This AI agent completes your To-Do list (plus 4 AI tools that’ll blow you away)

My First Million

Published: February 13, 2026
Duration: 57:20
Summary source: description
Last updated: Apr 25, 2026

Discusses openai, agents.

Summary

Get the cheat sheet: 5 AI tools to make your first million https://clickhubspot.com/ybo Episode 795: Sam Parr ( https://x.com/theSamParr ) and Shaan Puri ( https://x.com/ShaanVP ) show each other the cool things they’ve built with AI. — Show Notes: (0:00) 1 - Do Anything (18:32) 2 - Business bio wikipedia (30:11) 3 - Muse Art (42:13) 4 - Fit finder (47:20…

In this episode, the hosts explore cutting-edge AI tools and their potential to revolutionize everyday tasks, from creating personalized content plans and music to enhancing business operations and shopping experiences. They discuss AI's role in mass personalization, showcasing tools like "Do Anything" for automatin…

Key takeaways

AI tools are revolutionizing content creation and business processes by automating tasks and providing strategic insights, as demonstrated by tools like 'Do Anything' and custom-built internal systems.
Mass personalization in software is becoming a reality, allowing for highly tailored applications that cater to individual needs and preferences, such as personalized music creation and clothing size recommendations.
AI is significantly enhancing productivity and decision-making in businesses by integrating with existing systems to automate workflows, analyze data, and predict customer behavior, leading to substantial improvements in efficiency and profitability.

Why this matters

Understanding and leveraging AI tools can drastically enhance business operations, content creation, and personal productivity, providing a competitive edge in an increasingly digital and automated world.

Entities

ChatGPT Garrett Pipe Dream Goods Do Anything Claude Nebula Furcon Notebook LM Google Ray Dalio Fabian Glyph HubSpot Mercury Anthropic Suno Rick Rubin Justin Bieber

Strategic Intelligence Report▼

AI Tools Worth Knowing: A Practitioner's Show-and-Tell on Agents, Personal Software, and the Taste-Skill Decoupling Two practitioners share hands-on experiments with AI tools across content strategy, music creation, business intelligence, and personal productivity—offering a grounded view of where agentic AI is actually useful today versus where it remains aspirational. The audience is founders, operators, and knowledge workers evaluating where to place early bets on AI-augmented workflows.

The Agent Question: What Does "Useful" Actually Look Like?

The discussion opens with a pointed observation: agents are widely discussed but rarely demonstrated in practice. The first concrete example is a tool called Do Anything (doanything.com), built as a side project by the founder of Pipe Dream. The premise is a self-executing to-do list—users describe a task in plain language and the tool attempts to complete it autonomously. The demonstration involves asking the tool to analyze a YouTube channel's performance without providing login credentials or a direct link. Within minutes, the tool surfaced subscriber counts, average view performance relative to audience size (flagging a 3–5% click-through rate on new uploads as underperforming), and a diagnosis of why certain episodes outperformed others. It then generated a one-month content plan, proposed video titles, and offered to script individual episodes or develop thumbnail concepts—all from a single conversational prompt issued roughly 19 minutes before a recording session. The broader point: the value of agents is not in replacing human judgment but in compressing the time between intent and a usable first draft. The discussion notes that ChatGPT's "Pulse" feature extends this logic further—proactively surfacing relevant information without requiring a prompt at all.

Proactive Context Awareness: The Next Layer

A tool called Nebula, an early prototype from a founder named Furcon, is described as connecting to Slack, Gmail, and Google Calendar to generate meeting prep documents automatically and summarize GitHub commits without being asked. The discussion is candid about its current limitations—it is described as an "early, early prototype" not yet reliable enough for small teams where members already have direct visibility into each other's work. The use case sharpens for teams of 10 or more, where ambient awareness of what colleagues are doing becomes genuinely difficult to maintain.

Personal Software: The Vibe Coding Thesis

The discussion frames a macro argument: the news feed personalized content consumption; AI will do the same for software. The term "vibe coding" is used to describe building software by describing intent rather than specifying technical requirements—iterating conversationally until the output matches a felt sense of what was wanted. Two personal software examples are demonstrated. The first is a biography-to-Notion pipeline built with Claude (described as taking 45 minutes to construct). Users upload multiple ebooks of business biographies on a single subject, and the tool generates a structured Wikipedia-style page including: a financial timeline adjusted to 2025 dollar values, milestone analysis, and a "Founder's Playbook" section that cross-references the subject's decisions against the user's own business context and challenges. The Ted Turner example illustrates how the tool surfaces that CNN launched when Turner was in his 40s—a detail that carries different weight when rendered in a structured timeline than when embedded in narrative prose. The second example is a clothing size recommendation app, also built with Claude, that accepts body measurements, a product URL, and a screenshot of a retailer's size chart, then recommends the correct size with reasoning. The builder notes meaningful API costs (~$40 for a weekend of personal use) as a barrier to publishing it as a product.

Skill-Taste Decoupling in Creative Work

A music beat-generation tool called Museart is demonstrated live, with prompts like "hype basketball intro music for a high school team, make the crowd go wild"—no tempo, key, or instrument specifications required. The output is described as functional within seconds. The discussion then shifts to Suno, described as more capable for full song production, with one participant noting he no longer listens to mainstream music during workouts, instead using a shared playlist of AI-generated tracks. The conceptual frame offered: historically, producing creative work required both skill (technical mastery of tools) and taste (aesthetic judgment). AI decouples these. Taste alone is now sufficient to participate. The implication for businesses is that production costs for creative assets—music, presentations, content—approach zero for anyone willing to iterate conversationally. A related demonstration involves Google's NotebookLM generating a slide deck from a YouTube link to a podcast episode, including auto-generated graphics, in approximately 10 minutes. A separate tool called Glyph is described (not fully tested) as using AI video transition technology—originally built for short-form video—to animate between unrelated slides as if they were keyframes in a designed sequence.

Internal Business Intelligence: The CRM Agent

The most operationally mature example is an internally built customer success dashboard at a services company that scaled from zero to tens of millions in revenue in approximately two years. One engineer built the system over roughly three months. It connects HubSpot, Fathom call recordings, Slack, and accounting software to auto-populate customer records without manual data entry. Key capabilities: sentiment analysis of client calls to predict renewal likelihood, projected net revenue retention with confidence levels, identification of expansion opportunities by cross-referencing call transcripts against the company's service catalog, and automatic task creation assigned to the relevant team member. The system is described as the company's operational command center, with the explicit goal of maintaining service quality at scale without proportional headcount growth.

The K-Shaped Economy Framework

The discussion introduces a framework called the "K-shaped economy" to describe AI's labor market impact. Any job is described as a bundle of tasks. The argument: roughly 80% of those tasks will be replaced or substantially automated by AI; the remaining 20% will be enhanced—meaning workers who use AI will perform them dramatically better than those who do not. The divergence between adopters and non-adopters, not the elimination of jobs wholesale, is the predicted outcome. --- Key takeaways: - **Agents deliver clearest value at team scale**: Proactive, ambient AI tools that monitor Slack, GitHub, and calendars are most useful for teams of 10 or more; small teams with direct communication already have the context these tools synthesize. - **Personal software is the next frontier of personalization**: Just as algorithmic feeds replaced mass-appeal content, vibe-coded personal apps replace one-size-fits-all software—the biography pipeline and sizing app are early examples of software built for an audience of one. - **Taste is now the scarce input in creative production**: Music, slide design, and content strategy tools have reduced the skill barrier to near zero; aesthetic judgment and clear intent are now the primary differentiators. - **Internal AI tooling can substitute for headcount in services businesses**: A single engineer building a connected CRM-intelligence layer can enable a small team to manage client relationships at a scale that previously required significantly more staff. - **The opportunity set is expanding faster than competition is intensifying**: The "musical chairs" framing suggests that AI is simultaneously adding players and multiplying available chairs—making the absolute opportunity larger even as the field grows more crowded.

Show notes

Themes

openai
agents

openai agents

Original audio (RSS) ↗Apple Podcasts (show) ↗Official site ↗