Jun 14, 2025 · 59m · big-technology
Sam Altman’s Gentle Singularity, Zuck’s AI Power Play, Burning Of The Waymos
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the Big Technology Podcast, hosts Alex Kantrowitz and Ranjan Roy analyze major developments across the artificial intelligence landscape, including Sam Altman’s vision for digital superintelligence, Meta’s aggressive multi-billion-dollar talent acquisitions, Apple’s delayed AI rollout, and escalating public backlash against autonomous vehicles. They provide a critical look at the widening divide between Silicon Valley's grand technological promises and the economic, physical, and legal realities facing the tech industry.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 51.1% of the talking time here. How this is scored →
Alex as informed peer 6.3 Guest teaching 2.6 Guest disagreement 2.4 Alex pushing back 3.4
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 14:13 Ranjan rejects humanoid robotics hype
Ranjan openly breaks with mainstream tech consensus, firmly stating he does not understand the humanoid robot obsession and backing specialized warehouse machines instead.
Hardest push from Alex ▶ 55:21 Alex challenges the superficial vandalism framingAlex refuses Ranjan's claim that burning Waymos is just mindless vandalism, pressing him on why autonomous vehicles evoke deep societal anxieties around surveillance and automation.
Biggest teaching moment ▶ 25:08 Ranjan breaks down tool calling mechanicsWhen Alex asks for a definition, Ranjan clearly outlines the mechanics of tool calling via APIs and dynamic agent workflows that allow models to interact with the external world.
Alex holds their own ▶ 29:53 Alex synthesizes Meta's AI competitive strategyAlex demonstrates deep domain expertise by breaking down AI competition into compute, data, and algorithmic talent, citing Sergey Brin to defend Meta's acquisition strategy.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Sam Altman's 'Gentle Singularity' and Current AI Limitations | 7 | 2 | 2 | 4 | Alex counters Sam Altman's grand singularity claims by citing research showing a 30% AI coding adoption rate only yielded a 2.4% productivity increase. Ranjan agrees with the current application bottleneck but maintains optimism for enterprise integration over the next few years. | |
| Envisioning 2035, Brain-Computer Interfaces, and the Robotic Supply Chain | 5 | 1 | 1 | 2 | Alex and Ranjan engage in lighthearted banter over Altman's 2035 predictions, brain-computer interfaces, and runaway supply-chain robotics. Both express skepticism about invasive neural implants while acknowledging the steady spread of ambient wearables. | |
| Humanoid Robotics Versus Specialized Automation | 7 | 2 | 4 | 6 | Ranjan rejects the industry obsession with humanoid robotics in favor of specialized warehouse automation. Alex forcefully pushes back, arguing that foundation models with rich physical world models require flexible, human-like form factors to navigate environments built for people. | |
| Rise of the 'Idea Guy' and the Rebrand to Superintelligence | 5 | 3 | 3 | 3 | Alex notes the rising prominence of non-technical idea founders using AI coding tools, before questioning why the industry suddenly adopted superintelligence branding over AGI. Ranjan explains how Ilya Sutskever's branding allowed AI labs to pivot away from unmet AGI milestones. | |
| Model Advancements and the Critical Role of Tool Calling | 6 | 5 | 1 | 2 | Alex references reporting on OpenAI's O3-Pro model to emphasize how environment awareness matters for enterprise utility. He prompts Ranjan to explain tool calling, leading Ranjan to detail API executions and autonomous agent workflows. | |
| Meta’s Scale AI Deal and Zuckerberg’s AI Power Play | 8 | 2 | 2 | 4 | Alex presents a detailed strategic justification for Meta's $14.8 billion Scale AI deal, outlining compute, proprietary data, and elite algorithmic talent while citing Sergey Brin's previous appearance. Ranjan concedes that falling behind posed a larger existential risk to Meta's valuation. | |
| Meta AI Discover Feed Privacy Blunder and Data Labeling Future | 6 | 4 | 2 | 2 | Ranjan details Meta AI's Discover feed blunder, where private user queries and voice notes were broadcast publicly. In response to Alex's inquiry about Scale AI's motivation to sell, Ranjan outlines how synthetic training data threatens traditional human data labeling services. | |
| Apple’s Post-WWDC Strategy and Executive PR Defense | 6 | 3 | 3 | 3 | Alex and Ranjan critique Apple's post-WWDC media appearances, specifically Joanna Stern's interview with Craig Federighi. Both agree that Apple's dismissive messaging regarding AI delays and lack of self-awareness harmed its credibility. | |
| Generative AI's Impact on the Web and Copyright Battles | 7 | 3 | 2 | 2 | Alex shares analytics showing organic publisher search traffic dropped 55% over three years due to Google AI Overviews. Ranjan discusses Disney and Universal's copyright litigation against Midjourney, highlighting the visual replication of protected IP. | |
| Burning Waymos and Public Backlash Against Physical AI | 6 | 1 | 4 | 6 | Ranjan initially dismisses the burning of Waymo vehicles in Los Angeles as simple social media clout chasing. Alex pushes back vigorously, demonstrating that public unease with corporate surveillance and physical automation drives the intense public reaction. |
Statements from this episode (22)
Kantrowitz: Study Shows 30% AI Code Adoption Increased Productivity Only 2.4%
“There is a new paper that just came out there where they looked at a company going from zero to 30% of its code written by AI and a key measure of productivity only went up by 2.4%.”
Kantrowitz: Society Will Have Humanoid Robots Around by 2035
“I do anticipate that we'll have humanoid robots around, but the question is how good can the industry get them, and how safe can the industry get them?”
Roy: Specialized robots make more sense than humanoid form factors
“Like to me, applying the human form factor to robotics rather than actually having specialized robots that actually solve specific problems and are built, because again, right now you go to any automated warehouse, it's not humanoid robots moving around. It's …”
Kantrowitz: Humanoid robotics is optimal form factor for world-model AI
“When you build AI with a world model that understands physics, objects, how things work together, then you want to go humanoid robot, or maybe, you know, souped up robot that takes a humanoid form because all of a sudden you can be functional. Like the idea th…”
Roy: Founders have no excuse to pitch decks without working prototypes
“There's no excuse for that right now. Like anyone can build at least basic things right now and actually Many people, you do not have to be, have a full technical team to build a functional product. And that means that anyone with an idea should be able to act…”
Roy: Tech industry rebranded to superintelligence to dodge unmet AGI promises
“Clearly the messaging worked. Everyone started saying it. It absolved people from having to achieve AGI or when everyone is saying AGI is already here, yet life doesn't feel significantly different.”
Kantrowitz: Ilya Sutskever raised billions without releasing a product
“Well, Ilya obviously has raised billions without releasing a product.”
Kantrowitz: OpenAI o3-pro will launch first for $200/month subscribers
“OpenAI is releasing O three pro, which is a better version of the model. It's going to be available to those initially to those paying 200 dollars a month to open AI”
Roy: Tool calling will be the next decisive battleground in AI
“The next great battle in AI is tool calling. That's where we're going to see the maximum amount, like actually bringing these models and agent and agentic AI. That's all that matters is the ability for an action to understand its concept context and then take …”
Kantrowitz: Meta agrees to $14.8B 49% stake in Scale AI
“Meta has agreed to take a 49%
stake in the data labeling firm Scale AI for 14.8 billion.
Meta will send the cash to Scale's existing shareholders and place the startup CEO Alexander Wang, former Big Technology podcast guest in a top position inside Meta.
Meta…”
Kantrowitz: Zuckerberg's Scale AI deal is the exact right move
“So, to me, this is Zuckerberg clearly seeing an issue with his company and making, I would say, the strategic exact right move to fix it, unlike another company that we've seen within flagging
AI product that is, it seems like it's still in denial about what i…”
Kantrowitz: Meta is offering up to $100M to recruit top engineers
“There are reports that he's offered eight and nine figure cash amounts
Two top talent engineers to come over to Meta.
That's like in the tens, and maybe even up to a hundred million dollars to a person, not a company, a person to come over.”
Roy: Meta AI Discover tab publishes full user chats and voice recordings
“If you go to the Discover tab, it posts entireties of chats from people that probably don't realize that it's being posted, and it's like, as a social network, even crazier, it posts people's voices prompting Meta AI, so it has like an audio clip, audio messag…”
Roy: Manual data labeling by Scale AI will no longer be relevant
“What they did, their job, again, like actually having people manually tag, like large networks of people manually tagging data to make it more ingestible for a large language model or for a training, It's not going to be as relevant anymore. You can even now h…”
Kantrowitz: Siri's 15-year record shows Apple's AI is far behind competitors
“I think that's actually a fine strategy, but only if your AI works really well. And well, the state of Siri the actual ship stuff over the past 15 or so years suggests that it doesn't. They have to get their AI house in order and catch up”
Kantrowitz: Huffington Post and Washington Post Lost Roughly Half Organic Search Traffic
“Traffic from organic search to Huffington Post's desktop and mobile websites just fell by over half in the past three years. Nearly by that much at the Washington Post.”
Kantrowitz: Organic Search Web Traffic Fell 55% From 2022 to 2025
“Organic search traffic to websites declined by 55% between April, twenty-twenty-two and April, twenty-twenty-five, according to data from the company SimilarWeb.”
Roy: Search Engine Optimization for Display Ad Monetization Is Completely Dead
“Trying to game Google to get traffic, to show display ads, to make money is, that is beyond hospice. That is dead.”
Roy: AI Copyright Guidance or Resolution Will Arrive by Year-End 2025
“I, one of my predictions has been, like, we're gonna start to get some guidance or resolution, I think, by the end of this year, in terms of, like, how copyright will play out, and we need it.”
Roy: Midjourney Was Trained on and Directly Returns Copyrighted Content
“Midjourney, Clearly trained on copyrighted info and not an and end returns that info. That's a problem.”