Girish KotteBook a call
57 answers · 9 topics

The AI FAQ, from someone who builds.

Clear answers to the most common questions about AI, machine learning, LLMs and using AI for business growth, written by an AI architect who has built 4 startups.

Popular:
One source of truth
faq/data.ts57 answers, written once9 topics, each with its own description
This hub/faq
SearchFilters all 57
SchemaFAQPage JSON-LD
What search engines read
Every answer on this page is also in its FAQPage schema.
Topic 1 of 9

AI fundamentals

Core concepts and terminology in artificial intelligence

7 questions

Artificial intelligence is the field of computer science that builds systems capable of performing tasks that typically require human intelligence — understanding language, recognizing images, making decisions, and generating content. Modern AI is powered by machine learning, where algorithms learn patterns from massive datasets rather than following hand-coded rules. AI today ranges from narrow systems built for a single task (like spam filtering) to large language models that can write, reason, and code. It already powers search engines, medical diagnostics, autonomous vehicles, and the tools millions of knowledge workers use daily.

AI is the broadest term — any system that mimics human intelligence. Machine learning (ML) is a subset of AI where systems learn from data instead of being explicitly programmed. Deep learning is a subset of ML that uses neural networks with many layers to handle complex tasks like image recognition and natural language processing. Think of it as nested circles: all deep learning is machine learning, and all machine learning is AI, but not the other way around. In practice, most modern AI products use deep learning models, especially transformer-based architectures that power tools like ChatGPT and Claude.

Generative AI refers to artificial intelligence systems that create new content — text, images, code, audio, or video — rather than simply analyzing or classifying existing data. These systems learn patterns from training data and use that knowledge to produce original outputs. The generative AI revolution began with transformer models in 2017 and accelerated with ChatGPT's launch in late 2022. Today, generative AI can write marketing copy, generate photorealistic images, compose music, and build entire applications from natural language descriptions.

Large language models are AI systems trained on vast amounts of text data to understand and generate human language. They work by predicting the most likely next word in a sequence, but this simple mechanism produces remarkably sophisticated reasoning, writing, and coding abilities. Leading LLMs include OpenAI's GPT-4o, Anthropic's Claude, Google's Gemini, and Meta's open-source LLaMA. LLMs power chatbots, code assistants, content generators, and customer service automation. The 'large' refers to their parameter count — modern LLMs have hundreds of billions of parameters.

Natural language processing is the branch of AI focused on enabling computers to understand, interpret, and generate human language. NLP powers everything from search engines and voice assistants to sentiment analysis and machine translation. Traditional NLP used rule-based approaches and statistical models, but modern NLP is dominated by transformer-based language models that understand context and nuance far better than earlier systems. Key NLP tasks include text classification, named entity recognition, question answering, summarization, and language translation.

A neural network is a computing system inspired by the human brain's structure, made up of interconnected nodes (neurons) organized in layers. Data flows through an input layer, one or more hidden layers, and an output layer, with each connection having a learnable weight. During training, the network adjusts these weights to minimize errors in its predictions. Deep neural networks — those with many hidden layers — can learn incredibly complex patterns, which is why they excel at tasks like image recognition, speech processing, and language understanding. Nearly all modern AI breakthroughs are built on neural networks.

Narrow AI (also called weak AI) is designed for a specific task — playing chess, filtering emails, or generating images. It can outperform humans at its specialized task but cannot transfer that ability to unrelated domains. Artificial general intelligence (AGI) is a theoretical system that could match or exceed human-level reasoning across all intellectual tasks. Every AI system in production today is narrow AI. AGI remains a research goal, with estimates for its arrival ranging from a few years to several decades, depending on whom you ask. The path to AGI is one of the most debated topics in computer science.

Topic 8 of 9

AIO & GEO optimization

How to rank in AI Overviews and optimize for generative engines

6 questions

AIO (AI Overview Optimization) is the practice of structuring website content to appear in Google's AI Overviews — the AI-generated summary boxes that appear at the top of search results for informational queries. Unlike traditional SEO which targets blue-link rankings, AIO focuses on making your content the source that Google's AI cites when generating answers. Key strategies include using structured data (FAQ schema, HowTo schema), writing concise authoritative answers in the first paragraph of each section, building topical authority through comprehensive content coverage, and ensuring your site is crawlable by both traditional and AI bots.

GEO (Generative Engine Optimization) is the practice of optimizing content to be cited and referenced by AI-powered search engines and chatbots — including ChatGPT, Perplexity, Claude, and Google's AI Overviews. While traditional SEO focuses on ranking as links in search results, GEO focuses on making your content the source that AI models draw from when answering user questions. Key GEO strategies include publishing authoritative factual content, using structured data and schema markup, building citations and brand mentions across the web, maintaining an llms.txt file, and directly answering common questions in your niche with unique expert insights.

Google AI Overviews use large language models to synthesize information from multiple web sources and present a summarized answer at the top of search results. When a user searches an informational query, Google's AI reads content from high-authority pages, extracts relevant passages, and generates a cohesive answer with source citations. Sites that get cited typically have strong domain authority, well-structured content with clear headings and lists, comprehensive schema markup, and content that directly answers the user's query. As of 2026, roughly 47% of Google searches trigger AI Overviews.

Traditional SEO focuses on ranking your page as a blue link in search results through backlinks, keywords, and technical optimization. AIO (AI Overview Optimization) focuses specifically on getting cited in Google's AI-generated summaries at the top of search results. GEO (Generative Engine Optimization) is the broadest term — it covers optimization for all AI-powered answer engines including ChatGPT, Perplexity, Claude, and Google AI Overviews. All three share common foundations (authoritative content, structured data, strong E-E-A-T), but GEO and AIO place more emphasis on answer-ready content structure and AI crawler accessibility.

To get cited in ChatGPT and Perplexity answers, focus on four strategies: (1) Publish authoritative, factual content with unique data, expert insights, and clear sourcing. (2) Build a strong web presence with brand mentions, backlinks from authoritative sites, and Wikipedia references if eligible. (3) Add an llms.txt file to your site that tells AI crawlers about your content structure and key pages. (4) Allow AI crawlers in your robots.txt — don't block GPTBot, ClaudeBot, PerplexityBot, or other AI user agents. Content that gets cited most frequently is well-structured, factually dense, regularly updated, and comes from recognized domain experts.

The most effective schema types for AI Overview visibility are: FAQPage (structured Q&A that AI can directly extract), HowTo (step-by-step instructions), Article/BlogPosting (identifies author, publish date, and content type for E-E-A-T signals), Speakable (tells AI which content sections are best for citation), Organization and Person (establishes entity authority), and BreadcrumbList (helps AI understand site structure). Implementing multiple schema types on the same page is recommended — for example, an article page should have Article schema, FAQPage schema for any Q&A sections, and Speakable schema for key paragraphs.

More topics

Browse the rest.

7 topics
6 questions

AI tools & models

Understanding popular AI platforms, models, and APIs

7 questions

AI for business

How companies are using AI to grow, automate, and compete

7 questions

Building AI products

Technical concepts for developers and product builders

7 questions

AI marketing & GTM

Using AI for marketing, SEO, and go-to-market strategy

6 questions

AI in healthcare

How artificial intelligence is transforming healthcare and medicine

6 questions

AI for startups & founders

Practical advice for building and scaling AI startups

5 questions

AI ethics & future

Responsible AI development, risks, and what's ahead

Not answered here?

Ask it on a call

If your question is not in the 57, bring it to a free 30-minute call.

Book a call
Next step

Still have questions?

Ask on a 30-minute call, or take the AI scorecard to see where you stand.