MindForge AI vs GenAI Lab

MindForge AI makes you fluent in AI. Tokens, embeddings, context windows, RAG, the agent loop — the vocabulary and mental models that let you read a real tutorial, follow a real repo, and stop drowning in a field that reinvents itself monthly. Five minutes a day. EVERY PATH ENDS BY SENDING YOU SOMEWHERE ELSE Finish a path and it hands you a concrete thing to go build — in a real editor, on your own machine, with a time estimate attached. Not "keep learning!" — an actual assignment. This app is the warm-up and the memory layer; the real learning happens in the editor. FOUR PATHS, ONE ARC • AI Fundamentals — what these things actually are AI, ML & Deep Learning • LLMs & Prompt Engineering — how to talk to them LLM Foundations · Prompt Patterns That Work • RAG & Embeddings — how to give them your data Embeddings Explained · Your First RAG Pipeline • AI Agents — how to let them act The Agent Loop Six courses. Bite-size lessons that teach exactly one idea and bust one misconception beginners actually have — "RAG teaches the model" (it doesn't), "the model runs your tools" (it doesn't). Interactive quizzes build toward boss exams; pass one and earn a certificate with a completion code. EXPLAIN — A PLAIN-ENGLISH CONCEPT LIBRARY Nine core ideas — LLMs, neural networks, transformers, hallucination, embeddings, RAG, fine-tuning, agents, and MCP — each written out in plain English, browsable and searchable, with a "go deeper" technical version for when you want it. Hand-written, not generated. No chat, no model to wait on, no quota — free to read. READ CODE WITHOUT FEAR A built-in playground and a library of challenges with test cases. Run it, break it, fix it, see what the error actually means. The goal is literacy — being able to open a real repo and know what you are looking at. REMEMBER WHAT YOU LEARN Spaced-repetition flashcards (the same SM-2 method behind Anki) resurface concepts right before you'd forget them. Fluency is a memory problem, and this is the part of learning an app is genuinely good at. LEARNING THAT STICKS Streaks, XP, levels, daily quests, achievements, and leaderboards turn five minutes a day into a habit you keep. And Embry — the small anxious robot who runs the place — will have feelings about your streak. WHAT YOU GET, FREE Three lessons and fifteen code runs every day, plus the entire Explain library — free to read. The app is ad-supported and runs on your device — no account, no sign-up, no backend. Get fluent — then go build the thing. Terms of Use (EULA): https://www.apple.com/legal/internet-services/itunes/dev/stdeula/ Privacy Policy: https://helionixlabs.com/mindforge/privacy

GenAI Lab is an offline, hands-on course for understanding and prototyping generative AI systems, one experiment at a time. Learn how modern AI actually works by exploring it visually and interactively, right on your iPhone. No internet, no account, no tracking. A GUIDED CURRICULUM The course spans 8 tracks and 52 short missions covering: - GenAI foundations - Transformers and LLMs - Prompt engineering - Retrieval-augmented generation (RAG) - Image generation - Fine-tuning - Evaluation and safety - Multimodal AI and agents Every mission combines a concept check, a matching activity, and a guided Python code fragment so you learn by doing, not just reading. TRANSFORMER ARCADE Six short visual games and six boss challenges bring the internals of transformers to life: - Tokenization - Embeddings - Attention - Multi-head attention - Transformer blocks - Next-token generation Arcade rounds are untimed, offline, replayable, and use deterministic simulations so you can inspect exactly what is happening. LEARN BY EXPERIMENT The built-in simulators use deliberately small, deterministic examples so you can inspect attention weights, prompts, retrieval, diffusion controls, and agent loops step by step. Code activities validate guided fragments to build your intuition; the app does not execute arbitrary code. PRIVATE BY DESIGN Your progress is stored only on your device. GenAI Lab has no account, no ads, no analytics SDK, and makes no network requests. Learn anywhere, even completely offline. WHO IT IS FOR - Students and self-learners exploring how generative AI works - Engineers who want a clear mental model of transformers, LLMs, RAG, and agents - Anyone curious about the ideas behind modern AI, in a focused, no-pressure environment Questions or feedback? Email kshravi86@gmail.com.