Big Technology Podcast – Friday Edition Episode: "Why OpenAI Killed Sora, Did Apple Just Save Siri?, Meta’s Big Loss"
Host: Alex Kantrowitz
Guest: Ranjan Roy (Margins)
Date: March 28, 2026

Episode Overview

This episode dives deep into major recent shifts in the tech ecosystem:

• OpenAI’s decision to discontinue Sora, their consumer video generation product, and pivot towards enterprise and coding applications, with huge implications for the company’s strategy and the broader AI race
• Apple’s changes to Siri and what the new “open” approach to third-party AI might actually mean
• Meta’s (and YouTube’s) landmark legal loss regarding harms to youth, opening Big Tech up to further litigation
• Upcoming battle of next-gen AI models between OpenAI and Anthropic
• The shelving of OpenAI’s erotic chatbot (“adult mode”) and what that says about AI’s future
• Broader impacts on the tech industry, including the stock market, Microsoft’s AI stumbles, and more

The tone is insightful, bantering, and occasionally irreverent, with both hosts blending reporting, industry insight, and tech-wonk humor throughout.

Key Discussion Points & Insights

I. The Death of Sora: Why OpenAI Is Killing Video Generation

[01:45 – 14:00]

Summary:

• OpenAI is shutting down Sora, its video generation platform that once topped the App Store and drew significant attention.
• The real drivers are strategic focus and resource allocation as the company heads towards a possible IPO.

Core Reasons for Sora’s Demise:

• Lack of enduring consumer appeal: The novelty wore off quickly (“maybe most people just want to watch instead of create” – Alex, [02:11]).
• Homogeneity of output: Videos generated were bland, “the average of averages.”
• Strategic focus on enterprise and coding:
  • Direct from Greg Brockman (OpenAI President): “The Sora models… are a different branch of the tech tree than the core reasoning GPT series. …pursuing both branches is very hard for us. …the results in these text GPT style models that do all the reasoning…and we have to decide where to put the computer… to do that in a world model way would really limit the company's ability to progress, to make progress in the area they see as most promising. And that is why Sora is being deprioritized.” (Alex paraphrasing, [05:51–07:28])

Notable Moments:

• Ranjan’s nostalgic and irreverent “greatest Sora hits” – using AI-video for “a chicken and a horse running around a toilet bowl.” [04:17]
• “You burned an entire rainforest just to get the horse running around.” (Alex, [04:52])

II. OpenAI’s Strategic Pivot: The Agentic/“OpenClaw” Era

[08:02 – 14:35]

Summary:

• Both OpenAI and Anthropic are moving toward “agentic” AI: one-stack assistants (think: your desktop, phone, or work life managed by AI agents).
• No more split between consumer (OpenAI) and enterprise (Anthropic); now both are targeting knowledge work automation and personal assistants.
• Ranjan notes SaaS players (Writer, Notion, Sierra, etc.) are moving in, customizing models rather than just making API calls.

Quote:
“It’s not just OpenAI and anthropic… everyone has recognized that's the prize… more and more it's that a lot of tools to date it was just that API call. I think more and more people are going to start either customizing or fully training on the foundation model side.” (Ranjan, [09:41, 14:35])

Debate:

• Alex is skeptical that smaller players will match the scale or foundational strength of OpenAI/Anthropic.
• Ranjan predicts more companies will build/train unique models, increasing competition.

III. The New AI Model Race: Anthropic vs. OpenAI

[24:08 – 31:36]

Anthropic:

• Accidental data leak revealed “Claude Mythos” (aka Capybara), its new top model, described as a “step change” with much stronger coding, reasoning, and cybersecurity performance ([25:54]).
• Both hosts discuss the incremental/exponential nature of model improvements.

OpenAI:

• Next model codenamed “Spud” is nearly complete. Altman claims it can “really accelerate the economy” ([28:26]).
• Hosts riff on uninspiring code names and AI marketing hyperbole.

Quotes:

• “He’s literally—the team believes—can really accelerate the economy. …It's called Spud.” (Ranjan, [29:58])
• “Just as the world implements today's models…they’re building better models than these—some that they say are sizable leaps. Like, it is one of those moments where you sit back and just go, this is crazy.” (Alex, [29:27])

IV. Apple & Siri: Did Apple Really "Save" Siri?

[32:51 – 36:55]

Summary:

• Apple will let Siri connect with rival AI assistants (through the App Store) in iOS 27.
• Reality check: This integration won't make Siri itself smarter; it’s a conduit to other services and a way for Apple to take a cut from third-party AI subscriptions.

Quotes/Moments:

• “People were—and I initially took this as maybe Siri is saved—but then I realized that it's just going to be the same disappointing user experience…” (Alex, [34:13])
• “This made Spud look like pure poetry of a name.” (Ranjan, [34:13])
• “If this is just like, well, buy your Gemini through the App Store and then we're going to take a cut. This is troubling.” (Ranjan, [35:24])
• Both express pessimism about WWDC unveiling a true AI breakthrough for Siri.

V. Meta’s Landmark Legal Loss: The New Era of Platform Liability

[37:00 – 44:38]

Event:

• California court found Meta and YouTube liable for harming a young user, awarding multimillion-dollar damages.
• Critical precedent: It challenges Section 230 protections when it comes to platform design causing harm (not just content).

Implications:

• Potential for thousands of similar lawsuits.
• Market hammered Meta’s stock.
• New Mexico loss for Meta brings $375M extra penalty ([38:48]).
• The “algorithm is the cigarette”: Algorithmic recommendation, not just “posting photos,” is the harmful component ([43:03]).

Quote:

• “The finding validates a novel legal theory that social media sites or apps can cause personal injury. I mean that's a huge deal.” (Ranjan, [40:01])

VI. Social Media Algorithm Harm: Is the Algorithm the Real “Tobacco”?

[43:03 – 49:10]

Debate:

• Ranjan: “There is one culprit. It's the algorithmic recommendation of content. That's it.”
• Alex: Plays devil’s advocate, likening it to cigarette analogy—hard to prove one-to-one causality, but clear aggregate harm.

Notable Quotes:

• “That's like saying is smoking responsible for lung cancer, or could obesity or environmental factors and air quality. …We all use social media.” (Ranjan, [46:00])
• “The counter argument to Meta’s argument is…it’s not that smoking cigarettes lead directly to cancer, it’s that smoking cigarettes are a known carcinogen.” (Alex, [46:40])

VII. Tech Market Woes & Microsoft’s AI Dip

[52:24 – 55:03]

Summary:

• Tech stocks, especially the Big Five, are having their worst week in nearly a year.
• Major reasons: War worries, legal problems, and lack of near-term AI profits.
• Microsoft is notably underperforming despite its early OpenAI partnership; Copilot isn’t generating significant paid uptake, and AI innovation has stalled.

Quotes:

• “They have fallen behind. There’s nothing exciting coming out. …On the whole AI thing it is pretty crazy that they had…OpenAI…very early and now still they're not really anywhere notable.” (Ranjan, [53:56])

VIII. OpenAI’s Erotic Chatbot Shelved: “Adult Mode” RIP

[55:03 – 61:25]

Event:

• OpenAI indefinitely shelves “adult mode” after internal and investor backlash over potential harm and societal impact ([55:03]).
• Real concern: Can’t build trust with business and enterprise partners if “erotic chatbots” are in the lineup.
• Hosts riff whether other companies might take up the mantle using OpenAI APIs.

Quotes:

• “I don't want to be the morality police and say you shouldn't be able to…have cyber sex with your chatbot, but speaking of businesses that OpenAI shouldn't be in, this seems like one of them.” (Alex, [56:17])
• “Are we losing the weirdness of Sora and potentially erotic chatbots now that everyone's just making claws?” (Ranjan, [56:35])

Segment Closes with Humor:

• Discussion of “ChatWet”—a hypothetical adult chatbot app name, call-back to last week’s running gag ([59:55]).

Engaging and Memorable Moments

• Ranjan’s deadpan: “Sora is dead. Erotic chatbots are no more. Apple might be fixing Siri. This feels like a good week.” ([01:30])
• Ranjan: “You burned an entire rainforest just to get the horse running around.” ([04:52])
• On AI model names: “The name Spud did not jump out to you as what the hell is going on?” ([29:27])
• “This made Spud look like pure poetry of a name.” (on Siri news, Ranjan, [34:13])
• Ranjan’s advocacy for default reverse-chronological social media feeds as a fix for algorithmic harm ([45:10])
• Recurring banter about “dry chat” and “wet chat” as hilarious analogies for chatbot interactions ([61:00+]).

Important Segment Timestamps

| Time | Segment/Topic | |-----------|------------------------------------------------------------------| | 01:45–14:00| OpenAI kills Sora; deeper look at platform shift | | 14:00–18:07| OpenAI & Anthropic’s strategy convergence; agentic AI | | 24:08–31:36| Next AI model war: Anthropic (“Claude Mythos/Capybara”), OpenAI (“Spud”) | | 32:51–36:55| Apple/Siri “improvements” and analysis | | 37:00–44:38| Meta/YouTube found liable—Section 230 challenged | | 43:03–49:10| Social media algorithms as the “tobacco” of tech | | 52:24–55:03| Tech stock slide and Microsoft’s AI dilemma | | 55:03–61:25| RIP OpenAI “adult mode” – debate over AI liability and weirdness| | 61:40–end | Closing banter, next week’s preview |

Conclusion

This episode captures a major inflection point in Big Tech and generative AI:

• OpenAI and Anthropic are streamlining for the agentic AI assistant “endgame.”
• Consumer side projects—even viral ones like Sora—are out if they pull focus from the foundational race.
• Siri isn’t being fixed; Apple’s moving to extract revenue from third-party AI instead of advancing core capabilities.
• Meta, for the first time, faces legal precedent that algorithmic design can be ruled personally harmful.
• Next-gen model competition is speeding up, but real-world trust and use remain big barriers.

In sum: The AI arms race is sharpening, consumer fun is giving way to enterprise utility, and the legal and trust landscape for big tech is shifting underfoot—setting the stage for deeper changes ahead.