Software’s Epic Comeback, Meta’s AI Layoffs Blunder, South Korea Stock Market Chaos

Summary

Alex Kantrowitz and Ranjan Roy open with the software rebound, arguing that fears of an imminent SaaS apocalypse were overdone and that enterprise software disruption by AI will likely unfold on a much longer timeline. They dissect Salesforce’s rally and the Anthropic/Cloudforce optics, with Roy calling much of the announcement a marketing and PR move rather than a genuinely new technical integration. The discussion then shifts to Meta, where Reuters-reported internal testing of an AI-heavy org design showed more code output but not higher productivity, alongside rising reliability and security incidents. The episode closes with Meta’s $18 billion social-media addiction settlement and South Korea’s volatile AI-driven stock surge and selloff, which the hosts treat as a cautionary tale about leverage, retail speculation, and the risk of going all in too early on AI narratives.

Notes

• Software stocks rebounded sharply after earlier "SaaS apocalypse" fears.
• The software rebound is framed as a market correction, not proof AI disruption is gone.
• Salesforce, Adobe, and Microsoft were cited as symbols of the sector's recovery.
• Alex says investors were too quick to assume AI would kill enterprise software overnight.
• The software trade is described as a full round trip from collapse fears back to near-flat performance.
• Salesforce's stock pop was tied to strong earnings and improved guidance.
• The Salesforce-Anthropic Cloudforce announcement is portrayed as mostly messaging.
• Ranjan says the Claude/Salesforce integration was already possible before the announcement.
• Marc Benioff is credited with exceptional marketing skill.
• Ranjan reads Dario's presence as IPO-related optics and investor pressure.
• Alex says Cloudforce means using Claude to query Salesforce records.
• The strategic risk for Salesforce is the UI layer moving to an assistant.
• Roy says Anthropic still needs a Salesforce-disruption story to justify a huge valuation.
• Ranjan says Salesforce's governed data layer has value, but the long-term economics are uncertain.
• The hosts argue software disruption will be slow because human adoption lags model capability.
• Meta's pilot used small pods, AI tools, four-week sprints, and fewer management layers.
• The project name, Project OT, is described as a bad codename that may have doomed morale.
• Meta's internal data reportedly showed more code, but not proportionate productivity.
• Meta's AI coding raised infrastructure and security concerns.
• Roy thinks incentives may have pushed employees to optimize AI usage rather than output quality.
• Zuckerberg reportedly halted company-wide layoffs for the year but linked headcount to compute spend.
• Alex questions cost-cutting rhetoric from a highly profitable Meta.
• Meta settled child-addiction lawsuits for $18 billion over a decade.
• Roy says the settlement validates long-running ideas to demetricate social media.
• Ranjan calls the teen restrictions the right direction, even if Meta has business motives.
• Roy says a one-hour limit would be a competitive shot at TikTok and YouTube if they followed.
• Alex says YouTube has become a dominant entertainment network across media categories.
• Alex estimates friends-and-family content is a small share of Instagram time.
• Alex says the Reels feed is the core Instagram experience.
• Ranjan says YouTube Shorts dominate even inside YouTube.
• South Korea's market story is framed as an AI-bubble warning.
• Leveraged single-stock ETFs can wipe investors out on a large drawdown.
• Alex links South Korea's retail trading to agency-seeking behavior and wider social dynamics.
• Roy says bubble labels are relative, but South Korea looks like an early bubble-ending cycle.
• Alex closes with a warning about AI hype, turbulence, and going all in too early.