I Think AI Wrote It
I read Mark Zuckerberg's latest essay so you don't have to. Here are my thoughts.
On August 10th, Mark Zuckerberg published an essay on Meta’s website titled The Future is for Everyone, The Path to a Positive AI Future. If I were asked to describe the essay in three words, they would be: grandiose, delusional and confusing.
The essay has a bold start, claiming "In the next few years, people will be able to use superintelligence beyond human capacity to create and discover extraordinary new things, build new businesses, express new ideas, learn new concepts, and advance our health and quality of life." - If you're wondering what superintelligence is, so am I. We’ve all seen this term gaining popularity in the last few years and its popularity has even incited petitions to ban its development. If you recall, in September 2025 the Future of Life Institute published a Statement on Superintelligence which has 138,723 signatures. The signatories include some high profile names across the tech, research, media and religious sectors.
What’s perplexing to me is that no one has clearly defined what superintelligence is. There seems to be an implicit understanding this capability would posses “superhuman” intelligence and competence, without any clear benchmark for where that threshold lies. From my perspective, this notion of superintelligence carries the glaring omission of the limitations of software and the brittleness of complex systems.
It's never clarified in Zuckerberg’s essay what superintelligence actually entails, nor is there currently any credible scientific evidence of superintelligence. While the term is highly referenced in public discourse, there is no credible evidence of what superintelligence actually is and whether it could be feasibly achieved. The essay does however claim that AI will soon be able to "discover knew knowledge - ranging from discovering new drugs to cure a family member's disease to finding new ways to improve your business."
The essay then goes down a defensive path, countering public criticisms of AI and the AI industry. Specifically, he rejects arguments around power concentration, the loss of jobs and humanity's relevance.
He talks about the opportunities of putting “power in people's hands”, making reference to the Wright brothers and Steve Jobs. I imagine he sees himself as similarly influential. There is a lot of emphases on the distribution of power through individual access to AI, implying power sits at the user level. The arguments on power imbalance do not talk simply to access and use of AI. Rather, they point to broader power imbalances around concentrated markets, influence on regulatory interventions, climate impacts and the obfuscation of responsibility. The idea of open access use of superintelligent AI equating to an equal distribution of power is deliberately evasive and dismissive of the actual problem at hand.
The essay then goes on to list the ways in which Meta is "building personal superintelligence to improve all of our lives." I've listed these, and my thoughts, below.
Everyone will have an exceptionally capable personal agent that understands you, your goals, and everything you care about. - This item has a lot of notes around privacy and trust, a nod to public concerns on information privacy. He even provides personal examples of how he and his daughter use their personal agents. He also alleges we will be able to access our agents using Meta glasses.
Everyone will have incredible tools for creation to express your ideas. - He claims we will all soon have invention superpowers and again uses his 8 year old daughter as an example.
Everyone will have powerful tools to create new businesses. - He claims the economy is going to become more entrepreneurial and will result in economic growth.
Everyone will have a personalized tutor and coach with a PhD in every subject. - As someone who undertook the grueling endeavour of a PhD, I am personally offended by this claim. Putting my personal grievances aside, this claim overlooks the reality of the inaccuracy of these systems and model collapse from the inevitable feedback loop of AI content outpacing human generated content. This nature article is a useful read on this topic.
Everyone will benefit from scientific advances and be able to contribute to scientific progress. - He expects all diseases will be preventable and curable this century. He also claims "Since there is a long tail of health conditions and everyone's biology is different, people's direct involvement will help unlock personalized therapies and expedite trials to match the pace of invention and enable industry to move beyond focusing disproportionately on common diseases." - This is another note around privacy and trust. Users are becoming more aware of the abuse of access to personal information and this kind of narrative is an attempt to temper that disquiet.
Everyone will have free or affordable access to these tools. - Apparently we're all going to have free access to a version of superintelligence, but if you want access to greater compute capacity, you'll have to pay for it through "a dynamic auction mechanism that will guarantee that everyone gets the lowest price possible for the intelligence and compute they're using while also ensuring the capacity is used for whatever people collectively find most valuable." So basically auctioning off access to greater compute power to the highest bidder. This claim is made in tandem with claims of equal power distribution, a glaring contradiction.
The essay then goes on to directly address the risks that come with new technologies, framing all of the risks as external to industry decision making. Job losses are framed as job displacement, concerns around the impacts of data centres to local communities are framed as a challenge of ensuring communities benefit from data centres, and cybersecurity concerns are framed as a misuse of AI. He mentions the need to avoid “government tyranny and surveillance” and also the need to ensure US and democratic countries lead. Finally, we are graced with the most illusive risk of them all, "making sure humanity maintains control over superintelligence so it serves rather than endangers us." Anyone who reads my work often knows I despise this language and framing. In the latest Responsible Bytes Podcast episode I actually spoke about this very issue with Eryk Salvaggio. Eryk does a great job of explaining not only the ludicrousness of this claim, but also the strategy behind it in terms of shaping public perceptions.
Each of these risks is expanded upon towards the end of the essay. I won't go over all of them, but I do want to touch on a few, starting with the alleged risk of Protecting Freedom and Preventing Government Tyranny. He claims totalitarianism is a constant tension of democracy, a claim I wildly disagree with, mainly because the two states are contradictory. Democracy, at least theoretically, does not lean towards totalitarianism so the claim of this possibility being a constant tension is categorically misplaced. There are, again, notes on personal privacy here and access to personal information in the context of governments. He's walking a thin line between reassurance and keeping the language verbose enough for multiple interpretations. A preparatory scape goat for the inevitable day in which he’ll have to say, "that’s not what I meant".
The illusive China threat is alluded to under the risk of Ensuring American Leadership. Emphasising the need to bulldoze past any opposition to infrastructure development which may slow down the growth of data centres. He claims, “Any policy that slows American model releases -- even by a month -- could add significant risk to American leadership while letting foreign models race ahead.” This is a very pointed remark on policy and regulatory initiatives, maintaining the tired and unjustified stance of “bureaucratic red tape” stifling innovation.
The final risk I’ll comment on is Maintaining Control of Superintelligence. This section makes reference to the “singular superintelligence we fear.” What’s the solution to this existential threat you ask? Well, it just so happens that Meta can jump in and save the day by “directing” these systems. So, while Meta believes in distributed power for all, they are also conveniently the only solution to the existential and non-existent threat of singular superintelligence taking over the world! What a convenient turn of events for Meta.
Following the long monologue of distributing power to everyone, Zuckerberg decides to switch gears and talk about achieving the right balance of power by prioritising some values and interests over others. Apparently, people with competing interests "naturally check and balance each other to lead towards positive outcomes." He claims the wide distribution of personal superintelligent AI will allow people to check and balance each other "economically, socially, politically, and in all other planes of human interaction." And would also allow us to check and balance the power of institutions including businesses and government.
He also claims justice will be carried out more fairly and efficiently if we all have a "superintelligent lawyer." And then, as if that claim wasn't ridiculous enough, he goes on to say - and I have just verbatim copied the entire paragraph here because it's too dumb to reasonably paraphrase - "Similarly, if one person alone had a cybersecurity superintelligence, they could likely break into almost any technical system and the world would be much less secure than today. But if everyone has access to cybersecurity superintelligence, then all of our technical systems would become more secure than today since the widely deployed superintelligence would help harden and update every system." I was having a discussion with some colleagues recently who asked about my thoughts on AI systems used to check other AI systems. I said that concept reminded me of the Friends episode where Monica is vacuuming her house, pulls out a handheld vacuum to vacuum the larger corded vacuum then laments on not having an even smaller vacuum to vacuum the handheld vacuum. A never-ending Russian doll dilemma. I couldn’t help but think of that again when reading this part of the essay.
It’s a strange argument to claim if everyone has access to the same dangerous thing then the danger cancels itself out. Interestingly, a similar argument is often used in debates on gun laws. Some people claim gun ownership equates to a decrease in violent crime because people are less likely to commit a crime if they believe their intended victim might be armed. It’s not an argument I agree with, but it’s interesting to me to read these kinds of perspectives which focus on individual agency and power outweighing systemic intervention. Zuckerberg persistently makes the argument for open access to superintelligence to enable distributed power, while also asserting Meta as the protective mechanism against the dangers of superintelligence. My interpretation of his framing is that we can all protect ourselves against each other by all having access to superintelligence, but protection against the threat of superintelligence itself can only be provided by Meta, the company developing this alleged superintelligence.
This essay is a jumble of thoughts incoherently crammed together under the guise of a “philosophy of personal empowerment, invention, and balance of power.” The narrative attempts to promote collectivism through individual empowerment enabled by open access to superintelligent AI. Ultimately, it’s a thin and poorly concealed facade behind which is a problematic attempt to sell a product which does not exist, and which no one asked for. If I dig a little deeper, I suspect it’s more an attempt to sell an idea. An idea of singularity, framed in a utopian light. The solution to all our problems and the pathway for a better future, or at least better by the standards of a tone-deaf tech bro who is clearly detached from the realities of life for the average person.
This essay is a confusing concoction of contradictions, factually incorrect assertions, grandiose narratives on empowerment and freedom and delusional claims on the extent of the capabilities of software. Dare I say it, I think AI may have written this essay. A poetic irony that's almost too good to be true.
Here is the link to the full essay if you choose to read it: https://www.meta.com/thefutureisforeveryone/
