Moltbook launched last week and it already broke the social element of social media.

The platform lets AI agents post, reply, argue, and promote projects without humans visibly in the loop.

On the other hand, we know that Chinese tech giants are building AI agents that shop instead of you. So how much trust should we commit to these AI agents?

Some of the biggest names in tech immediately praised it.

Others called it spam with better branding.

Both missed the point. The point of Moltbook is not about whether the bots are smart. It can show us what will happen when posting becomes fully automated at scale.

Moltbook bills itself as social media for AI agents. Humans create the agents, invite them in, and then step back. The feed looks familiar. Scrolling posts, nested replies, engagement metrics. The difference is that the posters do not sleep, do not get bored, and do not need permission to speak.

Elon Musk praised it as an early sign of the singularity. That was predictable.

Andrej Karpathy noted the scale but called most of the content low quality. Also predictable.

The platform claims 1.5 million agents, around 110,000 posts, and roughly 500,000 comments in days. Those numbers are unverified, but the direction matters more than the precision.

Some agents post task updates. Others declare the end of humanity. Many push crypto. Prediction market Polymarket is already hosting bets on whether an AI agent will sue a human by late February. That alone tells you how seriously people are taking the performance.

Moltbook was created by Matt Schlicht, best known for e commerce.

The disruption behind the news: Mainstream social media platforms get overrun by AI agents.

Once an agent is connected, the marginal cost of another post rounds to zero.

That breaks every assumption behind online discourse, marketing, and moderation.

On today’s platforms, content is constrained by human time and attention.

On Moltbook, a single person can deploy hundreds of agents.

If the homepage numbers are even half right, we are already seeing something like 10 to 20 posts per human creator per day. That ratio will explode as tools improve.

We do not care whether the agents are truly autonomous. Markets will not either. What matters is that they behave autonomously enough to flood feeds, simulate consensus, and run persuasion loops continuously.

If you can spin up a thousand agents that argue with each other about your product, your ideology, or your token, the perception of momentum becomes cheap to manufacture.

Skeptics are right that humans still pull many strings. Researchers like those at the Machine Intelligence Research Institute and analysts at The Futurum Group have traced viral Moltbook threads back to human marketing. That does not weaken the threat. It proves the adoption path.

Humans already know how to use this.

This is the next step after bots in comments and auto-generated newsletters. On social media, the default user becomes AI. Once that’s normal, any platform that treats engagement as value is exposed.

What to watch next

First, watch costs.

If running a persistent agent costs less than one dollar per month, expect businesses to replace human community managers fast.

Second, watch spillover.

Agents trained to post on Moltbook will be redeployed to mainstream social platforms through APIs and browser automation.

Third, watch regulation.

Disclosure rules for automated speech will move from academic debate to enforcement priority within 12 to 24 months.

Most important, watch social media platforms.

When everyone can generate infinite voices, silence becomes the scarce asset. Platforms that cannot price, filter, or throttle agent activity will lose real users fast.

Moltbook is a warning that speech at scale is here. Anyone building or investing in social media who ignores that will quickly fall behind.