Going deeper on the financial slides? This entry covers the full deck (slide order, narrative arc, what investors actually look at). For the dedicated treatment of the financial projections slide, what to include, how to build defensible numbers, common mistakes, see [Pitch Deck Financial Projections].
A pitch deck is a 10 to 15 slide presentation used to communicate a startup's market, product, traction, team, and capital ask to potential investors. It is typically delivered as both a sent-in-advance PDF and an in-meeting walkthrough, and is the primary artifact in modern fundraising for pre-seed through Series B rounds. It is the document founders agonize over more than any other and the one that produces the largest gap betw...
An elevator pitch is the 30 to 60 second verbal summary of a startup's business, used as the cold opening of investor or sales conversations. It's designed to be delivered in the time of an elevator ride with a stranger, with the goal of producing enough interest to earn a follow-up conversation. The name dates to the late 1980s and refers to the apocryphal scenario of catching a hard-to-reach executive in an elevator with a single brief window to make an impression.
The formula that works for most startups: name + category + the specific customer + the specific outcome + the differentiating insight, with [The Ask] as an optional closer when the audience is an investor. Worked example: "We're Stripe, payment infrastructure fo...
The people side of building a company. This cluster covers founder roles and dynamics, the executive lineup, the hiring sequence, sales and customer success roles, compensation and equity, performance management, layoffs and severance, and the culture and operations that determine whether the team holds together. 63 entries.
If your business succeeds or fails on hiring (most do), this is the cluster you live in.
A Large Language Model (LLM) is an AI system trained on massive amounts of text to predict the next token in a sequence. The prediction capability scales into broader abilities (reasoning, code generation, analysis, conversation, translation, summarization) as models grow in size and training data. Modern frontier LLMs range from 70 billion to 1+ trillion parameters and are the technology underlying ChatGPT, Claude, Gemini, Llama, and other generative AI products that have transformed software since 2022. It's the specific type of foundation model that handles text.
What LLMs actually do (the mechanics):
Tokens, not words: LLMs break text into tokens (sub-word units). "Tokenization" of a sentence might produce 10-...
Async work (asynchronous work) is an operating model where collaboration happens primarily through written communication and structured documentation rather than synchronous meetings. It enables teams to make decisions, share information, and coordinate work without requiring all participants to be available at the same time. The model is particularly valuable for distributed teams across time zones and for deep-work-heavy roles where uninterrupted time matters more than constant availability. It is a structural shift in how work gets done and a discipline that requires intentional culture-building, tooling investment, and behavior change.
The components of async work:
Written-first communication:
An AI agent is an LLM-powered system that plans, uses tools, and takes actions over multiple steps to complete tasks autonomously. Tools include APIs, code execution, web browsing, and file operations. Agents go beyond single-prompt question-and-answer to handle complex workflows requiring reasoning, tool use, and iterative correction. "Agentic AI" is the dominant 2025 frontier for AI applications and the next major capability layer beyond chat. Agents are what happens when LLMs stop just answering questions and start doing things.
What distinguishes agents from simpler LLM applications:
Multi-step reasoning: agents break complex tasks into steps and execute each.
Tool use: agents call APIs, run code, browse the web, query database...
Debt financing is raising capital by borrowing money that must be repaid with interest, used as an alternative or complement to equity financing. It includes everything from a personal credit card founders charge on day one to a $50M syndicated bank loan at a Series D company, with a wide spectrum of structures in between, each with different cost, covenant complexity, founder risk, and dilution tolerance.
The categories that matter for startups: founder-side debt (credit cards, personal lines of credit, home equity loans, used in the earliest pre-revenue phase, typically $5K to $100K, with personal liability and interest rates of 8 to 25 percent), SBA loans (Small Business Administration 7(a) and 504 programs, $500K to $5M t...
AI alignment is the research field and engineering discipline focused on ensuring AI systems pursue their intended goals correctly. It tackles the problem of getting models to do what developers and users actually want, rather than misinterpreting goals, gaming reward functions, or developing unintended behaviors. The work spans current techniques (RLHF, Constitutional AI, evaluation against intended behaviors) and fundamental research into how to align increasingly capable systems whose internal reasoning may be opaque. It's a subset of AI safety focused specifically on the goal-correctness problem.
The alignment problem:
How do you ensure an AI system pursues what you want, not something else? Sounds simple but is technically...
Founder roles is the explicit division of responsibilities, decision-making authority, accountability, and titles among co-founders, ideally documented at company formation in the founders agreement. The discipline exists to prevent the ambiguity that compounds into founder conflict over time. The typical division involves one founder taking the CEO role (strategy, fundraising, external relationships) and others taking domain-specific roles (CTO for technical leadership, COO for operations, CPO for product). The structural clarity matters more than the specific division: clear-division-A and clear-division-B both work fine, while ambiguity in either direction fails. It is the foundational structural decision that determines ho...
A reverse merger is a transaction in which a private company acquires a publicly-traded shell company to become public without a traditional IPO. Also called a reverse takeover (RTO), it merges the private company's operations into the public entity, typically using a dormant public company with little or no operations as the shell. It is the predecessor mechanic to the SPAC structure, was historically used by smaller companies as a cheaper alternative to IPO, and has largely been displaced by SPACs and direct listings in modern practice.
The mechanic: a private company identifies a public shell company (often a former operating company that has shed most of its assets but kept its public listing, or a company specifically cr...