At Meta Connect this week, the company showcased its latest Ray-Ban Meta smart glasses lineup, introducing new display features, AI integrations, and even sport-oriented Oakley Meta Vanguard frames. But during the live keynote, two demos didn’t go exactly as planned (It happens to the best of us 🥲).
Instead of brushing it off as Wi-Fi trouble, Meta’s CTO Andrew Bosworth later gave a detailed explanation on Instagram of what really happened and why these kinds of edge cases are part of testing ambitious new technology.
The Cooking Demo
During the showcase, chef and creator Jack Mancuso put the glasses’ Live AI feature to the test. He asked the glasses for the first step in his recipe, but the system didn’t respond as expected. The AI skipped ahead, creating a confusing moment on stage.
According to Bosworth, this wasn’t a failure of the glasses themselves but rather the result of the demo environment:
“When Jack said, ‘Hey Meta, start Live AI,’ every single pair of Ray-Ban Meta glasses in the building tried to respond at once. They were all on the same access points, and because we routed that traffic to a demo server, it got overloaded. Essentially, we DDoS’d ourselves.”
This was a rare and very specific situation, only possible in a room full of identical devices listening for the same wake phrase. Bosworth noted that the team has already updated how demo traffic is handled to prevent this from happening again.
The Video Call
Later in the keynote, Mark Zuckerberg attempted to answer a WhatsApp video call on the new Ray-Ban Display glasses. The prompt to accept the call didn’t appear, leaving the demo incomplete.
Bosworth explained that this was caused by a race condition between the glasses’ sleep state and the incoming call notification:
“The glasses went to sleep at the exact same moment the call came in. That created a rare bug where waking the display didn’t properly trigger the notification. We’d never seen that happen before, but now that we have, we’ve fixed it.”
Why These Moments Matter
Rather than being setbacks, Bosworth positioned these glitches as valuable discoveries. Every new piece of technology has to go through the process of hitting unexpected edge cases and often, it’s live demos that uncover them.
“This is exactly why we keep testing and pushing these devices in real-world scenarios. Every time we find a bug, we fix it and make the product stronger for everyone who will use it daily.”
Building the Future of Wearable AI
Meta’s approach highlights an important truth for anyone building next-generation products: innovation isn’t about flawless stage moments, it’s about finding and fixing the issues that only appear under scale and pressure.
The Ray-Ban Meta glasses continue to push forward the promise of wearable AI, and the transparency from Meta’s leadership offers a reminder that progress in this space is built step by step, bug by bug, improvement by improvement.
At Botpool, we believe the same principle applies to the future of work. Testing, improving, and adapting is how freelancers and the platforms we use get stronger over time.
👉 If you’re a freelancer looking to showcase your skills, or a business that wants to hire top AI-powered talent, join us at Botpool and be part of the future of freelancing.
