Latest Interviews

Mandela's legacy: 25 years on

This analysis examines the deepening conflict surrounding Nelson Mandela's legacy and the "Born Free Generation" in South Africa, highlighting how unfulfilled post-1994 promises regarding housing, education, and economic justice have fueled systemic inequality. Decades of ANC rule have intensified these issues through corruption and Black Economic Empowerment policies that critics argue have only entrenched a new elite while failing to dismantle apartheid-era wealth disparities. Consequently, widespread social unrest and plummeting public trust underscore a critical divide between the state's historical pledges and the lived reality of a population facing high unemployment and a severe education crisis.

Should I Use a Dev Shop? - Michael Seibel

Michael Seibel

This analysis challenges the common startup strategy of hiring development shops to build an MVP, arguing that outsourcing frequently leads to budget overruns, timeline delays, and code that becomes obsolete during iteration. By contrasting the statistical failure rates of this approach with the superior speed and capital efficiency of teams that secure technical co-founders early, the discussion highlights how investors prioritize a team's ability to rapidly build and pivot over a polished but static product. Ultimately, while dev shops remain viable for non-VC business models, the recommendation for founders targeting high-growth valuations is to recruit a technical co-founder immediately despite the initial difficulty in doing so.

MIT 6.S093: Introduction to Human-Centered Artificial Intelligence (AI)

The presentation argues that learning-based artificial intelligence will supersede optimization models but requires "machine teaching" and continuous human supervision to ensure safety, fairness, and explainability. Key strategies include active learning algorithms that minimize data requirements, reward engineering to align systems with societal values, and uncertainty signaling through ensemble disagreements to trigger human intervention in high-stakes domains. These human-AI collaborations aim to overcome persistent perception challenges in face and emotion recognition while scaling autonomous technologies to societal levels where safety and symbiosis are paramount.

Towards a Decentralized Storage Network for the Blockchain Computer

Ali Yahya, Ben Fisch, Bruce Schneier

Decentralized storage networks solve the critical challenge of verifying file integrity across permissionless nodes by replacing institutional trust with cryptographic protocols like Proof of Retrieval and Proof of Replication. These mechanisms transform underutilized personal device capacity into a censorship-resistant infrastructure where economic incentives compel nodes to dedicate actual disk space rather than simulating storage. By enforcing high tightness constraints that disproportionately burden large centralized providers, the system ensures a scalable, fault-tolerant foundation for future decentralized applications.

Five Open Problems for the Blockchain Computer

Ali Yahya, Frank Chen, Ben Fisch

The event analyzes the emerging crypto paradigm where value shifts from raw performance to emergent trust, utilizing metrics like instructions per second and decentralized storage proofs to overcome physical network latency. It highlights critical infrastructure innovations, including Blockchain Distribution Networks and Proof-of-Stake security, that enable censorship-resistant finance and gaming applications. Positioned in the early stages of development, the sector aims to solve unsolved computer science problems by complementing centralized web models with systems optimized for trust rather than speed.

The Future of Decision-Making: 3 Startup Opportunities

Jad Naous, Frank Chen

Enterprise digital transformation is pivoting from manual migration to full automation, driving a fundamental shift from Business Intelligence to real-time Operational Intelligence. This transition empowers non-technical staff with self-service tools to make immediate, data-driven decisions while forcing specialized ETL layers to evolve for continuous monitoring. Consequently, startups targeting high-volume, low-margin sectors like construction and oil and gas are capitalizing on these efficiency gains to deliver value through vertical-specific solutions and deep domain expertise.

Google Photos Product Lead and Bump Cofounder David Lieb with Gustaf Alströmer

David Lieb, Gustaf Alströmer, Craig Cannon, Lamide Akomolafe

Founded in 2009 by David Lieb and co-founders, the contact-sharing app Bump achieved viral success through word-of-mouth, reaching 150 million downloads and securing $17 million in venture funding before pivoting to address its lack of a scalable monetization model. After transitioning to focus on photo sharing with its successor apps Flock and PhotoRoll, the technology was acquired by Google in 2011 and integrated with Picasa and advanced AI to create Google Photos. This acquisition transformed the platform into a free, AI-driven ecosystem that automatically backs up, organizes, and curates billions of users' memories, fundamentally shifting the product from manual sharing to proactive digital memory management.

How to Get and Test Startup Ideas - Michael Seibel

Michael Seibel

This session challenges the notion that startup ideas must be perfect at inception, using Justin Kan's co-founding of Twitch to illustrate how prioritizing deep personal connection to a problem over the concept itself drives resilience. Kan advises founders to maintain "problem books" rather than idea logs, validate issues through direct community impact, and rigorously handpick early users to test minimum viable products. Ultimately, the discussion emphasizes that successful entrepreneurs must fall in love with the customer's pain point rather than their initial product, ensuring the team remains uniquely qualified to solve a specific, verified need.

A CS Education That's Free Until You Get a Job - Austen Allred of Lambda School

Austen Allred, Craig Cannon, Steven Klaiber-Noble, David Kofoed Wind, Dave Dawson, Dayo Koleowo

Lambda School is an online vocational program utilizing a $48 million-funded Income Share Agreement model to train software engineers without upfront tuition, focusing on candidates who demonstrate high work ethic over traditional academic credentials. Led by CEO Austin Allred, the institution distinguishes itself through a unique admissions process, an in-house team of 37 interview sourcers, and a curriculum that integrates soft skills training from the third week to achieve over an 85% graduation rate. Having survived a previous Series A collapse and a Kickstarter book campaign, the company now projects a net financial advantage for its graduates compared to four-year degree holders while exploring the expansion of its ISA model into nursing education.

a16z Podcast | Incenting Innovation Inside, Loonshots to Moonshots

Safi Bahcall, Vijay Pande, Sonal Chokshi

This presentation defines "loonshots" as high-risk creative processes distinct from their destination goals, arguing that established organizations often suppress them due to misaligned incentives that favor political survival over innovation. By applying four structural control parameters—equity fraction, return on politics, skill fit, and management span—leaders can engineer a dynamic equilibrium that allows creative "artists" and execution-focused "soldiers" to coexist and successfully transfer ideas to the market. Historical case studies of figures like Vannevar Bush, Steve Jobs, and institutions such as Xerox PARC illustrate how shifting from "Moses" leadership to "gardener" models is essential for sustaining the organizational phase transitions required to turn early failures into transformative moonshots.

a16z Podcast | How Many Taps in the Apple (Plus) Tree?

Benedict Evans, Steven Sinofsky, Stephen Sanofsky

Apple is leveraging its installed base of over one billion iOS devices to execute a subscription stacking strategy that prioritizes human curation and privacy across diverse verticals including news, gaming, and financial services. By integrating acquisitions like Texture and partnering with high-profile creators for Apple TV+, the company aims to increase user stickiness and device switching costs without disrupting existing industry structures or forcing a unified mega-bundle. This approach targets markets with high spending power to incrementally improve complex user experiences while allowing individual services to achieve product-market fit independently.

a16z Podcast | For the Billions of Creatives Out There

Brian Koppelman, Marc Andreessen, Sonal Chokshi

Writer-directors Brian Koppelman and David Levine discuss their "write first, pitch later" methodology and specific strategies for maintaining creative control while navigating Hollywood power dynamics. The conversation explores how industry shifts have transformed mid-budget films like *Rounders* into expansive television narratives, emphasizing the depth required for modern storytelling in shows like *Billions*. Additionally, the pair details their 50-50 partnership model, the psychological tools they use to manage anxiety, and their philosophy that rigorous internal engagement outweighs external validation or immediate commercial success.

a16z Podcast | Lessons Learned from Chinese Education Startups

Connie Chan, Frank Chen

This analysis highlights China's $40 billion online education sector, where a mobile-first infrastructure and AI-driven cost reduction are transforming lifelong learning for the 26–35 demographic. Leading platforms like LingoChamp and VIP Peilin demonstrate how machine learning can replace human labor to achieve high gross margins while offering affordable, bite-sized content. Strategic recommendations emphasize shifting from ad-supported models to direct creator monetization and per-course pricing to sustain a market that views education as a primary household expense.

a16z Podcast | What's in the Water at the George Church Lab?

George Church, Jorge Conde, Hannah

George Church analyzes the 2018 Chinese CRISPR baby case as a pivotal, government-tolerated precedent for germline editing that bypassed ethical moratoriums through a technical strategy of *CCR5* knockout for HIV resistance. He predicts that, similar to the 1978 test-tube baby phenomenon, this "n of 1" success will accelerate adoption and regulation rather than halt research, shifting the field from treatment toward preventative genomic medicine. While Church advocates for robust privacy frameworks like zero-dollar genomes and encryption to overcome security barriers, his broader vision prioritizes environmental applications like mammoth de-extinction over non-therapeutic human editing.

Network Effects: Measure Them, Nurture Them (3 of 3)

D'Arcy Coolican, Li Jin, Frank Chen

Researchers established a comprehensive framework of 16 metrics to quantify network effects as dynamic variables rather than binary states, prioritizing leading indicators like retention cohorts and power user curves. The analysis defines network strength through the prohibitive cost required for competitors to reconstruct a platform's ecosystem, ranging from massive library investments to niche service replication. Ultimately, the study identifies pricing power as the definitive lagging indicator, where sustained higher take rates against competitors validate the platform's entrenched loyalty and multi-tenanting decline.