AI · Oct 14, 2026
OpenAI, Google pushing new AI models easily tricked into making nudes, researchers sayCan Muse overcome Meta’s trust issues?
Analysis of reporting by TechCrunch. How we work

Equity’s take on Muse is sharper than the Connect keynote: party tricks land; handing Meta your cards and inbox is the real wall.
Why it matters
Meta’s Muse agent stole oxygen at Connect while OpenAI and Anthropic chased models and enterprise money. Zuckerberg made clear AI features are going everywhere Meta already lives.
On TechCrunch’s Equity podcast, Anthony Ha, Kirsten Korosec, and Sean O’Kane argued that Meta may be playing to its actual strength—consumer habit—while rivals chase IPO-scale enterprise revenue. That bet only works if people trust Muse with the boring, sensitive stuff: credit cards, Gmail, subscriptions, the accounts that make an agent useful day after day.
Sean tried Muse. It found him unclaimed funds on day one. Cool. Also, in his words, more of a party trick than a habit. The sustainable pitch—Rocket Money–style subscription hunting and charge spotting—runs straight into Meta’s ads business.
From the desk
We’re glad someone built a consumer agent that doesn’t apologize for being consumer. Coding copilots and enterprise seats are real. So is an agent that lives next to Instagram and WhatsApp. Meta knows how to embed in daily life; pretending otherwise is lazy critique.
Trust is still the product. Sean’s detail that stuck with us: Muse didn’t immediately glue him to Threads, Instagram, and Facebook. It treated him like a stranger at first, which made him more willing to poke around. Then it started pulling for more context. That’s the arc. Novelty first, data gravity second.
Useful personal AI should cancel junk subscriptions and surface money you forgot about. We’re for that. We’re not for pretending Meta’s incentive stack is Apple’s. Sean’s comparison to a working Siri on a new iPhone is the right one: same class of intimate permissions, different business model. Meta will argue better ads from deeper knowledge. Wake us when users experience that as a feature, not a fever dream.
If Muse stays a cute demo plus one-off finds, Connect was a show. If it becomes the default place people wire finances and email, Meta’s trust deficit becomes everyone else’s risk surface. I’m watching whether the agent earns repeat use without demanding the keys to the kingdom—or whether “adults-only” Tamagotchi charm is just the on-ramp.
Context
Anthony Ha, Equity podcast write-up on TechCrunch, September 27, 2026. Muse is framed as Meta’s ground-up, phone-native agent in the vein of the OpenClaw-style work Meta integrated; the company also showed a cute, Tamagotchi-style device it insists is adults-only.
Who feels it
- Consumers
- Party-trick wins are easy; ongoing value means sharing financial and inbox access with an ads company.
- Meta
- Consumer differentiation vs. OpenAI/Anthropic only pays off if Muse clears the trust wall on sensitive account linking.
- Apple / rivals
- A more capable on-device Siri raises the bar: same agent promise, different default trust story.
What to watch
- Whether Muse retention depends on linking cards, Gmail, and social graphs
- User reports of ad targeting shifts after deeper Muse permissions
- How Muse Charm and adults-only framing fare once the novelty week ends
Companies: OpenAI, Anthropic, Meta
More on OpenAI
AI · Oct 13, 2026
Google finally releases Gemini 4 ArgonAI · Sep 22, 2026
Anthropic launches Claude Opus 5.5 with stricter safeguards for cybersecurityAI · Sep 30, 2026
Here’s how tech leaders will self-police AI safety under Trump’s dealAI · Sep 29, 2026
AI researchers put out videos saying superintelligence is ‘exactly as dangerous as it sounds’