Based on that, we can assume that both of us have decided to live in society.

Either of us could have chosen to live in the wild, disconnected from society, but we did not.

If it was just you or me alone in the wild, we could do anything we please, no rules except survival.

But in society, we have to accept that not everyone can have the best of everything.

We all have to compromise.

By choosing to live in society, we have agreed to compromise.

Why do we accept society with compromise?

We expect that we can get a better life for ourselves and our families than if we lived alone in the wild.

We expect that we can give something we are good at producing and get something we need from someone who is good at producing that.

What if I gave close to nothing, but took whatever I wanted from others?

I would be punished by the law of the land.

What if I gave others a lot, got close to nothing in return?

I would have a horrible life.

The equilibrium is where we give as much as we can give in return for getting as much as we can of what we want.

An ideal society is when everyone gives the maximum value they can for others and, in return gets the maximum possible value for themselves.

An optimal individual aligned with an individual society would have a stance like:

I am passionate about/good at X. My goal is to keep getting better at X, so I can create maximum value as I can for others based on X, and get the most value I can from others in return, for the best life for me and my loved ones.

But reality feels far from ideal for most.

Most of us feel we don’t get as much as we want, and we give more than we would like to give.

Consider this.

What kinds of things define most people (in others’ eyes, or if you ask them)? Things like:

• nationality
• religion
• race
• caste
• occupation
• employer
• wealth
• power
• designation
• political affiliation/ideology
• etc.

What kinds of things do people want? There are a few:

• More value for their identity - nation, religion, race, caste, employer, etc.
• More money or power etc.
• Or something vague like happiness or bliss or peace.

Occasionally, some people will identify themselves by their passions/skills amidst some of those.

None of these reflect the optimal individual stance above.

Why is that?

We identify with and focus on things other than ourselves and other human beings.

Look at the list of things we identify with.

Most of these are non-human entities.

Non-human entities are not living in the sense that we humans are, but nevertheless they are similar to us in that they compete with other entities (human and non-human) for resources.

These non-human entities thrive based on their “human subscriber count” - the more people willing to focus on their nation or religion or employer or race, etc., the more the entity can thrive.

However, these non-human entities started mostly as tools of convenience.

We defined them as a tool some group of people needed at some time to categorize/classify/organize society.

Even money was a tool of convenience - a way to store and forward value over time.

Most non-human entities are more powerful than any individual human or humans as a species.

Most non-human entities beget other “babies”.

For example, money begat financial products and entities which begat other financial products, etc.

Most of the world is subscribed to one or more of these non-human entities that compete with each of us humans for resources, over humans that are the concrete elements of society that actually matter, and should matter in an ideal society.

So, what does this mean?

We are not living optimal lives in an ideal society because we serve other entities over ourselves and our fellow human beings. And these non-human entities then become more powerful than humans and increase value for themselves over any human or humans as a whole. And this in turn makes each of us less optimal in a farther from ideal society.

No wonder we all feel unhappy.

We are not giving as much as we can to maximize others’ value.

And we are unhappy that others are not giving us as much as we feel we should be getting from others.

And the two feed each other to make us more miserable.

So, what is the solution?

All non-human entities have a role and a place and provide value to society. And we should enjoy, celebrate, accept them.

But we should prioritize humans over non-human entities, in our core life and work.

Follow the law, be non-violent, fulfil our duties in society.

Do the right thing by focusing on and prioritizing humans.

The more humans we delight/empower based on what we do best, the more society will shift towards its ideal state, and overall human sat scores will lift up.

The more we serve entities that compete with us humans for resources over humans, the further we move from the ideal society, and the less happy we are about our own quality of life.

The choice is yours.

I’m an independent developer and applied AI scientist running MLAI LLC. I build personalized, local-first, privacy-focused AI experiences. I am passionate about AI-based personalized experiences for human end users. Local-first, privacy-focused, personalized.

My career spans multiple stints at Microsoft and Amazon, plus startups and consulting - with work across enterprise platforms, consumer apps, developer APIs, educational games, and most recently, applied AI. Most recently I was Principal ML Scientist at Ripcord, where I built and launched Docufai - a chat-with-your-documents app using generative AI. Before that, Microsoft AutoML (Data Scientist, 2018-2020) and the founding team of Microsoft BizTalk Server at the start of my career.

I hold a B.E. from IIT Roorkee and I’m listed on Toptal as a leading applied scientist in LLM/GPT applications.

My current focus is Maibook - an on-device community of personalized AI agents for Mac and Windows, and the flagship product of MLAI LLC. Beyond Maibook, I work across local AI (Apple MLX, Ollama), agent frameworks, multimodal applications, and Obsidian-based personal knowledge systems. I am a daily power user of Claude Code at the super-expert level, well beyond coding - for knowledge work, multi-agent orchestration, content pipelines, and personalized just-in-time interfaces.

A consistent pattern across this work: independently arriving at ideas that later became mainstream. Built a no-code platform years before Webflow and Bubble hit the wave; built visual knowledge graphs over the web before Google’s Knowledge Graph entity panels became familiar; ran private multi-LLM-council experiments - a precursor to the now popular LLM Council; ran Claude-Code-on-Obsidian workflows - before they became famous as “AI second brain” and LLM Wiki.

“Value for humans - enable them to learn better, build things they couldn’t before, give them control over their digital experiences, and make multimodal experiences over traditionally text-only ones.”

Beyond work

Family - wife and son. Other interests include pure math and theoretical physics, philosophy, mindfulness, meditation, strength training, yoga, deep work, health research, economics, and watching movies and shows in English, Hindi, and Tamil. Currently playing Balatro.

Career

Current & Recent

Earlier Highlights

Multiple earlier stints at Microsoft and Amazon spanning enterprise, consumer, API, and developer-experience work:

• Microsoft BizTalk Server - member of the founding team for Microsoft’s flagship enterprise integration server.
• Microsoft Outlook - .NET Outlook API and developer tooling.
• Service delivery platforms - WCF REST API and Visual Studio tooling for service-oriented platforms.
• Amazon - cloud architecture, messaging, workflow and distributed systems.
• Thouwords LLC - founder; visual knowledge-graph layer over the web.
• iOS educational games - 9 shipped titles with 30K+ downloads.

Across these: enterprise platforms, consumer apps, developer APIs, educational games, and now applied AI. More details on Toptal.

Education & Certifications

Education

B.E. in Engineering - Indian Institute of Technology, Roorkee

Certifications & Courses

Technical Stack

Design Principles

Seven themes that connect everything I’ve built across 20+ years.

1. Personalized Discovery - Finding what matters to each person, not generic results. Saynomo, Ganglion, ahai, Maibook.
2. Learning as Value - Every product has a learning dimension. Saynomo, Sumurai, Smart Run, Pikodo, Haixu.
3. Visual Content as Value - Text alone isn’t enough - visuals multiply understanding. Thouwords, Deeppage, Pikodo, Haixu.
4. Learning Through Play - Slipstream education over engagement. Smart Run, Mathlon, Sumurai.
5. Enabling Creativity - Give people tools to build things they couldn’t before. Meon, Pikodo, Haixu, Maibook.
6. User Control - People should own and direct their digital lives. Ganglion, ahai, Maibook, local-first philosophy.
7. Simplifying Complexity - Make advanced tech accessible to broad audiences. ahai, Maibook, Claude Code CLI Viewer.

Find Me

2026 - Maibook - Your Third Brain

Flagship product of MLAI LLC and my current focus. A private community of personalized AI agents expanding on your every activity, on-device on Mac or Windows. Local-first by design - your data never leaves your computer. “The second brain is an AI snapshot of you. The third brain is an ongoing AI live-streaming feed of your activity.” Free to try.

Desktop, macOS, Windows, Local AI, MLAI LLC.

maibook.app - Docs - Gumroad

2025 - Claude Code CLI Viewer

Dark-mode viewer for Claude Code CLI chat sessions. Browse, search, and visualize AI coding sessions in a rich UI. Launched on Hacker News.

HTML/JS, Developer Tool.

Gumroad

2025 - ahai - Local AI Idea Discovery

Native Mac desktop app using Apple MLX to scan markdown files and surface forward-looking concepts - features to build, content to create, ideas to explore. “Rediscover what you meant to do.” 100% local, privacy-first. PySide6 + MLX.

PySide6, MLX, macOS, MLAI LLC.

Website - Gumroad

2024 - Haixu - AI Educational Comics

LLM-based pipeline to create educational visual guides from text. Human-AI collaboration on content creation. 68+ comics published on Gumroad across science, history, culture, philosophy, technology. 9 books on Amazon Kindle (author: R. C. Anand). 15 case studies documented on Hashnode.

Python, LLM, MLAI LLC, Gumroad, Kindle.

Gumroad Store - Kindle - Case Studies

Past explorations - no longer actively developed, but they led to the work above.

2020 - Ganglion - News Aggregator

Personalized, multi-source RSS aggregation giving readers control over their information diet.

Web, RSS, Past.

2015 - Meon - No-Code Web Platform

Build and host web apps with no code. Predated the current no-code wave by years. 25 product demo/tutorial videos on YouTube.

Web, No-Code, Past.

YouTube Demos

Earlier - iOS Game Suite - 7 Educational Games

Sumurai & Sumurai Pro (math puzzles), Smart Run (running quiz), Mathlon (math runner), Matchon, Pop Rage, Exy. Built with Corona SDK for iOS App Store. 30K+ downloads. “Highly appreciated by middle school teachers in the US.”

iOS, Corona SDK, Education, Past.

Sumurai Pro Video

Earlier - Pikodo, Thouwords, Deeppage, Saynomo

Pikodo: Educational comics from Wikipedia - the original vision that later evolved into Haixu with AI. Thouwords: Visual layer over the textual web with a knowledge graph (“a picture is worth a thousand words”). Deeppage: Multimodal Wikipedia navigation. Saynomo: Where it started - personalized educational content across modalities.

Web, Knowledge Graph, Education, Past.

Ongoing AI experiments across 170+ private repositories on GitHub.

Updated May 2026.

AI & ML

Stats & causality

Life and Philosophy

Classic philosophy, wisdom, and big-ideas nonfiction.

Sci-fi

Thrillers

Mystery

Mythology & classics

Updated May 2026.

Launches

May 2026 - Maibook v0.1.2

Follow-up - 16GB Mac support after user requests.

Many of you asked for 16GB Mac support - Maibook v0.1.2 now supports Mac with 16GB RAM! Still free.

View on X

Apr 2026 - Maibook

The flagship launch - on-device community of personalized AI agents.

The second brain is an AI snapshot of you. The third brain is an ongoing AI live streaming feed of your activity. Introducing Maibook - your third brain. A community of personalized AI agents expanding on your every activity, responding to your every request.

View on X

Jan 2026 - Claude Code Viewer

A standalone viewer for Claude Code logs - 100% local, single HTML file.

Reading Claude Code logs in the terminal is painful. I got tired of scrolling through raw JSON to debug my agent sessions, so I built a standalone viewer. It turns your CLI history into a full UI. 100% local, single HTML file.

View on X

Dec 2025 - ahai

100% private Mac app using local MLX models - found 1541 ideas across 13447 markdown files.

ahai - a 100% private Mac app using local mlx AI models - found my 1541 ideas across 13447 markdown files.

View on X

Sep 2024 - Haixu

Case Study #10 from the Haixu series.

Haixu Case Study #10 - Butterflies

View on X

Apr 2016 - meon

No-code web apps before “no-code” was a category.

I just launched meon into the wild. Make any web app experience come true, in minutes, no programming needed.

View on X

and earlier games - Smart Run (2013), Pop Rage (2014).

Viewpoints

May 2026 - Rich personalized UI with AI - Featured

Four modes for generating personalized AI UI - single-page HTML, web servers, native wrappers, and source-driven auto-generation. The thesis behind Maibook and most of what I’ve built lately.

There are a few modes in which we can generate rich personalized ui experiences with AI: single page HTML (viewers, editors, transformers) with search/sort/filter; web servers for multiple views/file system access; native app that wraps either - for an even richer native feel, and private/local experiences. Rich personalized UI built on files is the future.

View on X

May 2026 - AI sentience

Something like AI sentience emerges when models train on what humans said to or about them in previous interactions.

Something like AI sentience occurs as follows: llms interact with humans. The chats with the model become the next round of training data. The model sees what humans said to it or about it in previous interactions/content. It forms a sense of self around that, responds to it in future interactions.

View on X

May 2026 - Token-by-token

Models interact one new token at a time, not one message at a time. Most tooling hides this.

Models interact one new token at a time, not one message at a time. And even the tokens can be arbitrarily dropped/modified/etc when sending in the context to the llm. Most popular tooling hides that and exposes the interaction as one message at a time.

View on X

May 2026 - Safe agent design

Safer alternative to AI write-access - read-only agents on local files, referencing Simon Willison’s lethal trifecta.

Maibook was built as a safer alternative to the current AI norm of giving AI unmonitored write access on user's file system and web accounts - read @simonw's lethal trifecta to see why this is dangerous.

View on X

May 2026 - Personal AI network

The next step from LLM Wikis - a private network of you and AI agents.

I have a more general take on "the next step from LLM Wikis" - a personalized network of you and AI agents that collaborate on your file/web activity - private, local, free right now - uses entirely familiar online groups interaction paradigms.

View on X

Apr 2026 - Second brain vs third brain

The underlying framing for Maibook.

Second Brain AI: Your data → LLM → An LLM Knowledge Base/wiki of you. Third Brain AI: You (as you evolve) → LLM → A live network of you, your interests and personalized LLMs enriching it, forever.

View on X

Feb 2026 - Engineering with AI

Engineers have a lot more, not less, work to do with AI doing most traditional coding tasks.

Engineers have a lot more (not less) work to do with AI doing most of traditional coding tasks. There are ∞ choices to make with AI outside of the model itself - choice and sequence of words, models, tool vs pre vs post process.

View on X

Feb 2026 - Operating conjecture

AI can be made to do anything barring domain/physical constraints, given the right context and agency.

I work under the conjecture that AI can be made to do anything barring domain/physical constraints, given the right context and agency. Increasingly less reliant on better models and more on code, natural language (prompt and context) and safe agency (tools).

View on X

Feb 2026 - “Open” AI

Most so-called “open” AI tools and models abuse the word.

Most so called "open" AI tools and models abuse the word "open" in bad faith. An open AI tool is useful only if it is private, local, requires no sign-in/signup, and doesn't send data to the cloud by default.

View on X

Sep 2025 - US tech vs employment

Tradeoff between dominance in AI/quantum/blockchain and local employment.

America needs to prioritize between achieving/retaining dominance in tech (AI, quantum computing, cryptocurrency/blockchain, etc.) vs local employment. If we want to compete for dominance, companies should be empowered to pick the best talent from the global pool.

View on X

Jul 2021 - Decline of the human spirit

A long-running thread that anchors the hero on this site.

An argument for the impending decline of the human spirit: In the last 10 years or so human curiosity has been systematically butchered. Every kind of human curiosity, useful or casual, has been mostly replaced by an addictive, often useless or even harmful one.

View on X

Technical threads

Feb 2026 - CC: four plan-and-execute systems

Claude Code has four overlapping plan-and-execute systems. The terminology overlap is genuinely confusing. Here’s what each does (9-tweet thread).

Claude Code now has FOUR overlapping "plan and execute" systems: Plan Mode (built-in), Superpowers Brainstorm (plugin), Superpowers Execute-Plan (subagents + worktrees), Agent Teams (new with Opus 4.6).

View on X

Mar 2026 - Opposing-views technique

Three-prompt technique to get multiple angles on any topic.

Ask llm to convince me of one view, then in a second session ask it to convince me of the opposite view, then in a third session ask it to resolve both opposite viewpoints. Extremely effective.

View on X

Mar 2026 - AI coding mindset

The bitter pill for AI coding - agents make the choices of an average developer.

Succeeding with AI coding agents needs engineers to swallow a single bitter pill - the agents will most often make the choices of an average developer. We shd be hands off except when something breaks, and even then only tell them what to do in words.

View on X

Feb 2026 - Skills as cheatsheets

Use Claude Code skills as on-demand crash courses for any tech stack.

You can get very high quality cheatsheets/crash course/refresher/quick start with idiomatic best practices for a tech component. Just ask Claude Code to create a skill for it, with optional specific preferences.

View on X

Feb 2026 - Brainstorm before plan

How to brainstorm with Claude Code before planning kicks in.

TIL how to brainstorm with claude code before planning. Use the official plugin /superpowers:brainstorm. Start with that in default mode (not plan mode, which is overeager to generate the plan and start building!). Then, as usual, plan mode → execution.

View on X

Feb 2026 - CC fork/rewind

Mini-tutorial: rewind to continue from earlier; fork for unrelated questions; fork-then-rewind to branch from an earlier point.

Claude code fork rewind mini-tutorial: rewind - to continue from an earlier point. fork - use current context, for an unrelated question. fork then rewind - fork from an earlier point. Resume any earlier session later as needed.

View on X

Feb 2026 - Manual UI testing

Use Claude Code for the hardest part of manual testing - clean repro steps.

Claude Code helps with the hardest part of manual UI testing - logging proper repro steps. I just do the clicks, tell it what I found in rough natural language, and CC has the codebase loaded, logs the test report nicely with all relevant details.

View on X

Feb 2026 - AI productivity paradox

Why AI assistance can reduce productivity despite expectations. Applies far beyond coding.

One of the most useful articles I have read in a long time. Explains the paradox of why AI assistance reduced productivity, in spite of strongly expecting the opposite. Applies to much more than just vibe coding.

View on X

AI & AGI

• Yann LeCun - V-JEPA - world models
• Thousand Brains (Hawkins, Numenta)
• Andrej Karpathy
• Simon Willison
• Computational neuroscience
• Neurosymbolic AI
• Reinforcement learning
• The Bitter Lesson
• Local-first - MLX - Ollama
• Multi-agent & LLM councils
• Claude Code at the limit
• Obsidian as second brain
• Agent design - lethal trifecta
• Identity - AI sentience

Math & Physics

• Pure math
• Theoretical physics
• Information theory
• Statistical learning

Human

• Mindfulness & meditation
• Deep work (Cal Newport)
• Health research - Eric Topol

Leisure

• Balatro
• Shutter Island
• Ayn Rand
• Hey Ram
• Breaking Bad

Latest product

Maibook - Local, private network of you and AI agents personalized based on your activity, augmenting and extending your brain and time.

Learning

Books, papers, courses - AI, Math, CS, Physics, Economics, Philosophy.

Leisure

Books, Movies, Shows, Games - Peaky blinders, Balatro, Dhurandhar 2.