Best of nfx.com on Kiwi News
Most upvoted by Kiwi News readers
14 stories from nfx.com have been shared on Kiwi News since 29 June 2023, with 29 upvotes and 2 comments in total. The latest was shared on 22 May 2025. They are listed below, most upvoted first.
Text under a title is an AI summary of the linked article.
- Scientific approach to VC
NFX's guide to venture capital argues that VCs judge every early-stage pitch by one question: can this company return the fund? The article says investors test this through three criteria: whether the company can be big enough, whether it has defensible, differentiated "magic," and whether the founders are the right people. It explains that VC returns follow a power law, so seed investments often return 0 to 1x while a few must be large enough to return a fund of, for example, $100 million at 10% ownership, which requires a more-than-$1B outcome. The article also states that a standard fund pays about 2% annual management fees and 20% of profits, and that fund lifetimes run 10 to 15 years.
- Software with a Soul: AI's underestimated frontier
Software with a Soul, a term the article uses for AI products whose core value is "humanness," is presented as an emerging category of AI. The article says humanness is undervalued in founder and investor circles and that AI is making it possible to scale it. It names Similie, Prior Computers, Aaru, Anam, Beyond Presence and Delphi.ai as companies in this space. It predicts a split between anonymous automation, where humanness is secondary, and products where humanness is the product itself.
- The Network Effects Bible
Network effects, which make a product more valuable to each user as new users join, are presented as the best form of defensibility and the source of most value created in technology since 1994. The article states that founders who understand how network effects work will be better positioned to build category-defining companies. It describes 16 types of network effects ranked by strength, from physical networks like landline telephones to tribal and hub-and-spoke networks, and it explains concepts such as node density, link directionality, clustering and critical mass.
- The Wartime CEO
Nine leadership principles drawn from Israeli startup CEOs who led through six months of war are presented as useful for founders in any crisis or ordinary period. The principles were built from interviews with nearly a dozen CEOs of Israel's top multibillion-dollar companies, including Zeev Farbman of Lightricks, who said he had to redistribute work when only about one-third of his team was available due to reserve duty. Forter's Liron Damri said the company gave staff maximum flexibility so they could put personal safety first. NFX also announced Restart Startup Nation: FAST, an emergency investment track for early-stage Israeli startups, primarily in AI and biotech.
- The upgraded Go-to-Market playbook
The article argues that traditional marketing and growth channels have weakened, and that fast-growing companies now use a hybrid of product strategy and go-to-market strategy built on six tactics. It traces three phases: offline-to-online "pull" strategies (2000-2010) led by Trulia, Yelp and TripAdvisor, product-led growth with built-in virality (2010-2020) led by Slack, DocuSign, Dropbox, LinkedIn and Zoom, and a current "off-platform push" where companies build networks on third-party communities. Tactics include building in public, community-centric growth (Midjourney, Clay), self-service viral products (Cursor), prosumer targeting (ElevenLabs), open-source foundations (Snyk), and influencer marketing.
- Stablecoins are defense tech
Stablecoins are defense tech, and the author argues that digitizing the dollar through dollar-pegged stablecoins protects the dollar's position as the world's reserve currency. The author says the stablecoin market cap is near $150B and that 99% of stablecoin projects are pegged to USD today. The author also proposes that regulators trademark the USD by setting a government-backed definition of a digital dollar, with requirements such as fully backed reserves, auditable finances, and KYC-d founders.
- Does real identity matter for networks?
James Currier argues that networks whose profiles tie to real identity are more likely to endure and become valuable businesses, and that anonymous or pseudonymous networks tend to fizzle out. He points to LinkedIn, which launched real-name professional profiles in 2002, and Facebook, which did so in 2004, as the two most valuable online networks in the Western world. He says the early pseudonymous networks, including MySpace, Friendster, and Orkut, died out because their pseudonyms could be separated from users' real lives, leaving them weaker and easier for competitors to take. He adds that semi-pseudonymous networks such as Twitter, Reddit, and Snapchat survive, but their combined value is less than 10% of Facebook's market cap.
- Speed x AI
The article argues that speed is the main trait predicting early-stage founder success, and that most founders are slow because of their mindset rather than their tools or hours. It contends that founders' speed bars are set too low by past success and school-style deadlines, and it recommends resetting them to a benchmark 10 times faster. As an example, a web engineer at NFX built a working game in 21 days that reached 3 million users in 3 weeks, despite game developers saying it would take 12 months. The piece says founders should aim for 20 to 100 experiments per week and adopt mindsets such as borrowing existing ideas and dropping self-sabotage.
- Voice AI is Working. Here's Where It Wins First
Voice AI is becoming practical now, and the article argues that business customer service and sales calls are where it is winning first. It points to late 2024 as a turning point, when cloud APIs offered sub-300ms round-trip times, large language model integration, and falling costs. Numeo AI's VoiceFlow handles trucking dispatch calls, and the article says software costing $4,000 annually can replace a call center agent who costs $40,000 annually and works around the clock. It also names Boardy AI's career counseling and Autograph's voice-based life-story platform as examples of other early applications.
- Why Don’t Billionaires Invest In Longevity?
The author argues that billionaires and institutional investors are underfunding longevity research even though the underlying technology is ready, and that investing now would save lives and yield large returns. The article claims that 1% of today's billionaire wealth could give every person on earth more than 10 years of healthy life. It compares longevity to the WHO's smallpox eradication campaign, which cost about $300 million over 11 years, roughly 0.05% of today's billionaire wealth according to the article. It also cites several companies, including Altos Labs, New Limit, and Retro Biosciences, as working on cellular reprogramming.
- The Network Effects Manual: 16 Different Network Effects (and counting)
NFX presents a network effects map and manual covering 16 types of network effects, which it calls the strongest form of defensibility in the digital world. Its three-year study finds that network effects are responsible for 70% of the value created by tech companies since 1994. The manual traces the concept to AT&T chairman Theodore Vail's 1908 annual report, then covers Robert Metcalfe's law that network value scales with users squared and David Reed's 2001 argument that it scales exponentially.
- The "Intangible" Game: How VCs See You
A venture capitalist argues that investors often decide on a founder based on intangible qualities such as grit, a compulsion to build, authentic care about the market, and leadership, and that these perceptions often matter more than the technology or market. The author says investors' initial reaction to a pitch can determine the outcome, and that most VCs rarely state the real reason they pass. The author, who says they wrote 148 pitch decks as a founder and sees thousands of deals a year, cites Jensen Huang's statement that he would not start a company again as an example of how hard founding is. The article also quotes Naval Ravikant's line that you escape competition through authenticity.
- AI is like water
The article argues that generative AI application companies are like bottled water: many share the same underlying technology, so differentiation must come from brand, distribution, marketing and perceived value rather than the tech itself. It uses the history of bottled water to make the case, noting that Perrier repositioned a languishing French brand as upmarket in the mid-1970s, with sales rising from $3 million in 1975 to $200 million in 1979. It also cites Coca-Cola and Pepsi's Aquafina and Dasani, launched in 2001 using existing purification equipment, and Dasani's failure in the UK after it was found to be filtered tap water containing bromate.
- Ai First Marketplace
AI is opening a window for new marketplaces, much as GPS smartphones enabled Uber and Lyft, and it is expected to reshape marketplace architecture in four ways. AI automation can unlock supply in bottleneck fields like therapy, driving and tutoring, while low-cost AI copies could push the effective price of commoditized services toward zero. NFX-backed Tailorbird uses public visual data to generate architectural drawings, cutting a month-long renovation planning process to two weeks before routing users to bids for materials. The other features are better demand-side embedding, reimagined search covering earlier buying stages, and lower internal costs.