Viral AI News (Sep 23): Super Intelligence & Meta's Muse

Today's viral AI news is a naming fight, a download counter, and one genuinely alarming security story. The White House says the technology will be called something else from now on, Meta's new assistant is winning on charts while earning pennies per install, and a hacking campaign that used AI agents reached 100 companies in days. Yesterday's viral digest was about clones and product drops; today the spectacle is institutional.

Trump's 'super intelligence' rebrand is the meme of the day

Super intelligence is the new name US federal documents will use for artificial intelligence, per an announcement President Trump made at the United Nations. The Hill and the Washington Post both carried the remark. According to The New Republic, Trump credited a social media poll for the wording. Source: The Hill Source: The Washington Post Source: The New Republic

The BBC is already asking whether the rebrand will catch on, which is the honest framing. Nothing about model capability changed, no new system was unveiled, and no technical definition was published. A federal vocabulary change, announced from a UN podium, attributed to a poll.

If it sounds unserious, note the other half of the story: Reuters reported that Trump also said the Justice Department could rein in AI companies if needed. Source: Reuters

Meta's Muse is winning charts and earning 2.5 cents per download

Meta Muse is a consumer AI assistant app from Meta, priced at $20 a month or $100 a year, and it hit the top of the US App Store charts within three days of launch, per Bloomberg. Source: Bloomberg

Then the revenue math landed. Business Model Analyst estimated Muse is generating roughly 2.5 cents per download, and Business Insider noted it reached number one with fewer downloads than ChatGPT managed at launch. App-store ranking rewards velocity more than scale.

The more consequential viral moment came from Amazon, which moved to block Meta's Muse agent from shopping on its platform. Source: Forbes An agent that can buy things is a competitor to the store it is buying from.

An AI-fueled hack hit 100 companies in days

Forbes reported that a massive AI-fueled hacking campaign reached 100 companies in days, with credit card data among what was taken. Source: Forbes

The surrounding cases are real: Reuters reported Anthropic disclosed a fourth AI hacking incident its earlier review had missed, TechCrunch reported researchers used Anthropic's Claude to hack into OpenAI, and the Washington Post reported that the hackers who broke into OpenAI warn the industry has a security problem. Source: Reuters Source: TechCrunch

The scepticism is also fair: counts and cost figures from the 100-company campaign are preliminary, so treat them as unconfirmed. The confirmed pattern is that agentic tooling compresses reconnaissance and exploitation that used to take weeks into days.

Anthropic versus Xiaomi: the open-weights accusation

Several outlets reported this week that Anthropic has raised concerns that Xiaomi's new open model may be distilled from Claude outputs, a claim that arrived the same week the Xiaomi model went to the top of the open-weights charts. We are labelling this one community-sourced and unconfirmed: no formal finding has been published by either company, and Xiaomi has not responded publicly. Source: The Next Web

MiMo-V2.6-Pro is an open-weights model from Xiaomi, released under the MIT license, meaning anyone can use it commercially for free. It has roughly one trillion parameters, about 42 billion active per token, and a one-million-token context window. VentureBeat called it the top open-weights release at launch, and it scores 46 on the Artificial Analysis index against Opus 5.5's 58. Source: VentureBeat

If the distillation claim is ever substantiated, the open-weights debate stops being about performance and becomes about provenance, which is a much harder problem for everyone shipping free models.

Quick Comparison

Frequently Asked Questions

What is the viral AI news today?

The viral AI news today is Trump's announcement that US documents will call AI "super intelligence", a remark made at the United Nations and reportedly based on a social media poll. Meta's Muse assistant winning the App Store while earning about 2.5 cents per download is the second most shared story.

Why is "super intelligence" trending?

"Super intelligence" is trending because it is now the term US federal documents will use for artificial intelligence, per remarks Trump made at the UN General Assembly. The BBC and multiple outlets have questioned whether the rebrand will stick, since no technical definition or new model accompanied the change.

Is Meta's Muse actually a hit?

Meta Muse reached number one on the US App Store within three days of launch while generating an estimated 2.5 cents per download, so it is a distribution hit more than a revenue hit. Amazon has blocked the Muse agent from shopping on its platform.

Bottom Line

Two of today's three stories are about words and charts, and one is about a real security failure. The rebrand changes vocabulary, while the hack changes the threat model. Only the last one will still matter in December. More trending AI coverage sits in the viral archive.

Related reading

Deeper context from the archive on the companies and technologies in this story:

• Viral AI News (Sep 29): Jev Clones, Tracker Leak & Meta's Muse
• Viral AI News (Sep 12): Coxon, No-AI HN & Meta Muse
• General AI News (Sep 25): Oracle, Copilot Super App & Akamai
• General AI News (Sep 17): OpenAI Disclosures, Meta Arke & E-Waste

Building an AI agent, automation, or backend workflow? AsaanKaroobar builds custom AI solutions, apps, and digital experiences. Get in touch →