Social Round-Up 8 July

X (formerly Twitter) has launched a completely redesigned video editor and recorder user interface for its iOS app. Aimed at establishing itself as a true competitor to short-form giants, the updated UI introduces multi-language captioning and an integrated green screen recording overlay. X’s Head of Product, Nikita Bier, emphasized that giving creators native tools to publish original content remains a top priority, with even more video updates slated for the coming weeks.

Meta’s brand-new Superintelligence Labs division has launched its first official AI image generation model, dubbed Muse Image. Replacing the older Llama lineup across Instagram, WhatsApp, and Meta AI, this “agentic” model actually searches the web and reasons through prompts before rendering. Most notably, users can now @ mention other Instagram accounts directly in an AI prompt. By tagging a username, the AI scrapes that user’s public photos to seamlessly build their real-life likeness into the generated image. The model is also rolling out 30 new AI effects to Instagram Stories in the US.

Google is introducing a massive analytics upgrade for content creators called platform properties within Search Console. This new property type allows publishers and creators to directly monitor how their native social media and video posts are performing across Google Search and Discover. Launching initially for YouTube, TikTok, Instagram, and X, the system tracks total clicks, impressions, and exact search queries. Best of all, it works entirely for standalone social accounts, meaning creators do not even need to own or verify a traditional website to access the data.

In the era of automated “AI slop,” Reddit has radically upgraded its automated defenses to protect platform authenticity. By deploying advanced Large Language Models sitewide, Reddit is now blocking a staggering 23 million spam views per day and revoking nearly 2 million fake automated votes daily. These new safeguards have successfully slashed user spam exposure by 20%. Furthermore, the system has weaponized AI to address sitewide hate speech and harassment, dropping the average enforcement response time on harmful text down to under five seconds.

Meta is staring down a potentially catastrophic legal crisis as four U.S. states: California, Colorado, Kentucky, and New Jersey, seek a combined $1.4 trillion in penalties. The lawsuits allege that Meta intentionally engineered its social platforms to be highly addictive to teenagers while actively misleading the public on youth safety. The unfathomable penalty sum, which is calculated by multiplying state fine structures by the number of affected young users, is roughly equal to Meta’s entire market capitalization. Meta is currently fighting the claims, arguing there is zero proof they misled consumers since “social media addiction” remains an unestablished psychiatric condition.