This is your laptop... on AI

Summary of This is your laptop... on AI

by The Verge

1h 46m•June 5, 2026
Listen to Episode

Overview of The Vergecast

This episode of The Vergecast is a wide-ranging conversation about how AI is reshaping work, personal computing, and the future of devices. The main thread is a debate over whether AI agents belong in everyday consumer life or are mainly useful in enterprise settings, with examples from Google’s Gemini Spark, Microsoft’s Build announcements, and NVIDIA’s vision for AI-powered PCs. The hosts also dig into Apple’s smart glasses ambitions, Brendan Carr’s misleading broadband claims, and a very on-brand Meta security failure caused by replacing humans with AI.

AI Agents: Powerful, Useful, and Kind of Terrifying

Gemini Spark and the “personalized assistant” promise

  • David Pierce describes testing Google’s Gemini Spark and being stunned by how well it could build a hyper-personalized travel itinerary using data scattered across Google services.
  • Spark pulled details from:
    • Gmail
    • Google Photos
    • YouTube
    • Search history
    • Calendar and other Google data
  • The result was impressively useful but also unsettling, because it showed how much Google knows about a user’s life without explicit permission in every case.

The central debate: convenience vs. surveillance

  • The hosts frame the tension as:
    • Pro case: AI finally turns surveillance data into something genuinely useful.
    • Bear case: It feels like your phone has been listening to you the whole time.
  • Their takeaway:
    • The product can be good and still make people uncomfortable.
    • For many users, especially in consumer settings, this may be a dealbreaker.

Why this works better at work than in personal life

  • They argue AI agents make the most sense in enterprise settings because:
    • Work already runs on structured data.
    • Employees are incentivized to adopt tools that increase efficiency.
    • Corporate workflows are easier for AI to manipulate.
  • In personal life, by contrast, people don’t want their daily routines to feel like workflow optimization.

The Future of Computers: Local AI vs. Cloud AI

Jensen Huang’s vision for the laptop as an always-on assistant

  • NVIDIA’s Jensen Huang described a future where your laptop is essentially an AI assistant you can text remotely.
  • His vision:
    • The laptop should be running all the time
    • It should be available like a household appliance
    • AI should live locally when possible to avoid cloud costs

Why the hosts are skeptical

  • They argue this is mostly a business/workflow fantasy:
    • The example is basically PowerPoint automation
    • It assumes a device sitting open and ready to serve you
    • It doesn’t solve the broader question of what consumers actually want
  • Key point:
    • More powerful computers are useful.
    • That does not necessarily mean a new form factor is needed.

The real question: where does computing live?

  • The episode returns repeatedly to the unresolved split between:
    • Local compute on the device
    • Cloud compute powered by centralized models
  • Their view:
    • The industry is still confused about this.
    • Many companies are trying to solve the same problem from opposite directions.

Apple, Vision Pro, and Smart Glasses

Apple may be pivoting away from the Vision Pro thesis

  • The hosts discuss reporting that Apple may be moving away from a heavier headset-first strategy and toward lighter smart glasses.
  • Their read:
    • Vision Pro was the best version of “VR as a product”
    • But it was never enough to make headsets the default computing interface

Why smart glasses matter more

  • Smart glasses are seen as a more compelling long-term product because:
    • They’re lighter
    • They fit into everyday life more naturally
    • They can potentially support useful AR overlays without full immersion
  • But there are major hurdles:
    • Real-time visual recognition
    • Low-latency compositing
    • Notification design
    • Privacy concerns

Apple’s advantage and problem

  • Apple may be best positioned to make smart glasses work because it has:
    • Access to the iPhone ecosystem
    • Tight hardware/software integration
    • Strong control over notifications and cameras
  • But the hosts note that consumers may eventually see Apple-branded glasses as invasive rather than stylish.

Brendan Carr and the Broadband Price Lie

Carr claims Americans are paying less for internet

  • The episode’s “Brendan Carr is a Dummy” segment takes aim at Carr’s claim that Trump-era FCC policy lowered broadband prices.
  • The hosts call the claim false and misleading.

What the policy actually was

  • The Biden FCC had blocked “bulk billing” arrangements where landlords force tenants into a single ISP deal.
  • The hosts explain that these arrangements reduce competition and keep prices high.

Why the study Carr cited is bogus

  • Carr cited a study from the Bulk Broadband Alliance, which is a telecom-backed trade group.
  • Their critique:
    • It’s a sponsored study from companies that benefit from monopoly-like arrangements
    • Carr is parroting industry talking points while pretending to regulate them
  • Bottom line:
    • Internet prices have not gone down meaningfully
    • The FCC under Carr is removing transparency and competition protections

Meta’s AI Trust-and-Safety Failure

Instagram accounts were hacked through AI support

  • The episode closes with a cautionary tale:
    • Hackers exploited Meta’s AI chatbot system to reset passwords on hacked Instagram accounts, including reportedly Barack Obama’s account.
  • The hosts point out:
    • Meta had cut too many trust-and-safety and support staff
    • The replacement AI assistant was easy to trick
    • The company’s response was slow and vague

The broader lesson

  • This is presented as a perfect example of the risks of replacing humans with AI before the system is ready.
  • Their takeaway:
    • AI without human oversight can become a security liability
    • Meta is still prioritizing automation and efficiency over safety

Other Notable Moments

Gaming hype and 007 First Light

  • In the “Hype Desk” segment, Ross talks about:
    • Summer Games Fest
    • The new James Bond game 007 First Light
    • How good the tutorial and cinematic structure feel
  • He also highlights a few major gaming trends:
    • Sony’s single-player cinematic push
    • Marvel’s Wolverine
    • A new God of War game where you play as Laufey instead of Kratos

Supernatural is returning

  • The hosts end on a positive note with news that Supernatural, Meta’s VR fitness app, is being spun out into an independent company.
  • They’re optimistic this could let it evolve beyond Meta’s messy VR ecosystem.

Key Takeaways

  • AI is now good enough to be genuinely useful, but that makes its privacy tradeoffs harder to ignore.
  • Consumer AI is still struggling to find a product form that people actually want in daily life.
  • Work and enterprise remain the clearest fit for agentic AI.
  • Apple, Google, Microsoft, and NVIDIA are all still trying to define where computing should live: on-device, in the cloud, or somewhere in between.
  • The episode’s recurring theme: the technology may be improving fast, but the social and product questions are far from settled.