Are AI Glasses Over? + Big Technology AI Summit Audience Questions
TL;DR
Snap's new Spectacles triggered a blunt reality check: Kantrowitz says the AI device of the future may be the improved iPhone and Siri, while Ranjan Roy still believes AR glasses will happen eventually, just not with today's heavy, awkward hardware.
The audience's best AR use cases were social and work-related, not sci-fi: one attendee wanted lightweight glasses for shared movie experiences on planes and for managing massive multitasking setups, which landed better than vague promises about ambient computing.
Autonomous agent benchmarks are rising fast, but real-world use still lags: Kantrowitz cites METR-style progress from about 30 minutes of autonomous coding in 2023 to roughly 18 hours now, while Roy says benchmarks still overstate what agents can reliably do in messy enterprise settings.
AI chatbots are building far richer user profiles than search ever could: both hosts agree ads are likely coming, and Roy contrasts 'thousands of three-to-five-word Google searches' with long therapy-like conversations that expose far more personal context.
The panel keeps returning to a gap between demos and actual utility: travel booking remains the default agent example, which Roy openly mocks, even as he admits tools like Perplexity Computer can already monitor hotel prices and email you when rates drop.
On big tech responsibility, neither host expects altruism to win over competition: Roy notes Anthropic's public benefit corporation status, but both hosts say the race is too fierce to count on companies solving society's hardest problems out of principle.
The Breakdown
Alex Kantrowitz opens by arguing the AI device might just be the iPhone, not your face, after Snap's new Spectacles landed with a thud and Evan Spiegel's demo became a punchline. The live audience then pushes the conversation into the real stakes: autonomous agents that may soon work for "18 hours," AI systems that know us better than search ever did, and whether big tech will use that power for ads, public health, or mostly competition.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.

Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.

Playbook
The Art of Tasteful Prompting
Learn how tasteful prompting helps you move beyond generic AI output by shaping context, style, and judgment from the start.