JSN

Designing-Building

Insights from venture capital investors, e.g., Marc Andreesen, Steve Jurvetson, YC partners and others to evaluate 'does this product have legs?' and 'how am I doing?'. The following list may even reveal enough clues to identify the concept.

☺ Tracking well:

  • is an all-in-AI product
  • product could not have existed, nor been realistically envisioned, 6-8 months ago
  • present tense pain point: addresses FOMO about job loss due to AI
  • future tense pain point: will become impossible to manage without a structured tool or product
  • is global – every person, in every place on earth, can use it
  • is based on the evolution of human work due to AI, that may manifest as a transformation in the world economy
  • assumes the progressive adoption of LLM capabilities that currently exist but are vastly under-utilized
  • useful to a vast range of users: every day lay-people to venture capital portfolio managers
  • has network effects with incentives across three distinct personas
  • exploits a legacy model of startup ventures to successful companies
  • has a moat similar in nature to Visa and the payment networks. Promotes a standard
  • costs nothing to get started. costs at higher levels remain negligible to lay-people
  • is inexpensive to operate
  • succeeds financially at scale
  • has security and confidentiality core to its architecture
  • does not host sensitive, confidential customer data
  • founder is advisable
  • founder with experience. Previously founded and built a company to profitability in nascent market
  • founder has worked inside, observed success factors within a global payments company
  • strategy and build process fully leverages AI tools up and down the stack (see tools)
  • no known direct competitors. Related: visible.vc/ capboard.io/ > counter-positioning: my AI agent version of these incumbent dashboard tools could cannibalize their own products
  • taste and judgement – more important than ever, since technical development is no longer a bottleneck
  • deep workflow control: embedding deeply into a business's operations so the product becomes impossible to rip out (see next item)
  • confidentiality of data and security requires deployment in standalone agent (e.g., Harvey for legal provides security and confidentiality as a core feature)
  • sticky product, useful, simple, effective
  • peer-to-peer
  • non-hosted customer data avoids legal risk
  • most established companies aren't seeing returns from AI, but we know solo founders and small teams building with AI turn are increasingly profitable quickly without investment. AI's sweet spot appears to be not evolution of incumbents, but a revolution of our economy away from enterprises and SMBs to long-tail founders building products and minting money

☹ Tracking unwell:

  • solo founder without a co-founder (currently scanning for co-founders)
  • founder that is father to a small child (joint custody)
  • founder that is not extremely "messianic" (e.g., Elon Musk)

AI-native companies will outnumber human-managed companies. Today's management stack assumes humans produce status updates, reports, KPIs, and board materials. Autonomous companies don't. A new communication standard is inevitable.

Don't give us your data. Keep it where it already lives. We'll standardize how trusted conclusions move between businesses.

The standard by which AI-run businesses communicate decisions, health, and intent with one another

The primitive is not the assertion itself (this is confidential info). The primitive is the exchange of an assertion via a protocol.

Not status updates. Executive communications. Business-to-business trust.

Expecting a metamorphosis of the economy in the next 3-10 year, AI will engender a new breed of business owners. The product needs to be useful today, modeling best business practices and also must be future-proofed for where the world is headed. Today's customers may largely be human business owners and venture capital investors communicating via the Executive Assertion Network.

Today
Business Owner  <------>  VC Investor
             \           /
              \         /
          Executive Assertion Network

These early customers don't go away, as traditional business operating models will not disappear, but we need to name them more generically. In the future, one person may start and own autonomous agent-run businesses and also oversee them as a sort of portfolio managers. That means the customer maybe be referred to as:

Portfolio Executive

Future
Portfolio Executive
          \  
Agent overseer <------>  Agent autonomous businesses (many)
            \           /
             \         /
    Executive Assertion Network

A portfolio executive might be any of the following:

  • VC partner
  • Angel investor
  • Solo founder
  • Family office
  • Acquisition entrepreneur
  • Holding company
  • Search fund
  • AI-native conglomerate

In many, if not most, cases the customer will be wearing two hats:

  • Owner: founder or acquire, then manage operating success of each agent-run business
  • Portfolio exec: oversee all of the agent-run businesses of which they are owner

They're doing the same thing from different sides: communicating and monitoring vitals of their businesses

What changes when a designer can stay close to strategy, interaction, content, implementation, and shipping without handing the work across a hard boundary?

Tools

Building with Codex, Cursor, Figma, Vercel, Supabase, CloudFlare and a growing set of AI-native tools. The point is not to cosplay as an engineer. The point is to keep product judgment close to the material of the product itself: flows, data, interface states, content, edge cases, and the quiet details that make it trustworthy.

Working Thesis

Designer-builders are not replacing teams. They are changing the altitude at which teams can operate. This one can make the abstract concrete, make the concrete better, and prove how product strategy funnels all the way down to pixels and implementation and back up the stack instantly in fast, repetitive cycles.