Skip to main contentSkip to navigation
ThisIsHowItWorks.in

Complex systems, clearly explained.

An independent visual publication explaining the invisible protocols, networks, infrastructure, and mechanisms that run our world.

Explainers

  • How UPI Works
  • Offline UPI Mechanisms
  • All Explainers (Archive)
  • Topics & Roadmap
  • Search Index

Publication

  • About Publication
  • Editorial Principles
  • Changelog
  • RSS / Atom Feed

Legal & Contact

  • Privacy Policy
  • Terms of Use
  • Editorial & Legal Notice
  • Contact Us

Connect

  • Instagram
  • Discord Community
© 2026 ThisIsHowItWorks.in. All rights reserved.
Durable technical understanding built from first principles.
ThisIsHowItWorks.in
ExploreTopicsAbout
  1. Home
  2. /Topics
  3. /Computing & Artificial Intelligence
Artificial IntelligenceTopic Knowledge Hub

Computing & Artificial Intelligence

From foundational machine learning to modern generative architectures: self-attention, token sampling, and autonomous agent loops.

6 Published Explainers
22 Systems Analyzed
40 Core Concepts
1 Inquiry Path
Recommended Starting Point

How AI Agents Work

The architectural loop behind autonomous planning, tool execution, and stateful problem solving

Core Question Answered
“What actually happens behind the scenes when an AI agent solves a problem on its own?”
Agent RuntimeReAct PatternFunction CallingWorking Memory Context
Read Explainer

Inquiry Learning Paths

Structured sequences designed to take you from foundational mechanics to complex interactions.

From Next-Token Prediction to Autonomous AI Agents

A step-by-step inquiry path tracing how statistical word prediction scales into goal-directed agents that use tools in the real world.

1

How Large Language Models Generate Text

Foundational mechanism: Next-token autoregressive generation

2

Why AI Chatbots Sometimes Make Things Up

Failure mode: Statistical plausibility vs. factual grounding

3

How AI Agents Work

Autonomous architecture: Closing the ReAct tool execution loop

4

How an AI Agent Decides Which Tool to Use

Decision layer: Schema matching and constrained token decoding

All Published Explainers

Every verified first-principles analysis in this domain.

Explainer

How AI Agents Work

The architectural loop behind autonomous planning, tool execution, and stateful problem solving

First-Principles ExplainerRead
Explainer

How Large Language Models Generate Text

From raw prompt strings to token embeddings, self-attention calculations, and next-token probability sampling

First-Principles ExplainerRead
Explainer

How Search Engines Actually Work

Web crawling, inverted indexing, PageRank link graphs, and neural vector retrieval across 50 billion pages

First-Principles ExplainerRead
Explainer

How an AI Agent Decides Which Tool to Use

The mechanics of schema matching, token probabilities, and semantic tool routing

First-Principles ExplainerRead
Explainer

Why AI Chatbots Sometimes Make Things Up

The mechanics of hallucination, probability vs. truth, and the engineering behind grounded AI responses

First-Principles ExplainerRead
Explainer

Why Search Results Differ Between People

Geographic IP Anycast routing, session history vectors, freshness ranking, and A/B experiment flight allocation

First-Principles ExplainerRead

Inquiry Roadmap & Research Pipeline

Next-order causal questions in this discipline currently undergoing source verification and mechanism synthesis.

2 Queued Inquiries
Upcoming InquirySource Verification

“Why Neural Networks Were Invented”

Investigate the conceptual origins and developmental history of artificial neural networks, examining early bio-inspired computational models (McCulloch-Pitts, Rosenblatt), linear separability constraints identified by Minsky and Papert, and the emergence of multi-layer architectures trained via gradient-based backpropagation.

McCulloch-Pitts artificial neuronRosenblatt perceptron weightsMinsky-Papert XOR linear separability crisis
Upcoming InquirySource Verification

“How Transformers Changed AI”

Examine the architecture of the Transformer model (Vaswani et al.), analyzing how the self-attention mechanism replaced recurrence and convolution, enabling parallel computation over sequence contexts and altering scaling behavior in deep learning.

self-attention mechanismquery key value dot productselimination of sequential recurrence

Technical Systems & Protocols Analyzed

Hardware, protocol switches, and central clearing houses examined in this hub.

Large Language ModelSandbox EnvironmentTool API GatewayTransformer Neural NetworkTokenizer BPE EngineInference Server (KV Cache)Distributed Web Crawlers (Googlebot)Inverted Index ShardsDocument Forward IndexLink Graph EngineServing Cache & Query AggregatorCross-Encoder Neural Re-rankerFunction RegistrySemantic RouterVector DatabaseRAG PipelineRLHF Reward ModelGeolocation Edge ResolverShort-Term Session Memory StoreReal-Time Freshness Ingestion StreamExperimentation & Traffic Splitting FrameworkMobile-First Rendering Pipeline
← All Topics & DisciplinesHome Gateway →