[BidClub_]

PERSON DIRECTORY

Martin Casado

Martin Casado appears in 28 indexed conversations across The a16z Show, 20VC, Invest Like the Best. This directory brings every appearance, source, TL;DR, digest, and transcript into one searchable feed.

28 EPISODES4 SHOWS
28 episodes
Language
The a16z ShowEN · 33 min

What It Takes to Run a Personal Assistant Around the Clock

Ken LiuMartin CasadoJoel de la Garza

Second aims to be a portable store of personal context that specialized agents can query, with background monitoring linking events to priorities without prompts.Ken Liu estimates ten always-on agents could cost $1,000 daily at “Opus 5.5” pricing versus $60 with DeepSeek Flash, making model routing central to the economics.Local data control and confidential GPU inference shape the security design, but zero data retention remains a policy promise and prompt injection cannot be definitively solved.

Invest Like the BestEN · 65 min

What Happens When the AI Boom Slows Down

Patrick O'ShaughnessyMartin Casado

Frontier models captured 80%+ of dollar-weighted demand over three years, yet Casado estimates an individual model’s commercial lifespan at just three to six months.Startups may still pay 10 times more for the smartest model, sustaining pricing power even as coding workflows saturate.If demand saturates, value could shift to open source, a transition Casado says “nobody knows” how it would unfold.

The a16z ShowEN · 45 min

What iPhone Packaging Reveals About the Robotics Opportunity

Josh GruensteinMartin CasadoRaghu Raghuram

Hand-assembled iPhone packaging illustrates Gruenstein’s claim that nine of ten US factories and warehouses lack robots.Tutor’s $16-an-hour offer requires no upfront customer CapEx; contracts take 40–60 days and deployment about 40 more, Gruenstein says.Paid work builds its data loop, but ongoing costs now exceed hardware costs, making inference, teleoperation and reliability central economic tests.

The a16z ShowEN · 59 min

Why AI Can Write Code But Can’t Keep It Running (Yet)

Jad NaousMartin Casado

AI coding has moved beyond autocomplete: Martin Casado built data-lake auto-promotion production code in 10 days versus roughly six months historically, a 10- to 20-fold gain Jad Naous says favors engineers already strong in systems.Grepr’s 90% telemetry-cost reduction offers a concrete commercial wedge, while anomaly-led diagnosis works internally and with a few customers; autonomous cluster management remains “not yet.”

The a16z ShowEN · 32 min

Recursive Self-Improvement and the Future of AI

Martin CasadoDhaval AdjodahOwen He

RSI founders Dhaval Adjodah and Owen He report early signs of performance rising as resource use stays flat or falls, with the model itself as optimizee.They cite retraining GPT-2 in hours for about $100 on the same GPU and data, and target inference efficiency across the stack.Casado has not seen equivalent GPT-5 evidence, leaving scalability beyond fast-evaluation tasks unresolved.

The a16z ShowEN · 42 min

How Jev Turns AI Into Software That Gets Things Done

Ben HorowitzMartin CasadoDiogo Almeida

Jev targets AI’s missing software layer: a natural-language primitive that joins a state machine and selects actions with confidence, making software itself more capable rather than merely speeding code production.Its economic unlock is reliability—“the same intelligence every time”—which could let incumbent SaaS vendors automate high-return workflows across existing distribution, while state consistency, security, cost, speed, and strong guarantees remain unresolved.

The a16z ShowEN · 55 min

Why AI’s Next Breakthroughs Could Come from Outside the Big Labs

Erik TorenbergAaron LevieMartin CasadoSteven Sinofsky

Agent swarms expose a security gap: controls built for trustworthy, capacity-constrained humans fail when software exhausts APIs and confuses good tasks with bad ones.That implies upgrades across identity, authorization, observability, operating systems, networks, and granular data permissions.Meanwhile, Jevons-style option selection could shift innovation beyond frontier labs by making models faster, cheaper, and more accurate interfaces for traditional software.

The a16z ShowEN · 67 min

Why Would AI Companies Want to Slow Down?

Ali GhodsiMartin CasadoSarah Wang

Frontier AI’s immediate enterprise risk may be cyberattack acceleration: CVE-to-exploit time fell from 2–3 years in 2018–19 to “basically hours” now, while most organizations lack automated detection and threat hunting.Ghodsi says recursive self-improvement shows none of four required conditions; enterprises need context and cost control more than smarter models, though a frontier freeze would be disastrous to the labs.

The a16z ShowEN · 44 min

Why World Models Could Change Robotics, 3D, and Creativity

Fei-Fei LiJustin JohnsonBen MildenhallMartin Casado

World Labs’ Atlas introduces novel-view prediction as a foundation-model primitive, unifying 3D reconstruction and generation through camera-conditioned outputs.Three iPhone shots can replace 100–300 room photos, a claimed 50–100× capture reduction, while Atlas targets robotics’ data bottleneck.The open commercial test is industrial-grade editability and control without degrading quality; dynamics are claimed latent, but this remains a milestone beyond entertainment.

The a16z ShowEN · 54 min

How AI Is Reinventing Computing from Chips to Power

Ben HorowitzMartin CasadoRaghu RaghuramErik Torenberg

a16z’s new AI infrastructure fund captures a founder migration into hardware, with top-founder hardware pitches rising from roughly 3–5% to “north of 20% or 30%.”Hyperscaler capex, booked-out GPUs and resale premiums support opportunities across chips, power and cooling, while grid shortages, regulation and uncontrolled agent spending remain constraints.