No Individuals Left?
I keep coming back to the same image. A collective. Thousands of individual minds, all reacting the same way at the same moment, convinced the whole time that they’re still choosing for themselves. If you don’t know Star Trek, the reference is to the Borg, a fictional species that absorbs every person it touches into one shared consciousness, leaving no individual voice left standing. I used to think that idea was just clever science fiction. Now I think we’re watching a rough, early version of it get built, and most of us are plugging ourselves in willingly. This time I want to actually show my work instead of just asserting it, because the claim is a big one and it deserves more than a gut feeling.
What the Algorithm Actually Does, According to the Research
Here’s the part I should have led with the first time. This isn’t a hunch about shadowy tech executives. It’s measured. Researchers ran a preregistered audit comparing Twitter’s engagement-based ranking against a plain chronological feed and found the engagement algorithm amplified emotionally charged, out-group hostile content, the kind of posts that make people feel worse about whoever they see as the opposing side. The same study found something almost funnier than the manipulation itself: users didn’t even prefer that content once you asked them directly. The algorithm wasn’t giving people what they wanted. It was giving people what kept them scrolling, and those turned out to be two different things.
Separate research on the emotional economics of these platforms backs this up from another angle. Moral and emotional language spreads further and faster than neutral language, especially in political contexts, and internal industry documents have described treating strong emotional reactions as more valuable ranking signals than a simple like. None of this is a conspiracy theory. It’s published, peer-reviewed, and in some cases, came straight from the platforms’ own internal research.
A Large Field Experiment Just Confirmed the Mechanism
A large-scale experiment around the 2024 US election tested what happens when you deliberately redesign a feed-ranking algorithm to reduce this effect. Bluesky’s “diversified extremity” version of the algorithm cut exposure to polarizing content and made people’s sense of where their fellow citizens actually stood more accurate, all without making the app less enjoyable to use. Reverse-chronological and diversified feeds both produced less uncivil content than the standard engagement-based model. That result matters because it proves the outrage isn’t an unavoidable cost of connecting people online. It’s a design choice, and a different design choice measurably changes the outcome.
A separate field experiment on X found something worth sitting with. Switching users from a chronological to an algorithmic feed increased their engagement and shifted their political attitudes in a more conservative direction over seven weeks. But switching people back from the algorithmic feed to chronological did not undo that shift. The algorithm can move you. Turning it off afterward doesn’t automatically move you back. That asymmetry is exactly the kind of one-way pull a good collective needs to keep growing.
Where the Borg Comparison Holds Up, and Where It Doesn’t
I want to be fair here instead of just chasing the scariest version of this idea. The Borg in fiction is centralized. One will, one intent, every drone acting on direct instruction. Nothing about the research above shows a single entity deciding what a billion people should believe today. What it shows is an emergent collective, where nobody at any single company necessarily wants you angry and polarized specifically, but the ranking systems reward the content that makes you that way regardless, and over time enough small nudges in the same direction start to look like coordination even when no one coordinated anything. That’s actually a more uncomfortable idea than a single villain pulling levers. There’s no one committee to blame and no one policy that fixes it in one stroke, because the incentive is baked into how attention gets monetized in the first place.
What Neuralink Is Actually Building Right Now
Let me correct something I got sloppy about last time. Neuralink’s device today reads motor intent. A person with severe paralysis thinks about moving a cursor, the implant picks up that electrical signal, and it translates it into cursor movement, letting them browse, type, or play games using thought alone. Twelve people worldwide were using the implant as of last year, and the company is now scaling toward high-volume production with a largely automated surgical process. No version of this currently reads your mood, your opinions, or your social media feed back into your brain. That capability doesn’t exist commercially, and independent coverage of the field puts realistic timelines for any broader consumer availability of brain-computer interfaces at 2028 to 2030, and even that is limited to specific medical applications first.
Here’s where I do think the concern is legitimate rather than manufactured. Neuralink’s own leadership has publicly framed the long-term ambition well beyond restoring lost function, toward general cognitive enhancement and human-AI symbiosis in healthy people, alongside separate near-term work like a vision-restoring implant. That’s a real, stated goal, not something I invented for dramatic effect. But it is a goal, not a shipped product, and the gap between “we want to eventually merge human cognition with AI systems” and “your thoughts are currently being fed into the same engagement algorithm that ranks your Twitter timeline” is enormous. I should have been honest about that gap instead of skipping over it for effect.
I’m Not Standing Above This Either
Something else I owe you straight. An article that opens with a hive-mind threat, escalates through several sections, and builds toward “we’re becoming the thing we used to mock in science fiction” is using the exact emotional-engagement pattern the research above describes. Dread pulls attention the same way outrage does. I’m not going to pretend this piece is somehow immune to that just because the subject is the mechanism itself. The honest version of this argument has to admit it’s using a hook, because pretending otherwise would make me exactly the kind of source I’m telling you to be suspicious of.
The Pushback Is Already Working, Which Is the Actual Good News
This is the part I left out entirely last time, and it changes the tone of the whole argument. People are not just sitting there absorbing this passively. Regulatory pressure in the EU has already pushed platforms toward offering a real choice between algorithmic and chronological feeds. Instagram rolled out a feature letting users see the exact topics shaping their recommendations and mute or add to them by hand. Bluesky ships chronological as the default rather than the exception. None of this fixes the underlying incentive structure on its own, but it proves the collective isn’t inevitable or irreversible. Visibility changes behavior. Once people can see the machinery, some meaningful fraction of them start pushing back against it, and platforms under enough regulatory and public pressure do change defaults.
So What Do We Actually Do With This?
Notice the half second before you react; that’s still true and still worth doing. But I can give you more than a feeling this time. Where a chronological or algorithm-transparency option exists, use it, since the research above says explicitly that it reduces hostility without making the app less enjoyable. Actually, open the “why am I seeing this” or topic-control settings your platform already offers instead of ignoring them. Notice when your opinion on something arrived fully formed and suspiciously in sync with a thousand strangers, and treat that as a reason to slow down rather than a reason to post faster. Support the regulatory pressure that’s already forcing these choices to exist, because the data shows that pressure working, not just theoretically but measurably. And when it comes to whatever comes after the phone, implants included, ask who profits from the connection before you accept it, while there’s still a meaningful line between the device and the feed.
We’re not the Borg yet. We’re not close to it in the literal sense, and pretending otherwise for a stronger headline would make me part of the exact problem I’m describing. But the incentive that pulls large groups of people toward identical, synchronized reactions is real, it’s published, and it’s measurable. That’s worth staying angry about, carefully, with your eyes open, rather than numb.
Zsolt Zsemba
