Zuckerberg's Hilarious Reaction to First Guinness Sip Goes Viral
Zuckerberg's First Guinness Sip Goes Viral

Mark Zuckerberg tried Guinness for the first time during an appearance on John Collison's Cheeky Pint podcast, and his reaction was far from impressed. The 42-year-old Facebook mogul took a small sip of the Irish stout, pulled a disgusted face, and quickly set the pint down, pushing it away.

Viral Clip of Zuckerberg's Reaction

In the video, Zuckerberg sits next to his chief of AI, Alexander Wang, who discusses Meta's newest AI model, Muse Wonder. Zuckerberg is quiet throughout the brief clip, but at one point, he samples the drink, looks up at the ceiling, and swills it around his mouth before setting it aside.

The clip was posted across social media, including X and Reddit, and has since gone viral. One person reposting the video said, "Mark Zuckerberg pretending to enjoy Guinness is my new favourite thing."

Meta's AI Announcements

Zuckerberg recently made headlines after announcing that Meta's new bot would be at the "frontier" of the AI race. He claimed Muse Wonder is as powerful as the latest technology from OpenAI, the brand behind ChatGPT, and Anthropic, the developer of Claude.

Meta will also roll out Muse Wonder Full, which the company describes as the "largest and most capable model we have trained." Zuckerberg said the bots would "allow much more sophisticated behaviors that are at the frontier of what is being done in the AI field."

Anthropic's Report on Houthi Weapons

Meta's competitor, Anthropic, recently admitted that its technology had been used to create missiles used by militants who attacked King Khalid International Airport in Riyadh, Saudi Arabia. The attack by Houthis killed more than a dozen people and wounded hundreds.

Anthropic published a report admitting that the Houthis had launched three weapons programmes using Claude AI. The technology provided homing guidance, allowing missiles to reroute themselves if they missed targets. The Houthis used Claude Code to help develop technology for guiding, controlling, and navigating the weapons. Anthropic blocked most of the AI requests, but not all of them.

The report said: "Threat actors based in northern Yemen [were] running three weapons development programmes: a guided rocket that used a commodity, phone-class flight computer with final phase homing guidance, a multi-stage ballistic missile with a stated range goal of more than 1,500 miles and a multi-variant missile, referred to as the R2000, that included a hypersonic glide vehicle variant."