Aug 8, 2026 · 58 min
Listen from 7:14
Listen at 7:14
Google bets on world models as Apple and Microsoft face AI pressures
Demis Steps Down, Apple’s Memory Problem, Microsoft’s Clever Trick
The episode connects Google’s strategic choices with the financial and operational costs AI is imposing across the technology industry.
- 1Demis Hassabis’s shift toward scientific research and world models could advance AI while leaving Google exposed in current applications.
- 2Apple’s memory-cost pressures may force higher iPhone prices as services growth slows and its hardware economics face strain.
- 3Microsoft’s data-center accounting change sharpens questions about AI spending, cloud demand, and reliance on major AI labs.
Don't miss
The hosts debate whether Google has deliberately ceded the current AI race to pursue world models and the breakthroughs it believes AGI requires.
The brief
Demis Hassabis is stepping down as Google DeepMind CEO, prompting a debate over whether his greater focus on scientific research and world models helps AI but hurts Google’s application strategy.
With M. G. Siegler, Alex Kantrowitz examines Apple’s memory-cost pressures, slowing services growth, and the possibility that higher iPhone prices could protect the company’s margins.
The discussion turns to Microsoft’s decision to extend data-center depreciation from 15 to 25 years, a change that lowers reported costs as AI spending accelerates.
Alex and Siegler question whether cloud growth depends too heavily on OpenAI and Anthropic, raising concerns about circular financing and the durability of AI demand.
The central tension is Google’s apparent tradeoff: pursuing world models and long-term AGI breakthroughs while competitors gain ground in large language models, coding agents, and applications.
What was said on this episode
25 statements · 6 positive · 12 negative · 1 mixed · 6 neutral
Demis’s shift toward scientific and world-model research benefits AI development.
“I actually think it's good for AI that Demis will be focusing more on the scientific side, probably more on the world model side, and not deeply in LLMs.”
Listen at 4:13
Success by either LLM developers or Demis’s group would advance AI technology.
“if one of them succeeds, whether it's the Anthropic OpenAI group or it's the Demis group, ultimately the technology progresses.”
Listen at 4:39
Google is currently behind competitors in coding agents.
“Google is clearly behind on the coding agents.”
Listen at 5:00
Google has not made significant progress in coding-agent applications.
“Google has not made noise on that front.”
Listen at 5:11
Google’s leadership shakeup may or may not produce beneficial results.
“whether this effectively high-stakes shakeup pays off.”
Listen at 6:34
Apple will struggle as memory costs increase.
“Apple is going to struggle as these memory costs go up.”
Listen at 9:48
Apple’s device components have become substantially more expensive.
“the components are now super expensive that they need to use for all their devices.”
Listen at 11:49
Apple underestimated the importance of artificial intelligence.
“within the company, they underplayed the importance of AI.”
Listen at 13:01
Apple will raise iPhone prices.
“I do think that they're going to for sure raise prices.”
Listen at 21:06
John Ternus may modify Apple’s App Store model.
“I wouldn't be shocked if at some point one of John Ternus's first moves is to change the App Store model a bit.”
Listen at 24:07
Apple will continue resisting EU platform demands in the short term.
“I wouldn't expect them to do anything in the short term with the EU other than continuing to fight”
Listen at 26:06
Apple’s device subscription strategy is strategically sound.
“This, I think, is strategically very smart of Apple.”
Listen at 29:20
Apple Upgrade will be important for managing device price increases.
“I think it's gonna be vital again for all the stuff we were talking about with the price increases”
Listen at 30:32
Microsoft’s accounting change does not reduce total spending.
“they're spending the same amount overall.”
Listen at 39:27
Microsoft shifted some data-center spending from CapEx to OpEx accounting.
“it changes it from going under the CapEx line to moving it over to the OpEx line.”
Listen at 39:31
Microsoft’s accounting change is a strategically savvy move.
“I think it's a savvy move and I think it gives them a little bit more headway.”
Listen at 40:54
Microsoft’s 25-year accounting life applies to data-center facilities.
“25 years, from my understanding, is the actual facility itself.”
Listen at 42:02
Data-center server lifespans commonly range from three to seven years.
“they range, yeah, like you noted, from 3 years, some are 6 years, some are 7”
Listen at 42:30
A slowdown at Anthropic or OpenAI would affect major cloud providers.
“if anything were to happen to slow down Anthropic or OpenAI, that has a cascading effect that sort of happens across the industry”
Listen at 45:50
Anthropic and OpenAI will eventually operate their own clouds.
“they will eventually obviously just have their own clouds.”
Listen at 48:02
Google’s 3.5 Pro model is delayed by months.
“the reporting now suggests that it's months delayed.”
Listen at 51:26
Google’s AI organization is experiencing internal cultural conflict.
“there's a bit of a culture clash.”
Listen at 51:55
Google mishandled development of its 3.5 Pro model.
“they really messed up with 3.5 Pro.”
Listen at 53:08
Google is prioritizing world models over large-language-model development.
“they're willing to sort of take a backseat on LLM development in order to push forward with that.”
Listen at 55:30
Anthropic and OpenAI may reach recursive self-improvement through current LLM progress.
“they could find that where they are in LLMs right now is actually going to trip over into the recursive self-improvement stuff”
Listen at 56:32
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.