This AI Agent Completes Your To-Do List (Plus 4 AI Tools That'll Blow You Away)

5 AI Tools That Make You Feel Like You're in 2050 by My First Million (Sam Parr and Shaan Puri), 56 minutes.

A guy built a custom clothing size app with Claude Code that tells him exactly what shirt size to buy from any store on Earth, and he did it in a single weekend for 40 bucks. That might sound niche, but it represents something enormous happening right now: the death of one-size-fits-all software and the birth of software built for a single human being.

Sam Parr and Shaan Puri sat down to play a game they called Blow My Mind with AI, each bringing their favorite tools that feel genuinely futuristic. Three big ideas emerge from this conversation that will change how you think about work, creativity, and what software even means anymore.

The Agent That Actually Works

Shaan opens with a rant that will resonate with anyone who has tried and failed to get an AI agent to do something useful. He had been on a crusade to find one agent, just one, that could do real work without constant babysitting. He found Do Anything dot com, built by a guy named Garrett who also runs an autonomous underground delivery company in Austin.

Here is what blew his mind. He typed a single lazy sentence asking it to analyze the My First Million YouTube channel like a world-class content strategist. He did not even bother giving it the channel link. Within minutes, it came back with a state of the union report that cut through vanity metrics and identified that only three to five percent of their 869,000 subscribers were clicking on new uploads. It identified outlier episodes and explained why specific titles worked. Then it generated a full one-month content plan with video essay formats, thumbnail strategies, and actual video titles like How to Build a One-Person Million-Dollar Portfolio in 2026.

Shaan said what excited him most was not this specific tool but the direction everything is heading. He pointed to ChatGPT's new Pulse feature, which proactively researches things you care about without being asked, and a tool called Nebula from his friend Furkon that reads your Slack, Gmail, and calendar to automatically create meeting prep docs and surface important changes. The key insight is that we are moving from a world where you prompt AI to a world where AI figures out what needs to be done by reading what is going on around you.

Your Voice Is Now Your Keyboard

Both hosts lit up when talking about Whisper Flow, which Sam described as the app that made him barely type anymore. It works across every app on your phone, not just inside ChatGPT or one specific tool. You hit a button and start talking, and it captures everything with perfect formatting.

What makes this different from basic speech-to-text is that it learns your personal style. Sam noticed it started formatting his messages the way he naturally writes, breaking up text, removing periods that feel too formal in texts, and adding bullet points when he lists things. If he says actually scratch that while rambling, it edits on the fly.

Shaan described a workflow that sounds like it belongs in a science fiction novel. He goes for a walk with his dog, pops in an AirPod, and just talks for 30 minutes, rambling about ideas for a blog post or book chapter. Then he hands the raw transcript to AI and says remove the rambly parts and structure this into an actual book chapter. He gets back something that is 85 to 90 percent finished without ever looking at a screen. His dad would not even recognize it as work.

Notebook LM Turns Podcasts Into Presentations

Shaan dropped what might be the most underrated AI demo of the episode. He took a single YouTube link to a podcast featuring Andrej Karpathy, the former head of AI at Tesla and one of the best thinkers about AI alive, and dropped it into Google Notebook LM. Without watching a single second of the hour-and-a-half podcast, he clicked one button and it generated a complete slide deck.

The result was a fully structured presentation called Summoning Ghosts: The Decade of AI Agents, with auto-generated graphics, quotes pulled from the conversation, and organized sections. Sam's jaw visibly dropped. The images were not gibberish. The slides were legible and coherent. Some were overly complicated, sure, but Shaan made the key point: this is so much better than listening to Ray Dalio talk for an hour. You get the gist, and then you might actually be motivated to listen to the full thing.

Vibe Coding a Basketball Anthem

Shaan introduced what he calls vibe coding for everything, and the music demo might be the most fun moment in the episode. He coaches a high school basketball team and wanted hype intro music. He typed hype basketball intro music for a high school team, high energy buildup, make the crowd go wild. No musical terminology. No beats per minute. No key signatures.

Within minutes, the tool created a full beat with stadium brass, 808 bass, and a drum buildup. When he wanted changes, he typed lower in tone, more masculine. Is it Grammy-worthy? No. But the point is devastating: this use case would never have existed before. He would never hire someone to make this. But if it takes two minutes and costs nothing, why not?

He then pivoted to Suno, the AI music platform, and this is where things got genuinely impressive. He and his personal trainer no longer listen to mainstream music during workouts. It is all music they have made for themselves. He played an original track that sounded like something from an Euphoria club scene, and it was indistinguishable from a professionally produced track.

As Shaan put it: skill and taste used to be inseparable. Now they are decoupled. You do not need the skill. You just need the taste.

The 40 Dollar Personal Stylist

Sam's demo was unexpectedly brilliant. He has a problem a lot of people share: he is between sizes, and buying clothes online is a gamble. Years ago, he physically measured every part of his body and kept the measurements in a Google Doc. He used to literally send this measurement sheet to high-end stores. Their response, as he joked: we have given this to security, you are now on a list.

Using Claude Code, he built a personal app in a single weekend. You enter your measurements, paste a product link, take a screenshot of the size chart, and it tells you exactly what size to buy with full reasoning. He trained it on actual books about how clothing should fit. It has been especially transformative for buying vintage clothes on eBay, where sizes from different eras and countries mean nothing. Japanese sizing runs so different he wears 3X there.

The cost was about 40 dollars in API fees for a weekend of building and using it. He has already ordered three items based on its recommendations and every one fit.

The AI Command Center Running a Tens-of-Millions Business

Shaan saved the most serious demo for last, and it might be the most commercially significant. His company scaled from zero to tens of millions in revenue in about two years, and instead of hiring a massive customer success team, one developer built an AI command center in three months.

The system connects to HubSpot, Fathom video recordings, Slack, and accounting software. Nobody manually enters anything. For each customer, it auto-generates status updates, contract details, and a renewal probability based on sentiment analysis of actual client calls. It reads the transcripts, scores the sentiment, and predicts expansion revenue.

There is an AI agent embedded in each customer record that you can just talk to. Ask it what expansion opportunities to pitch and it cross-references public info, call transcripts, and the internal service catalog. It creates follow-up tasks automatically. Anyone in the company can see what has already been brainstormed and what actions are in flight.

The K-Shaped Economy and Your Career

The conversation closed with a framework that ties everything together. Shaan described what he calls the K-shaped economy. A job is a bundle of tasks. About 80 percent of those tasks will be replaced by AI and 20 percent will be enhanced. If you do not use AI, of course it will take your job, the same way computers took jobs from people who refused to use them.

But the most practical advice came last. Shaan talked to the CEO of a billion-dollar company who appointed an AI general manager, not a human, just AI managing projects. Just by improving customer support with AI, making developers more productive, and removing junior programmers who were bad at using AI, they doubled their profit margin. At a billion dollars in revenue, doubling your profit margin is staggering.

Shaan's final point is the one that matters most for anyone feeling overwhelmed. You do not need to be in the top one percent or even 10 percent of AI users. If you are already good at something, business, content, whatever, just getting to the 50th percentile of AI ability is enough. That multiplies against your existing skills. The wrong game is competing with AI-native 22-year-olds. The right game is being dangerous enough with AI that it amplifies what you already know.

Key Takeaways

The era of prompting AI is ending. The next phase is AI that reads your context and acts without being asked.

Voice-first workflows are fundamentally changing how people create. Walking and talking produces better output than sitting and typing.

Personal software is the new frontier. Apps built for one person, one body type, one business are now trivially cheap to create.

Google Notebook LM is quietly one of the most powerful AI tools available and almost nobody is using it to its potential.

Skill and taste have been decoupled forever. You no longer need technical ability to create music, presentations, or software. You just need to know what good looks like.

The K-shaped economy means 80 percent of tasks get replaced and 20 percent get enhanced. The question is not whether AI will change your job but whether you will be on the winning side of the K.