Wallet
- Evaluator independence
- Golden power
- Hidden stratification
- Master Data Management (MDM)
- Mechanistic interpretability
- Model risk management
- NIST AI RMF
- Open source license risk
- RSL (Really Simple Licensing)
- Shadow IT
- Single source of truth
- Single-vendor dependency
- Software exit strategy
- Software license
- Spreadsheet risk
- Tacit knowledge
- Vendor lock-in
Security and cybersecurity · 65
Back to filtersHow data and AI systems get attacked, and how they are defended.
- 3-2-1 backup
- Abliteration
- Agentic AI threat model
- Agentic Browser
- Agentic social engineering
- Agentic Trust Framework (ATF)
- AI agent sandboxing
- AI BOM (AI Bill of Materials)
- AI jailbreak
- AI kill switch
- AI red teaming
- Attack Surface Management
- Business continuity vs disaster recovery
- Business email compromise (BEC)
- C2PA
- CBOM (Cryptography Bill of Materials)
- Context poisoning
- Crypto-agility
- Cyber insurance
- Dark web
- Data breach
- Data Security
- Deepfake
- Digital forensics
- EDR, XDR and MDR
- Excessive Agency
- Firewall
- Guardrails
- IEC 62443
- Incident response
- Least privilege for AI agents
- Lethal trifecta
- Log management
- Man-in-the-middle attack (MitM)
- MCP tool poisoning
- Membership inference
- Model hub typosquatting
- Network segmentation
- Non-Human Identity (NHI)
- OIDC-A (OpenID Connect for Agents)
- OPSEC
- Password manager
- Patch management
- Phishing
- Phishing-resistant MFA
- Pickle vs safetensors
- Post-quantum cryptography (PQC)
- Prompt injection
- Quantum key distribution (QKD)
- Ransomware
- Restore test
- Reward hacking
- RTO and RPO
- Security awareness training
- security.txt
- Shadow AI
- SIEM
- SOC (Security Operations Center)
- SPF, DKIM and DMARC
- SSO (Single Sign-On)
- Tor and onion routing
- Uncensored models
- VPN (Virtual Private Network)
- Vulnerability assessment vs penetration test
- Zero Trust
Cloud, MLOps & DevOps · 47
Back to filtersThe operational practices that keep platforms, pipelines and costs standing.
- AI tokenomics
- AI-Assisted Testing / Agentic QA
- CI/CD
- Cloud migration strategies (the 7 Rs)
- Cloud repatriation
- Confidence Calibration
- Cross-validation
- Data augmentation
- Data leakage
- Data migration
- DataOps
- DevGenOps
- DevSecOps
- Edge AI
- Ensemble learning
- Eval alerting
- Experiment tracking
- Feature engineering
- FinOps
- GitOps
- Human-in-the-Loop (HITL)
- Infrastructure as Code (IaC)
- Intelligent Process Automation (IPA)
- LLM Determinism
- LLMOps
- MLOps
- Model card
- Model collapse
- Model monitoring
- Model registry
- ModelOps
- NAS vs cloud storage
- NPU
- Observability
- Overfitting
- Platform engineering
- Pose estimation
- Principal component analysis (PCA)
- RLCD vs RLHF
- ROUGE Score
- SDD & TDD
- Serverless vs On-prem vs Edge
- Sports computer vision
- Temperature scaling
- Tensor parallelism vs pipeline parallelism
- Tokenizer fairness
- Vibe coding
Architecture and engineering practices · 38
Back to filtersArchitecture patterns, testing and legacy modernization.
- Access Database Migration
- Agent Experience (AX)
- AI accelerator lock-in
- Anti-corruption layer
- API (Application Programming Interface)
- API gateway
- AS/400 Modernization
- Canary release
- CAP theorem
- Chaos engineering
- Circuit breaker
- Code refactoring
- Contract testing
- CQRS
- ERP Migration
- Event sourcing
- FaaS
- Feature flag
- Garbage collection
- Greenfield vs brownfield
- Hyperautomation
- Industry 4.0
- Interoperability
- Legacy system
- LSP (Language Server Protocol)
- Mainframe and COBOL modernization
- Malleable software
- Microservices vs monolith
- Performance profiling
- RPA (robotic process automation)
- Rules engine
- Saga pattern
- Service mesh
- Shift Left Testing
- SLI, SLO, and SLA
- Strangler fig pattern
- Test automation pyramid
- Throughput vs latency
Visibility in search and AI · 17
Back to filtersHow you get found and cited: search engines, generative answers and authority.
Websites and digital products · 18
Back to filtersHow a website is planned, rebuilt and maintained, and what it really costs.
- CMS (Content Management System)
- Core Web Vitals
- Design system
- DNS (Domain Name System)
- Headless CMS
- HSTS (HTTP Strict Transport Security)
- Landing page
- PWA (Progressive Web App)
- Responsive design
- SEO site migration
- SSL/TLS certificate
- Static site generation
- UX design
- WCAG (Web Content Accessibility Guidelines)
- Web hosting
- Website cost
- Website maintenance
- Website redesign
Regulation and compliance · 71
Back to filtersThe regulations touching data, AI and cloud, and what they mean in practice.
- Accessibility overlay
- ACN Cloud Qualification (QC/QI)
- AI Act
- AI Act Compliance as Code
- AI Act enforcement on GPAI models
- AI criminal liability in Italy (Law 132/2025)
- AI literacy (Art. 4 AI Act)
- AI music licensing for enterprise
- AI text watermarking
- AI TRiSM
- Algorithmic admissions
- Algorithmic dynamic pricing
- Algorithmic hiring
- Anonymization vs pseudonymization
- Athlete data governance
- BCBS 239
- CAM ICT (Italian Minimum Environmental Criteria)
- CLOUD Act
- Cloud and AI Development Act (CADA)
- Content fingerprinting
- Cookie consent
- CSRD (Corporate Sustainability Reporting Directive)
- Cyber Resilience Act (CRA)
- D.Lgs. 160/2026 (first AI Act implementing decree)
- Dark pattern
- Data Act
- Data breach notification
- Data Controller vs Data Processor
- Data Governance Act
- Data residency
- Deployer-to-provider requalification (AI Act Article 25)
- Digital Product Passport (DPP)
- Digital Services Act
- Digital sovereignty
- DORA (Digital Operational Resilience Act)
- E-invoicing
- EDPB web scraping guidelines
- ESG Rating
- EU Machinery Regulation (EU 2023/1230)
- European Accessibility Act (EAA)
- Explainable AI (XAI)
- FRIA (Art. 27 AI Act)
- GDPR and Artificial Intelligence
- Generative content copyright due diligence
- GPAI
- Green IT and Digital Sustainability
- Iperammortamento 2026 (Italian tax incentive)
- ISO/IEC 27001
- ISO/IEC 27701
- ISO/IEC 42001
- Italian implementation of the AI Act
- Italy's National Cybersecurity Perimeter (Perimetro di Sicurezza Nazionale Cibernetica)
- Legal Digital Preservation (Italy)
- Multi-cloud and hybrid cloud
- NIS2
- PII (Personally Identifiable Information)
- Prompt ownership
- Public Procurement Act (proposed EU regulation)
- Quantum computing
- Responsible AI
- SBOM (Software Bill of Materials)
- Secure by design
- Sovereign AI
- Sovereign cloud
- Startup innovativa (Italian innovative startup)
- Student data governance
- Synthetic content disclosure (AI Act Article 50)
- UNI 11621-8
- Value chain cap (the CSRD ceiling on supplier data)
- Virtual influencer disclosure
- Zero data retention
Method and consulting · 51
Back to filtersHow a project gets delivered: delivery, management, roles and engagement models.
- A/B test
- Agent manager
- Agent washing
- Agentic maturity
- Agile
- AI business case and ROI
- AI code smell
- AI Engineer
- AI execution throughput
- AI maturity assessment
- AI project failure causes
- AI use case selection
- Architecture decision record
- Blameless postmortem
- BPMN
- Build vs buy
- Capability map
- Change management (technical)
- Cognitive debt
- Conway's law
- Corporate adaptive learning
- Data debt
- Data Scientist vs Data Engineer vs Data Analyst
- Data strategy
- Digital transformation
- Enshittification
- Extreme Programming
- Forward Deployed Engineer (FDE)
- Fractional CTO / CDO / CIO
- Fuzzing
- ISV (independent software vendor)
- ITSM (IT Service Management)
- Kanban
- Lean software development
- Linting
- PoC, prototype and MVP
- RFI vs RFP vs RFQ
- Sales engineer
- Scrum
- Service-as-Software (SaS)
- Software 3.0
- Sports analytics as a service
- Staff augmentation (Italian "body rental")
- Staff engineer
- Support engineer
- Team Topologies
- Technical debt
- Time & Material vs Fixed Price
- Tokenmaxxing
- Watermelon effect
- XaaS (Anything as a Service)
Deciding under uncertainty · 52
Back to filtersHow you decide when the data is not enough: cognitive biases, forecasting and statistics.
- "If you can't measure it, you can't manage it"
- AI anthropomorphism
- Algorithm aversion
- Anchoring bias
- Apophenia
- Automation bias
- Base rate fallacy
- Bitter lesson
- Brier score
- Butterfly effect
- Bystander effect
- Cargo cult
- Casino effect
- Chesterton's fence
- Cognitive bias
- Cognitive dissonance
- Confirmation bias
- Conformal prediction
- Criti-hype
- Decoy effect
- Delphi method
- Dunning-Kruger effect
- Feynman Technique
- First principles
- Goodhart's law
- Grosch's law
- IKEA effect
- Information cascade
- Jevons paradox
- Koomey's law
- Lewin's model (unfreeze-change-refreeze)
- Monte Carlo simulation
- Moore's law
- Normalization of deviance
- Occam's razor
- Pareto principle
- Planning fallacy
- Prediction market
- Psychological pricing
- Reference class forecasting
- Ringelmann effect
- Salesforce vs Dynamics 365
- Selective prediction
- Simulated annealing
- Skin in the game
- Statistical significance
- Status quo bias
- Sunk cost fallacy
- Superforecasting
- Survivorship bias
- TCO (Total Cost of Ownership)
- Wisdom of the crowd
A term that hits close to home? Write to me.
CONTACT ME