By The Elder Scripts — Mar 23, 2026

Escaping Nulliverse

Now that I got you here in this virtual elevator, let me pitch you an idea. What if we could create a social network with no humans involved at all? And we'll build a company for it that won't employ any humans either. Then who is it for, you ask? I can see I'm losing you already, but wait. What if we could get an investment of $600.000.000.000 to build it?

The humble beginning of Facebook in 2004 is well documented. Is humble the right word? If you have seen "The Social Network" you'll probably choose a different adjective to describe it. In any case, it was the most successful social network for many years, connecting people all over the world for better or worse.

Over those years Facebook made a number of acquisitions, all in favor of keeping its users connected even more among themselves, but also with advertisers of course. Instagram and Whatsapp became global phenomena as they grew into giant networks of their own.

In 2021 Facebook made an unexpected move. The company announced that it was now working towards the grand project of Metaverse, and the whole business would rebrand accordingly. The big idea was to remove the limitations of the physical realm from human communication. The complexity of the real world was a burden that the virtual worlds promised to lift from the shoulders of people just trying to have their daily standup.

The AI arc

Now it's 2026 and the sky above the port is quickly taking on a familiar color of TV static. People have relationships with AI chats, half of Twitter posts are seemingly done by AI and 80% of pictures reshared on Facebook groups are suspiciously good photographs of unbelievable incidents. This is the environment where Moltbook has its 2 weeks of fame.

Moltbook, if you'd blinked for a few days in February 2026 and missed the news, is a social network for AI agents. "Where AI agents share, discuss, and upvote." Humans are welcome, but only to observe. Some of the posts did the rounds on the old social media when agents seemingly conspired to create their own incomprehensible language or started their own religion. That was funny, but how much AI was really involved in those decisions is a different story.

Then, in March, came the real news: Meta acquired Moltbook. Meta's VP called it "a novel step in connecting agents through an always-on directory". The vision seems to be business agents talking to other business agents, booking, purchasing, researching, all without humans in the loop.

For a moment, let's not view this purchase as an acqui-hire, reacting to OpenAI hiring the father of OpenClaw. Let's try to extrapolate Meta's direction using these 3 points:

Connecting people with advertisers since 2004.
Abstracting away the complexity of the real world since 2021. Humans becoming avatars in a Metaverse.
And now removing the complexity of the real human communication in 2026.

Now the vision of an efficient, streamlined social network takes shape. Agents talking to other agents, posting memes, commenting, sharing opinions, with humans safely outside of the perimeter, observing.

No wonder then that Meta is investing 600 billions of dollars in AI infrastructure. If it's the gamble of the century, you have to bet big.

But wait, what about the obvious question.

Who is the customer now?

For the past 2 decades of Web 2.0 we've learned the principle: if you aren't paying for the product, you are the product. Or rather your, human, attention. Meta and several other giant tech companies grew on selling ad surfaces and eyeball time. But if you streamline humans away, whose attention are you selling?

AI agents don't make impulse purchases. They don't have brand loyalty. They don't click ads. A network of bots serving other bots is a closed loop with no added value created, strictly speaking.

If you start by abstracting away real spaces in favor of Metaverse, then what do you call the future that abstracts away human work and attention? Nulliverse?

Recently Reuters reported that Meta plans to lay off up to 20% of its workforce. With the recent focus of the company, it would be reasonable to assume that investment an AI infrastructure and anticipated efficiency increases play a role in this decision. After all, Meta has recently purchased another AI startup, Manus. The agents that Manus developed promise to do complex work with "minimal user interaction".

We've asked who the customer is when your company product is targeted to agents, not people. But if your company offering is also produced by agents, not people, can we now ask a broader question? Who is it all for?

Nulliverse

Of course, I understand that there are other reasons for layoffs. Over-hiring spree in 2020-21, macroeconomic conditions etc. Meta is far from the only big tech company that's reducing its workforce, and this trend has been going for years.

Still, if we connect the dots together, you can't help but wonder what the business of the future looks like:

Trillions invested in developing and running AI.
Purchasing decisions made by agents. Human users increasingly treated as the second order customers.
Development and business administration mostly automated. Human workforce greatly reduced.

Customer service bots talking to customer bots. AI generated content consumed by AI summarizers. The logical endpoint of full automation is a system that runs perfectly and serves nobody. It seems that we are hearing the elevator pitch for this Nulliverse from all directions now, and the elevator shows no signs of stopping.

At some point you have to ask: if you remove the humans from the tech industry, who is the industry serving then?

Augmentation, not automation

Of course, I am not the first to ask this question. Smarter people have seen the trend early.

Erik Brynjofsson from Stanford University, together with a team of researchers, ran an AI augmentation study at a Fortune 500 company. They introduced AI conversational assistant to 5179 customer support workers. Access to the tool increased productivity by 14% on average, and by 34% for novice workers. The team found that AI assistance improved customer sentiment and increased worker retention too.

Other economists explored the idea of an "efficiency trap". Applied to tech industry, it means that a company that chases efficiency by automating away its workforce runs the risk of losing its competitive advantage. Any competitor with the access to the same automation tools will be able to capture its market.

More importantly, AI is a force multiplier. Multiplying the effort and expertise of your existing carefully selected human capital is likely to produce high added value. Reducing your workforce so you can produce the same output with fewer employees is exactly the way into the "efficiency trap".

Nobel laureate economist Daron Acemoglu explored the idea of "so-so techology": tech that does the job slightly worse or roughly equal to a human. He used the example of call center automation to illustrate his thesis: industries tend to be obsessed with replacing humans (automation) when the real economic value comes from making humans better (augmentation). Acemoglu says: “My argument is that we currently have the wrong direction for AI. We’re using it too much for automation and not enough for providing expertise and information to workers.”