Meta’s AI Glasses Controversy, Apple’s New Laptop, and More

Hey folks,

It’s Monday, so let’s quickly catch up on the some of the biggest and most interesting AI updates from the past few days.

Meta AI Glasses Footage Reviewed by Contractors

Meta’s AI-powered smart glasses are facing scrutiny after reports that videos captured by the devices were reviewed by contractors in Kenya. The footage was reportedly used by human annotators to help train AI systems.

According to the investigation, some videos contained highly sensitive situations, including private moments inside homes. While Meta reportedly blurs faces before sending footage to reviewers, workers claimed the system does not always work properly, occasionally exposing identifiable information.

The report highlights growing concerns about privacy as wearable AI devices with cameras become more common.

Claude Discovers More Firefox Bugs Than Human Teams

Researchers at Mozilla found that the AI model Anthropic Claude Opus 4.6 identified 22 vulnerabilities in Firefox within two weeks, and around 100 bugs overall.

Fourteen of these vulnerabilities were classified as high severity. This shows how AI can accelerate the process of identifying software security issues.

However, the model was less effective at exploiting those vulnerabilities, successfully turning only two into working exploits. Researchers say existing safeguards make real-world exploitation more difficult.

Apple MacBook Neo Benchmark Reveals Performance

Benchmark results for Apple’s upcoming MacBook Neo have appeared online, offering an early look at the device’s performance. The laptop runs on the same A18 Pro chip used in the iPhone 16 Pro.

According to Geekbench scores, the MacBook Neo achieved 3461 in single-core and 8668 in multi-core performance. These numbers are very close to the iPhone 16 Pro’s results, which is expected since both devices use the same CPU architecture.

The laptop includes slightly fewer GPU cores than the iPhone version of the chip, which explains the small difference in graphics scores.

Scrape Websites. Get Clean AI-Ready Data

Geekflare API lets you scrape any website and instantly convert it into clean, structured data for your AI workflows. Just enter a URL, and it returns the content in Markdown, JSON, or HTML. You can also capture screenshots and extract LLM-ready data without worrying about blocking, CAPTCHAs, or messy page code. It’s a simple way to collect reliable data for RAG, AI agents, and automation.

Turn ANY Website into LLM Data in seconds

Cheers,

Keval, Editor