IGNITE EXTENDED / LESSON 01
What must be true before an idea deserves more time and resources?
A one-page opportunity-and-uncertainty brief.
Humane, founded by former Apple designers Imran Chaudhri and Bethany Bongiorno, raised $230 million before shipping its first product. The Ai Pin went on sale in April 2024 at $699 plus a $24 monthly subscription, as a screenless, voice-driven wearable meant to reduce phone use. Reviews were harsh: slow and unreliable answers, short battery life, a projector that was hard to control, and a device that grew warm. In May 2024, about a month after launch, Bloomberg reported the company was seeking a buyer at $750 million to $1 billion. By August 2024 reports put sales near 10,000 units against a goal of 100,000, with daily returns exceeding sales and roughly 7,000 devices still in customers' hands; the company had spent more than $200 million over about six years. In February 2025 HP agreed to buy most of Humane's assets, including its software platform and about 300 patents and applications, for $116 million. The Pin stopped working on February 28, 2025, and only recent buyers received refunds. A celebrated team, a large vision and ample capital did not substitute for evidence that a specific customer had a recurring problem the product solved better than the phone already in their pocket.
Sources: TechCrunch: Humane's $699 Ai Pin is now available (April 11, 2024) · Engadget: Wearable Ai Pin maker Humane is reportedly seeking a buyer (May 2024) · 9to5Mac: Humane Ai Pin woes worsen as recent returns exceed sales (August 7, 2024) · Android Central: Humane will disable Ai Pins after $116 million HP buyout (February 2025)
Map Humane onto the venture life cycle. Which assumption in the chain customer-problem-alternative-adoption-payment was weakest, when could it have been tested cheaply, and why might a well-funded team with investors seeking a very large outcome have postponed that test?
A customer can want a product that still fails as a business. The customer must experience a problem, recognize the proposed benefit and have a workable way to obtain the solution. Someone must also pay enough to cover the activities required to deliver it. Each connection needs evidence.
The venture life cycle changes which uncertainty deserves attention. A prototype can address feasibility. An actual purchase tests more than a statement of interest. Repeated purchases begin to test whether the company can retain customers and deliver consistently. Funding a later stage before resolving an earlier uncertainty increases the cost of being wrong.
A probability describes what a person expects to happen. A preference describes how much that outcome matters to them. Two decision makers can agree on the probability of success and choose different actions because they face different consequences.
A founder may have personal savings and several years of work concentrated in one company. A fund can spread exposure across investments and may need a very large outcome to affect its overall return. The VMware discussion uses this difference to examine market choice and investor incentives. An apparent disagreement about strategy may conceal a disagreement about objectives.
Use this sequence to locate the uncertainty that deserves attention before the next commitment. A venture can move backward when new evidence challenges an earlier assumption.
Identify a person, a recurring difficulty and the alternative used today. Example: clinic staff repeatedly re-enter appointment details. The proposed app is a possible response, not yet a demonstrated business.
Check whether the need, adoption process and economics fit together. Observe the work, test a prototype and ask who can authorize payment. Interest alone leaves several of these questions unresolved.
Assemble the people, cash, rights and partners required for the next test or delivery milestone. Hire against a specific capability gap rather than treating headcount as evidence of progress.
Deliver the promised result repeatedly and examine retention, capacity and cash. A successful first customer does not establish that acquisition and service can be repeated economically.
The next slide distinguishes an attractive idea from an opportunity. That distinction determines what the evaluation stage must actually test.
Use this distinction when deciding what to investigate. A proposed product names a possible response; an opportunity connects a need to adoption and viable delivery.
A possible solution: “an application that automates reports.”
A customer has a recurring problem, an alternative is inadequate, and an organization can deliver value economically.
Illustrative application: “automate reports” leaves the buyer, task and economics open. Observing clinic managers spend an hour reconciling appointment changes identifies a specific situation. Testing whether they will adopt and pay for a workable alternative investigates the opportunity.
Keep the solution provisional while checking the need, purchase process and delivery cost.
An operating milestone records an accomplishment, such as a completed prototype or a new sales hire. A learning milestone resolves a question important enough to change the next decision. The two can coincide, but activity by itself does not reduce uncertainty.
A useful test states an assumption, a method of observing it and an action for different outcomes. The response should be specified before the result arrives. This reduces the temptation to reinterpret every outcome as encouraging. A test also needs a limit on time and resources so that learning does not become an indefinite reason to postpone a decision.
A director, an investor and an advisor can all contribute expertise, but they occupy different relationships with the company. Advice may influence a decision without carrying formal authority. Equity may provide economic rights without giving its holder sole control over operating decisions.
The entrepreneurship material asks who is at the table because incentives and authority shape the advice a founder receives. A founder should understand the person’s role, the decisions they can approve and the outcomes they prefer. This also helps identify a missing capability: operational experience, customer access and governance oversight are different contributions.
The source separates traits associated with entrepreneurs from traits associated with successful outcomes.
A useful decision process connects a choice to a belief that could be wrong. The goal is to identify which uncertainty could change the action.
Specify feasible alternatives and the immediate commitment. Example: spend $5,000 recruiting pilot customers now, or spend one week investigating the onboarding problem first.
Write the causal assumption behind each alternative. Here the team believes poor onboarding, rather than weak demand, explains why interested customers do not start using the product.
Choose an observation that separates explanations. Watch target customers attempt setup unaided and ask why they stop. A general satisfaction survey would reveal less about this particular disagreement.
Decide what to do for different findings, name the owner and set a review date. If customers finish setup but still decline the offer, revisit the demand assumption.
A test becomes useful when its possible results lead to different actions. More information without a decision consequence may not justify its cost.
The same person can offer useful expertise and hold formal rights, but those contributions must be distinguished when interpreting advice.
A formal governance role with decision and oversight responsibilities.
Expertise and support within an agreed scope; a useful opinion does not automatically carry formal authority.
Illustrative application: an adviser recommends a cautious pilot because installation is uncertain. A director may have a formal vote over financing. An investor may prefer faster growth. Ask about evidence, objectives and authority separately before deciding how each view should affect the choice.
Role clarity explains both whose approval is needed and why well-informed people may recommend different actions.
Click either page to read it at full size. In the reader, use ← → to switch between these two pages.
Selected reading: PDF pages 1, 2
The opening case places Diane Greene at a transition: VMware has customers, growing revenue and new senior executives. Its next challenge is to create a management system capable of supporting growth.
Greene must integrate experienced executives, strengthen sales and introduce processes into an engineering-centered company.
Delegation transfers functions that Greene previously handled personally. New capability also creates a need for clear authority and coordination.
The excerpt describes products, enterprise revenue components and several channels. Product expertise operates within that wider commercial system.
A profitable, growing business can still face organizational uncertainty. Success at one stage changes the problem management must solve next.
Selected reading: PDF pages 1, 2
These opening pages establish the situation. They do not report whether the new management team ultimately solved the integration problem.
A senior executive’s prior achievements do not establish how that person will work in this company’s stage, culture and decision process.
Adding a sales leader changes more than headcount. Responsibilities, information flows and relationships with engineering need explicit attention.
The archived entrepreneurship discussion compares entering a smaller workstation market with pursuing the larger, more uncertain server market. Existing progress and adoption evidence favor one path; scale and investor objectives may favor another.
Which evidence would help choose, and whose objectives should be made explicit?
In March 2025 Y Combinator managing partner Jared Friedman said about a quarter of its Winter 2025 batch had codebases roughly 95% generated by AI, while stressing the founders were highly technical. A small team can now produce a convincing demo in days.
When feasibility is cheap to show, a working product is weaker evidence that the venture is an opportunity. The assumptions that remain expensive to resolve are adoption, willingness to pay, distribution and trust, which is where Relay's first test belongs.
Humane raised $230 million and sold about 10,000 devices against a 100,000 goal before selling its assets for $116 million. Abundant funding can postpone, not replace, the evidence that customers want the product over their current alternative.
Sources: TechCrunch: A quarter of startups in YC's current cohort have codebases that are almost entirely AI-generated (March 6, 2025) · 9to5Mac: Humane Ai Pin woes worsen as recent returns exceed sales · Android Central: Humane will disable Ai Pins after $116 million buyout
Fictional exercise: a team can launch a narrow product in 3 months or a broad product in 9 months. It has 6 months of cash. Two narrow-market buyers have agreed to a pilot; broad-market demand is untested. The founder values independence and a potential investor values scale.
Write the decision, separate known facts from beliefs and preferences, and propose one useful test before committing.
Maya Chen ran operations for a 6-location PT group in Austin and watched front-desk staff spend about an hour a day phoning patients to fill cancellations. Daniel Okafor can build SMS and scheduling software. Their illustrative clinic: 4 therapists, about 280 visits a week, 15% late cancellations and no-shows, about 42 empty slots, $110 average reimbursement. If Relay refills 40% of those slots, a clinic recovers about $96,000 a year of gross revenue for a $399 monthly fee. None of these figures has been observed in a pilot yet, and three clinics have agreed to try a manual version.
Write a one-page brief: the chain of assumptions that must hold for Relay to be an opportunity, the evidence for each, the single uncertainty that most deserves the next dollar, and a time-boxed test with a decision rule set before results arrive.
Relay's $96,000 figure assumes the clinic fills nothing today. The front desk already fills 10-15% by phone, so the defensible claim is the increment over that alternative: about $66,000 a year.
Engineering the EHR integration before testing whether patients accept texted slots spends the most expensive resource on the least informative question. Test the cheapest decisive assumption first.
Roughly 38,000 US PT locations (illustrative) times $4,788 a year describes a ceiling, not a demand signal. It says nothing about who will buy first, through which channel, or at what acquisition cost.
A one-page Relay opportunity-and-uncertainty brief: assumption chain, incremental value over the phone list (about $66,000 a year; $96,000 gross), the patient-acceptance test design and its decision thresholds.
Next lesson: Clear business communication & audience engagement
Check your understanding
Question 1
A Relay pilot clinic has 42 empty slots a week. Its front desk already fills 12.5% by phone; Relay's pilot fills 40%. At $110 per visit, roughly how much additional annual revenue does Relay create compared with the status quo?
Incremental fill is 40% - 12.5% = 27.5%: 42 x 0.275 = 11.55 visits x $110 x 52 = about $66,000. $96,000 compares Relay with filling nothing, which overstates the value because the customer's real alternative is the phone list.
Question 2
Three clinics agree to try Relay. Daniel wants to spend two months building the EHR integration before the pilot so the test is realistic. Maya wants to text patients manually first. Which argument best supports Maya's sequence?
The argument is about which uncertainty to resolve first. If patients ignore texted offers, the integration is wasted. 'Manual is always first' is too broad: when feasibility is the binding uncertainty, building may be the right test.
Question 3
After five interviews, every clinic owner says Relay sounds great and they would definitely use it. What has the team learned?
Interviews can support the problem hypothesis. Stated enthusiasm costs nothing and does not show adoption, payment or a repeatable sale. A paid pilot with a defined success criterion is a different, stronger test.
Question 4
A founder and an investor both estimate a 10% chance that a venture reaches a $500 million outcome. The investor wants to pursue it aggressively; the founder prefers a smaller, steadier plan. What is the disagreement mainly about?
They agree on the probability, so more data will not settle it. The fund spreads risk across many investments and needs large outcomes; the founder has savings and years concentrated in one company. Name the objectives first, then debate estimates.
Question 5
Relay's founders list their goals for the next quarter. Which of these is a learning milestone, one whose result would change the next decision, rather than an operating milestone?
A learning milestone resolves a question important enough to change the next decision. Shipping a service or hiring records activity; it may be necessary, but it does not by itself reduce uncertainty about whether Relay creates value.
Question 6
VMware entered the smaller workstation market before pursuing the larger, more uncertain server market. Which reasoning best fits how the case frames that early decision?
The case pairs two parts: probabilities (where adoption was more predictable) and preferences (founders versus investors seeking scale). 'Small is safer' ignores that a narrow start is only good if it builds a credible path to the larger market.
Workbook
Module output: A decision journal, a short explanation and a working agreement.
Library
10 appear in the slides with notes; the rest are kept here for reference.
Source S03 · Feinberg · 19 pages
Source S17 · Feinberg · 23 pages
Source S16 · Feinberg · 32 pages
Source S01 · Feinberg · 14 pages
