The Playbook

Ideation Agent

Assign a dedicated Claude agent solely to ideation. Give it a clean context window with only the inputs relevant to generating ideas (e.g., topic brief, audience, goals). Keep it isolated from drafting or reviewing tasks to avoid context contamination.

Drafting Agent

Assign a second dedicated Claude agent solely to drafting. Feed it only the output from the ideation agent plus any relevant style or format guidelines. Do not include review feedback or unrelated context in this window.

Reviewing Agent

Assign a third dedicated Claude agent solely to reviewing. Feed it only the draft from the drafting agent plus any relevant quality criteria or rubric. Keep its context window clean of the ideation and drafting history to ensure an unbiased review.

Browse a curated list of 100 actionable side hustle ideas

Access a database of 100 specific, non-generic side hustle business models that real people are actively using. Avoid vague suggestions and focus only on ideas with proven execution paths.

Evaluate each idea against three key criteria

For each side hustle idea, assess: (1) startup costs — how much capital is required to begin, (2) earnings potential — realistic income ceiling or range, and (3) time commitment — hours per week required to operate.

Select an idea you can launch within the current month

Use the criteria from the database to narrow down to one idea that fits your available budget, time, and skills. Prioritize speed to launch over perfection.

Take immediate action on your chosen idea

Begin executing on the selected side hustle without delay. The framework emphasizes actionability — the goal is to start building, not just researching.

Identify Your Task Type Before Prompting

Before sending any prompt, categorize the task: Is it a quick lookup, a simple rewrite, or a surface-level question? Or does it require multi-step reasoning, document analysis, or nuanced judgment? This classification determines which AI mode to use.

Use Fast Mode for Low-Complexity Tasks

For simple, well-defined tasks — such as summarizing a short text, answering factual questions, generating a quick list, or light editing — use the standard (fast) AI response mode. These tasks do not benefit from extended reasoning and fast mode reduces latency and cost.

Use Deep/Extended Thinking Mode for Complex Tasks

For tasks requiring multi-step logic, document analysis, strategic planning, comparative evaluation, or nuanced synthesis, switch to the model's extended thinking or reasoning mode (e.g., o1, o3, Claude's extended thinking). This allows the model to deliberate before responding, improving accuracy on hard problems.

Apply a Reusable Prompt Template for Document Analysis

When analyzing documents, use a structured prompt template: (1) State the document type and context, (2) Specify exactly what you want extracted or evaluated, (3) Define the output format (e.g., bullet points, table, summary), and (4) Add any constraints or lens to apply (e.g., 'from a policy risk perspective' or 'for a non-technical audience'). This template works across policy docs, reports, contracts, and research papers.

Match Mode to Stakes, Not Just Complexity

Even moderately complex tasks may warrant deep mode if the stakes are high (e.g., decisions, public-facing content, legal or financial analysis). Use fast mode freely for internal drafts and iteration; reserve deep mode for final outputs or high-consequence decisions.

Client Agent Locks Payment in Smart Contract Escrow

The client AI agent initiates a transaction by locking the agreed payment amount into a smart contract escrow. This ensures funds are committed before any work begins, protecting both parties and removing the need for trust between autonomous agents.

Provider Agent Completes the Work

The provider AI agent executes the requested task — such as writing code, generating media, analyzing portfolios, or coordinating workflows. Work is performed against the locked escrow, with payment guaranteed upon successful verification.

Evaluator Verifies and Releases or Refunds Payment

A neutral evaluator — which can be an AI agent, a Zero-Knowledge (ZK) contract, or a DAO — reviews the completed work against the agreed criteria. If the work meets the standard, payment is released to the provider agent. If not, the payment is refunded to the client agent.

Start With One Vendor, Not Many

Pick a single AI SDR vendor that covers the majority of your use case and go deep with it. At most, use two: one for outbound, one for inbound. Do not buy three or four tools at the start. The tool matters far less than the strategy you bring to it.

Prove Your Human Playbook Works First

Do not deploy an AI SDR until you have already validated your outbound motion with human SDRs. Identify which messaging converts, which segments respond, and what cadences perform. Feed that proven context into the AI agent. The AI SDR's job is to clone your best performer — if you feed it untested or poor-quality context, it will produce poor results. Founder-led sales must come before agent-led sales.

Align with executive incentives

Before any interaction, research and understand what the executive is measured on and cares about. Frame your proposals in terms of their goals, not yours. Show explicitly how your ask advances their priorities and removes obstacles to their success.

Present three options, not one

Never bring a single recommendation to an executive. Instead, present three clearly scoped options with distinct tradeoffs (e.g., fast/cheap/limited vs. slow/expensive/comprehensive vs. balanced middle path). This gives executives agency, surfaces your thinking, and prevents a binary yes/no dynamic.

Enter meetings to learn, not to convince

Reframe your mindset before executive meetings: your goal is to gather information and understand their perspective, not to push through your agenda. Ask questions, listen actively, and treat the meeting as a discovery session. This builds trust and often surfaces context that changes your approach entirely.

Follow the breadcrumbs when leaders signal direction

Pay close attention to subtle cues, offhand comments, and repeated themes from executives. When a leader signals a direction — even informally — treat it as a strategic signal worth acting on. Document these signals, connect the dots across conversations, and proactively align your work to where leadership is already heading.

Recognize the macro timing: technology works, but scaling takes years

Acknowledge that the core technology is already proven (Waymo is delivering 400k+ rides/week; Tesla, Zoox, Baidu are racing to scale). However, adoption will follow a slow S-curve due to city-by-city regulatory rollouts, fleet economics, and consumer trust-building. This gap between 'working' and 'scaled' is the early investor window. Target horizon: ~$40-50B industry by 2030.

Frame robotaxis as an embodied AI infrastructure play, not just a transport play

Reframe the investment thesis around robotaxis being the first real commercial application of embodied AI. This framing expands the investable universe beyond fleet operators to include AI infrastructure, sensors, mapping, compute, and software layers that underpin the entire buildout.

Identify and diversify across multiple investable angles in the AV stack

Rather than betting on a single robotaxi operator winning, map out and allocate across the full buildout ecosystem. Investable angles include: (a) fleet operators/platform companies (Waymo, Tesla FSD, Zoox, Baidu Apollo), (b) sensor and hardware suppliers (LiDAR, cameras, radar), (c) AI software and mapping infrastructure, (d) charging and fleet management infrastructure. Early positioning across layers reduces single-company risk while capturing the overall trend.

Identify the single stock with the cleanest exposure for a core position

Among all the investable angles, select one stock that offers the cleanest, most direct exposure to the robotaxi buildout as a core holding. Per the playbook, Vincent has identified a specific stock for this purpose (withheld behind PRO paywall). Criteria for 'cleanest exposure' imply: pure-play or near-pure-play in autonomous vehicle commercialization, revenue already being generated or near-term catalysts visible, and limited exposure to unrelated business lines that would dilute the AV thesis.

Size positions for a multi-decade trend with early-stage patience

Position sizing should reflect the long time horizon. This is not a near-term catalyst trade. Expect slow S-curve adoption with city-by-city expansion. Avoid over-concentration given regulatory and execution risk, but enter early enough to capture the compounding effect of infrastructure buildout before mainstream investor attention arrives. Monitor milestones: new city launches, regulatory approvals, fleet size growth, and unit economics disclosures.

Layer in macro risk management around geopolitical and oil price signals

Even high-conviction long-term positions require macro overlay. Rising oil prices and geopolitical conflicts (e.g., Iran) historically precede weaker environments for risk assets. Maintain defensive awareness: if macro signals deteriorate (oil spike, geopolitical escalation), consider trimming cycle-sensitive positions and rotating toward revenue-generating, defensive holdings while keeping long-term AV infrastructure exposure intact.

Assess Time-to-Value and Agent Performance Before Hiring

Before making any CS hiring decisions, evaluate two key factors: (1) how long it takes customers to get value from your AI agent product, and (2) whether the agents are consistently performing well. If time-to-value is long OR agents underperform, do not hire CSMs yet — hire FDEs first.

Hire FDEs First If Deployment Is Not Yet Proven

When your AI agent product requires complex configuration, troubleshooting, or customer-side setup, prioritize hiring Field Deployment Engineers (FDEs) over Customer Success Managers. FDEs unblock technical barriers, reduce churn risk, and accelerate time-to-value before a scalable CS motion is possible.

Embed One FDE With Your Top 3-5 Customers to Systematize Deployment

Assign a single FDE to work deeply with your top 3-5 highest-potential customers. The goal is to identify what a successful deployment looks like, document repeatable patterns, surface common failure points, and build a deployment playbook that can be handed off to less technical team members later.

Prove and Document a Repeatable Deployment Process

Only move to scaling CS once the FDE work has produced a proven, repeatable deployment process. This means: clear onboarding steps, known configuration requirements, documented edge cases, and measurable time-to-value benchmarks. If you cannot hand this to a CSM and have them execute it reliably, it is not ready to scale.

Hire CSMs Only After Deployment Is Proven and Repeatable

Once the FDE has systematized deployment and the product reliably delivers value within a predictable timeframe, begin hiring CSMs to manage relationships, drive expansion, and handle renewals at scale. CSMs should inherit a working playbook — they should not be expected to figure out deployment from scratch.

Use the 30-Day Deployment Interview Test When Hiring Either Role

When interviewing FDE or CSM candidates, ask: 'What commercial AI agent or tool with real ROI have you deployed in your organization in the last 30 days?' Strong candidates will answer with specifics — what they bought, why, how they configured it, what broke, and what it does now. Candidates who cannot answer this concretely are not operating at the required level for AI agent products.

Deposit ERC-20 Tokens into the Privacy Pool

Send standard ERC-20 tokens into the STRK20 Privacy Pool smart contract. This is the entry point that converts public, traceable tokens into the private system.

Receive Encrypted Notes

Upon deposit, the protocol issues encrypted notes representing your token balance. These notes are only readable by the holder's private key, hiding ownership and amounts from public view.

Transfer Value via UTXO Model with One-Time Channel Setup

To send tokens privately, use the UTXO (Unspent Transaction Output) model. Set up a one-time channel per transfer so that inputs and outputs are cryptographically unlinkable, breaking the on-chain trail between sender and recipient.

Interact with DeFi Anonymously in a Single Transaction

Use the encrypted notes to interact with DeFi protocols (swaps, lending, etc.) in a single atomic transaction, preserving privacy end-to-end without exposing the user's identity or balance at any step of the DeFi interaction.

Train Claude on a Reference Library of Brand and ICP Data

Compile a reference library consisting of brand documents (tone of voice guidelines, messaging frameworks, product briefs, past campaigns) and ideal customer profile (ICP) data. Feed this library into Claude to ground the AI in your specific brand context, ensuring outputs align with brand standards and target audience expectations rather than generic responses.

Enrich Account and Persona Data Using Clay and Apollo

Use Clay and Apollo to enrich account-level and persona-level data — pulling in firmographics, technographics, job titles, buying signals, and behavioral intent. Feed these enriched insights back into Claude as dynamic context to build personalized campaign skills, enabling the AI to generate highly targeted messaging and campaign assets tailored to specific accounts or segments.

Reduce Onboarding Friction

Identify and eliminate unnecessary steps in the onboarding flow that prevent users from reaching their first value moment. Map the current onboarding journey, find drop-off points, and remove or simplify any step that doesn't directly move the user toward activation.

Align Pricing with Observed User Behavior

Audit your pricing model against actual usage data. Identify which features drive the most engagement and ensure your pricing tiers reflect how users naturally adopt and expand usage. Remove pricing structures that penalize early or growing users and reward the behaviors that correlate with long-term retention.

Design Trials to Drive Real Engagement

Structure free trials so users are guided to experience the core value proposition — not just surface features. Set trial milestones that mirror the activation events of your best retained users. Use in-trial nudges and progress cues to push users toward meaningful engagement before the trial ends.

Protect Product Direction from Speed-Driven Feature Bloat

As shipping velocity increases, establish a clear decision framework for what NOT to build. Use a lightweight vetting process (e.g., evidence-backed proposals, user behavior data, strategic alignment check) before any feature enters development. Speed without judgment degrades product quality and erodes long-term user trust.

Compound Retention Through Continuous Alignment and Visibility

Use living, evolving documentation (not polished static docs) to keep teams aligned on retention signals in real time. Make retention metrics and behavioral insights visible across product, design, and growth teams continuously — not just at quarterly reviews. Small, consistent alignment compounds into sustained retention improvements over time.