What was the viral Substack fanfiction about AI doomsday and market impact?

The Substack piece imagined a 2028 global AI crisis where initial AI-driven efficiency improves margins, followed by consumer constraints, leading to a 'death spiral' of demand and layoffs. It caused broad market selloffs in financial stocks before facing scrutiny over attribution and virality.

Who is Derek Thompson and what did he say about AI narratives in the market?

Derek Thompson argued that much of the AI market debate is science-fiction masquerading as analysis because there is insufficient real-time macroeconomic data. He emphasized that the conversation is a marketplace of competing narratives rather than solid analytics.

What did Jason describe about deploying AI agents in his business?

Jason described training multiple OpenClaw agents to automate repetitive tasks (like lead generation, email, and scheduling) and said these agents operate around the clock, improving efficiency by 10–20% without adding headcount, effectively redeploying humans.

What is Kimmy 2.5 and why is it significant in the discussion?

Kimmy 2.5 is an open-source AI model referenced as capable of performing a large percentage of tasks. The group uses it as an example of cheaper, on-premise AI tooling that reduces token costs and enables faster deployment of AI-enabled workflows.

Who were announced as keynote speakers for the All-In Summit?

Two keynote guests were announced: Dan Loeb (Third Point founder) and Sarah Frier (OpenAI CFO). The conversation framed these names as high-profile figures for the event.

What was the gist of the Bill Gurley book plug?

A plug for Bill Gurley’s book Running Down a Dream was made, with a call to buy multiple copies to share with younger people and a parent, framed as inspirational and practical for future leaders.

What happened with the SCOTUS tariffs ruling?

SCOTUS rebuked the administration on tariffs, suggesting there are multiple bases in law to sustain tariffs and indicating that refunds to importers may be required in some cases, while leaving room for future tariff mechanisms to remain in place.

What is the Yamanaka-factor biotech update discussed in Science Corner?

The segment described Life Biosciences’ phase-one trials delivering Yamanaka factors via viral delivery to the eye to rejuvenate retinal cells and potentially restore vision, marking the first human applications of Yamanaka factors in this context.

What does the episode say about the 'organization vs. pie' view of the economy?

The hosts argue the economy is not a fixed pie; technology acts like rain in a garden, expanding productivity and the economy as a whole, potentially creating new jobs and opportunities rather than simply redistributing wealth.

Software Stocks Implode, Claude's Hit List, State of the Union Reactions, Trump's Tariff Pivot

All-In Podcast

Entertainment7 min read82 min video

Feb 28, 2026|433,418 views|8,028|1,245

chamath david sacks david friedberg jason calacanis all in podcast tech news politics election stocks stock market palihapitiya

Save to Pod

Key Moments

On this page

TL;DR

AI market nerves, viral fiction, SAS shift, and data-center energy policy.

Key Insights

Markets have shifted from a 'when cash flows disappear' mindset to an 'if cash flows are durable' mindset, driving lower multiples (PE, revenue) and higher discount rates (WACC) as a hedge against uncertainty.

A viral Substack fanfiction about an AI-driven death spiral sparked a market reaction, but experts argue no one can know AI's macro effects with certainty; narratives matter but should be grounded in data.

AI productivity could transform software and knowledge work, but not necessarily eliminate jobs; demand for software engineers may rise as tooling lowers barriers and enables new roles.

Practical AI deployment—building and managing agents—can yield meaningful efficiency gains (e.g., 10–20%+) without headcount growth, reshaping labor dynamics and a company’s operating model.

Data-center expansion faces energy policy challenges; a 'ratepayer protection' pledge seeks to keep residential prices stable while enabling scale, with behind-the-meter generation as a potential solution.

Global competition for data-center infrastructure intensifies; places like the Middle East are expanding aggressively, while US policy must balance incentives, grid costs, and national economic interests.

INTRODUCTION: CONSPIRACY CORNER AND MARKET TALK

The episode begins in classic All-In fashion with a tongue-in-cheek conspiracy corner, then pivots to rigorous market discussion. The crew notes that a viral AI-age Substack piece—depicting a 2028 death spiral from rapid AI adoption—captured attention and moved markets, but they stress the difference between entertaining narratives and robust analytics. The conversation then anchors on concrete AI developments—Anthropic's Claude plugins, code security, and database modernization—that have real implications for valuations. The mood is skeptical but curious: not forecasting doom or doom-mongering certainty, but unpacking how investors price uncertain futures when AI progress could alter cash flows, cost of capital, and competitive dynamics.

WHEN TO IF: SHIFT IN VALUATION AND RISK MODE

The core market thesis is that we have moved from a 'when' world (will these cash flows endure?) to an 'if' world (will these cash flows endure at all?). Hedge funds are aggressively degrossing, shrinking long and short bets to reduce risk, which creates broader downward pressure on tech equities. The guests explain the math of discounting future cash flows: as uncertainty rises, the market re-prices earnings, cash flows, and growth, driving lower price-to-earnings multiples, lower revenue multiples, and higher WACC. They emphasize that the market now asks whether cash flows are durable over the long horizon, not just whether they exist, which reshapes risk premiums and investment theses.

DOOMER NARRATIVES AND MARKET ANALYSIS: VIRAL FICTION VS. REAL DATA

A key thread is the tension between doom-and-gloom narratives and empirical analysis. The viral Catrini/AI death-spiral piece is scrutinized for attribution and potential market impact, with a counterpoint: none of us truly knows the macro trajectory of AI two or even ten years out. Derek Thompson’s argument—that AI discourse often reads like science fiction masquerading as analysis—frames this as a marketplace of competing narratives rather than a settled forecast. The panel also revisits SaaS economics, noting that traditional predictability (ARR, net dollar retention) is challenged by AI’s potential to alter pricing, churn, and growth dynamics, forcing investors to tolerate greater uncertainty.

THE SAS CONTRADICTION: DRIVING PRODUCTIVITY OR DESTROYING MARGINS

The discussion delves into whether AI preserves or disrupts the SaaS model. AI may unlock productivity, but it also complicates valuation that used to rest on predictable ARR multiples and strong net retention. The panel cites data points—Anthropic’s high-engineer salaries despite claims of obsolescence, job-market trends in software engineering, and the paradox of increased productivity without immediate job losses. They argue that SaaS was historically a dependable annuity, but AI disrupts that predictability, potentially compressing margins while opening new pricing models and product paradigms. The debate highlights that a broader rethinking of software economics is underway, not a simple conclusion of doom or bloom.

ANTHROPIC UPDATES AS MARKET SIGNALS: PLUGINS, SECURITY, AND COBALT

Anthropic’s product milestones—Claude Co-Work with Thomson Reuters, legal plugins, Claude Code Security, and database modernization—are framed as tangible signals rather than abstract hype. While these moves support broader AI adoption in enterprise workflows, the stock market reacts to both the news and the broader uncertainty about AI’s long-run impact. The conversation clarifies that such updates can boost productivity and enterprise credibility, yet they do not settle questions about whether AI will compress or expand total demand for software, nor how companies will repricing risk in a rapidly evolving landscape.

THE DOOM OR ABUNDANCE: DEREK THOMPSON AND THE MANAGEABLE UNCERTAINTY

A central exchange centers on how to ground expectations amidst divergent narratives. Derek Thompson’s critique—AI’s macroeconomic effects are highly uncertain and often rendered as literary speculation—serves as a counterweight to doom fantasies. The panel acknowledges that scenario planning is essential but emphasizes the need for data-driven approaches. They discuss how AI could enable a productivity leap without triggering massive layoffs if jobs re-skill around agent-based systems, creating new value rather than merely replacing workers.

AI-ENABLED KNOWLEDGE WORK: FROM CODING TO COORDINATING AGENTS

The conversation shifts to a practical realm: AI tooling shifts the knowledge-work landscape. Instead of eliminating engineers, AI can empower a broader set of workers to produce more with the same or fewer people. The participants discuss how agents can be trained to perform tasks, augment decision-making, and automate routine work. They highlight the potential to redeploy personnel into higher-value roles and to launch new services or products with greater speed, effectively expanding the productive capacity of organizations without proportional headcount growth.

OPENCLAW AND THE NEW WORKFLOW: HANDS-FREE OPERATIONS IN PRACTICE

A concrete example is provided: teams using Claude-based agents to scan competitors, map advertisers, and sync data across CRM tools. Agents can summarize weekly activities, manage emails, and generate performance insights, turning previously manual tasks into continuous automation. The result is 10–20% efficiency gains and a reallocation of human talent toward design, strategy, and governance of AI systems. The segment emphasizes that the real challenge is building scalable processes and training agents to handle complex workflows, not simply adopting new software.

THE AGENTS WAVE: CASE STUDIES OF AUTOMATED ROUTINES

Beyond internal efficiency, examples extend to content production and marketing—Automating podcast clipping, subtitle creation, and distribution workflows; analyzing thousands of transcripts to identify relevant ads; and generating weekly performance reports. These case studies illustrate a broader trend: knowledge work can be systematized through agents that learn, adapt, and optimize over time. The discussion also notes the need for governance—ensuring agents are aligned with business goals, secure, and auditable—while recognizing the potential to transform entire departments.

DATA CENTERS AND ENERGY POLICY: BEYOND THE GRID

The State of the Union energy policy discussion centers on ratepayer protection promised by policymakers to keep residential electricity prices stable as data centers expand. The debate covers the economics of behind-the-meter generation, grid modernization, and the potential for surplus energy from data centers to feed back into the grid. The panel warns of utility capex dynamics and regulatory risk that could offset some efficiency benefits. The takeaway is that policy design will determine whether data-center growth translates into net economic gains for consumers or higher fixed costs.

GLOBAL COMPETITION AND POLICY TENSIONS: US VERSUS THE WORLD

The conversation broadens to geopolitics and the global race for data-center capacity. They compare the U.S. approach to energy policies with aggressive expansion in the Middle East, including Saudi Arabia and the UAE, where data-center investments are treated as strategic economic opportunities. The panel argues that delaying or over-burdening regulatory processes risks ceding the data-center boom to other jurisdictions. Texas and other investor-friendly environments are highlighted as examples where rapid buildouts are possible, reinforcing the view that policy must balance resilience, grid integrity, and economic opportunity.

FUTURE SCENARIOS: PATHS AHEAD, RISK MITIGATION, AND TAKEAWAYS

The closing lens is forward-looking: with AI accelerating productivity, how should individuals and firms prepare? The panel emphasizes risk management through diversification of bets on technology, careful governance of token economics, and a willingness to redesign compensation and hiring strategies around agent-based workflows. They stress that scale-driven data-center growth will hinge on policy alignment, energy economics, and international competition. The takeaway is a pragmatic optimism: embrace AI-enabled productivity while proactively addressing market uncertainty, workforce transformation, and the policy environment.

Mentioned in This Episode

●Tools & Products

●Books

●People Referenced

Common Questions

The pledge would have major AI data centers fund their own power needs behind the meter so residential electricity rates don't rise. It suggests large tech companies should cover the incremental costs of their energy usage rather than households bearing the burden, and it includes the potential for data centers to backfeed excess capacity to the grid as they scale.