“When the chef said, 'Hey Meta, start Live AI,' it started every single Ray-Ban Meta's Live AI in the building," he said. "And there was a lot of people in that building."

You don’t think of that before the demo?? What’s fucking wrong with you?

DDOSing yourselves explains slow response or no response, but the chef supposedly DDOSing the dev server doesn't explain away the AI arguing with him on what part of the cooking process he's in. That sort of hallucination is baked in and frankly unfixable.

I loved this episode of Silicon Valley. So glad they brought the series back.

"When the chef said, 'Hey Meta, start Live AI,' it started every single Ray-Ban Meta's Live AI in the building," he said. "And there was a lot of people in that building."

"That obviously didn't happen in rehearsal," Bosworth said. "We didn't have as many things."

It wasn't the only major blunder Meta encountered during its keynote.

Since Meta routed all traffic to its "dev server," including from all of the headsets in the building, "we DDoS'd ourselves, basically," Bosworth admitted, referring to a common cyberattack strategy known as a "denial-of-service attack" that attempts to bring a network down by overwhelming it with phony internet traffic.

The article missed the best part.

When the AI failed to help with the recipe, they blamed the wifi.

And Zuckerberg backed it up! He also blamed the wifi.

Like dude the AI is reaching skynet BC it responded to you, the wifi is clearly working.

The sad thing is, having a quality camera and some smart features integrated into your glasses could be useful. Think about a hands free video call to your mother, helping you with a recipe. Or making a cool sport video without a gopro. Taking a quick photo or checking map directions during your vacation, without having to constantly hold your phone in your hand.

So the possibility is there, but insted we get it from the biggest privacy nigthmare company on the planet, and have it packed full of useless AI shit. No way I would ever touch this

Thanks for posting, this is really cheerful!

I'm almost positive that what actually happened is that the response the chef cut off included an instruction to mix the ingredients, but he cut it off before it could say that. I bet the LLM doesn't get any feedback about whether or not the voice module got through the whole response, so it couldn't account for him not having heard it. He probably could have gotten back on track by saying something like "I haven't mixed the ingredients yet, let's start over."

The ingredients list it was going through before he cut it off probably mentioned the pear too, which would have been a good time for him to say "We don't have a pear, is there a sauce we can make using only the ingredients visible here?" but he was afraid of going off script.

Those glasses are so fucking dorky.

hmm, does this mean that all those expensive AI scientists will be laid off now? 🫤

I’m going to be sorry for those guys for trusting this company.

C’mon guys, put these things on your faces so I can stop giving Apple and Google money.