Zuckerberg's AI manifesto draws skepticism from TechCrunch writers

Meta CEO Mark Zuckerberg published a 6,500-word essay this week titled "The Future is for Everyone," promising a future where, in his words, "everyone will have an exceptionally capable personal agent that understands you, your goals, and everything you care about." On TechCrunch's Equity podcast, writers Anthony Ha, Kirsten Korosec, and Rebecca Bellan discussed why the vision is meeting resistance, including a column by their colleague Russell Brandom, TechCrunch's AI editor, who called the manifesto "exactly why people don't like AI."

Bellan offered a cynical read: Meta is not winning in the closed frontier-model race, and is not clearly winning in the open-model race either, so it is trying to win on personal empowerment instead. The plan is for people to run Meta's own AI on their own devices. Zuckerberg's letter describes an assistant called Glimmer for managing schedules, drafting messages, and organizing files, always on and working with or without an internet connection. A separate, more capable model, Muse Spark, stays under tighter company control and gives Meta a revenue path from people or companies who want to scale up compute for larger projects. Bellan noted that last year's similar letter pointed toward wearables and glasses as the hardware, while this year it is unclear what the final hardware form factor will be, left up to the individual to choose.

Ha added context: Meta made large AI investments, including reports last year of steep pay offers for AI scientists, yet the company rarely comes up when people list the leading frontier labs, and it is not the company people think of for a personal AI assistant, despite users already interacting with Meta AI inside Instagram and Facebook. Bellan said that should be taken with skepticism: Zuckerberg has a history of promising his products will empower people, and on the social media side that promise turned into, in her words, "ragebaiting and advertisements, and not connection" rather than the outlet for friendship he had described. On the consumer AI side, she said, what people actually see from Meta AI is "a lot of creepy-ass chatbots." Her conclusion was that the negative reaction is largely about the messenger.

Korosec said the manifesto "didn't strike the right note for a lot of people," calling it Pollyannish for casting AI as an unqualified tool for humanity without acknowledging the costs that come with it, which she said explains the scale of the pushback. Bellan described Zuckerberg as positioning himself as an anti-Dario, in contrast to the frontier labs' talk of pacing development and pulling back on safety grounds amid recent cybersecurity incidents. She framed his stance as refusing to slow down, saying Meta cannot afford to give ground to China and that slowing down would only hurt the individuals AI could empower. When Bellan tried to download Muse Glimmer to her MacBook, she could not: it requires specific hardware she could not name, which she said makes it "not actually that accessible to the average person." Korosec summed it up: "It's not for everyone. Not yet."

Ha pointed to one line in the essay he found fair: Zuckerberg's challenge, "If you are an AI doomer and you believe that that's the future that AI is going to bring, why are you building this stuff?" His own takeaway was less about AI's promise than about people who make dire predictions and keep building anyway. He also raised the messenger problem directly, saying public suspicion of Zuckerberg and Meta could grow as a new Facebook movie approaches release, and tied the reaction to a broader pattern of viral pile-ons whenever AI executives predict the future, citing Sam Altman's idea of ChatGPT-generated podcasts about a listener's own kids, which was met with "why don't you just talk to your kid?" Ha said roughly half of Zuckerberg's manifesto is abstract talk of unleashed creativity that will not move a skeptic, while the more concrete parts, about personal coaches and assistants, have some appeal mixed with parts that do not land.

Key facts

Mark Zuckerberg published a 6,500-word essay this week titled "The Future is for Everyone," promising a personal AI agent for everyone.
TechCrunch's AI editor Russell Brandom wrote that the manifesto is "exactly why people don't like AI," a view aired on the outlet's Equity podcast by Anthony Ha, Kirsten Korosec, and Rebecca Bellan.
Meta's pitch centers on Glimmer, an always-on personal assistant, alongside Muse Spark, a more capable model Meta keeps under tighter control and monetizes for larger-scale compute use.
Rebecca Bellan said she could not download Muse Glimmer to her MacBook because it needs specific, unnamed hardware, calling it "not actually that accessible to the average person."
Bellan traced the skepticism partly to Zuckerberg's own record: his past promise of social connection instead produced, in her words, "ragebaiting and advertisements, and not connection."

Why it matters

The essay is Zuckerberg's second straight annual manifesto staking out a lane for Meta in AI: not competing head-on for the closed frontier or clearly winning the open-model race, but betting on personal, on-device AI agents as its point of difference. Bellan frames it as Zuckerberg positioning himself as an anti-Dario, refusing the caution around pacing and safety that other frontier labs have voiced, at a moment when trust in AI companies' promises is already being publicly tested.

Who it affects

The story is aimed at anyone weighing Meta's AI pitch against rivals: consumers deciding whether to use Meta AI, Instagram, or Facebook's built-in assistants, and readers tracking how Meta positions itself against labs like Anthropic and OpenAI. It also reflects on TechCrunch itself, since the discussion draws on its own AI editor's column alongside the podcast hosts' commentary.

How to use it

Per the essay as described on the podcast, Glimmer is meant to run always-on across personal devices, with or without an internet connection, handling schedules, message drafts, and file organization. Muse Spark is the more capable, more tightly controlled model, aimed at people or companies scaling up larger compute-heavy projects. Muse Glimmer, at least as Bellan tried it, cannot be installed on a MacBook and needs specific hardware she could not identify; no price or release details for either model are given in the source.

How solid is it

This is a discussion and opinion piece, not new reporting: three TechCrunch writers giving their read on Zuckerberg's essay, building on a column by a fourth colleague. Claims like "Meta doesn't tend to come up" among frontier labs or that Meta is "not winning" in open or closed models are the panelists' own characterizations, not independently sourced data, and no user numbers, revenue figures, or market-share figures are cited for any Meta AI product.

Risks and caveats

The source leaves the relationship between "Glimmer" and "Muse Glimmer" unclear, using both terms without saying whether they are the same product. The "recent cybersecurity incidents" Bellan alludes to are not identified, and neither Zuckerberg's essay nor last year's earlier letter is given an exact publication date, only "this week" and "last year." Readers should treat the panel's assessments as opinion and industry commentary rather than verified reporting on Meta's AI roadmap.

“exactly why people don't like AI.”

— Russell Brandom, TechCrunch AI editor