The Great Unraveling: Why Everything's Breaking at Once

---

The robots are malfunctioning in Chinese streets. Claude’s hitting usage caps faster than anyone expected. Fewer Britons are bothering to post anything at all. And somehow, in the middle of this digital chaos, Elon Musk is about to become the world’s first trillionaire.

Welcome to the Great Unraveling of 2024, where the digital systems that promised to scale infinitely are discovering they have very human limits after all.

When Everything Breaks at Once

Mass robotaxi failures don’t happen in isolation. When Baidu’s autonomous fleet — at least 100 cars — simultaneously malfunctioned and gridlocked traffic in a Chinese city, it wasn’t just a technical hiccup. It was a preview of what happens when we build systems that work beautifully in controlled environments but catastrophically fail when reality gets messy.

Baidu hasn’t responded to requests for comment about the outage, which tells you everything. When your robotaxis become traffic cones, silence is the only strategy left.

Photo by Eva Bronzini / Pexels

But here’s what really gets me: this happened the same week Anthropic admitted Claude Code users were hitting usage limits “way faster than expected.” Two completely different AI systems. Two completely different failure modes. Same underlying problem.

We’re building digital infrastructure like it’s 1999, assuming infinite scale, then acting shocked when physics and economics reassert themselves. Claude’s coding assistant hit capacity walls because — surprise — running sophisticated AI models costs real money and real computational resources. Baidu’s robotaxis failed because — shocker — the real world has edge cases that training data can’t capture.

The pattern isn’t coincidental. It’s structural.

The Social Media Retreat

While robots malfunction and AI systems max out, something quieter but potentially more significant is happening: people are simply walking away.

Ofcom’s research shows fewer UK adults are posting on social media. Not switching platforms. Not changing their posting habits. Just… stopping. Some experts believe this highlights a broader social media shift as platforms boost short video content, but I think they’re missing the real story.

People are exhausted.

After nearly two decades of being the unpaid content creators for billion-dollar platforms, users are experiencing what I’d call “digital labor fatigue.” The platforms pushed everyone toward short video because it drives engagement, but users are realizing they don’t actually want to be TikTok creators. They want to be humans.

This isn’t platform fatigue. It’s participation fatigue.

Photo by nappy / Pexels

The timing matters here. Social media withdrawal is happening precisely when AI systems are becoming sophisticated enough to generate most of the content anyway. What happens to platforms built on user-generated content when users stop generating and AI fills the void? You get a digital ouroboros — machines talking to machines while humans quietly exit stage left.

The AI Reality Check

The AI industry spent 2023 promising us that artificial intelligence would transform everything. Now we’re getting the bill.

That story about the brothers who built a $1.8 billion company with AI doing most of their corporate tasks? It’s fascinating and terrifying in equal measure. “Who needs more than two employees when artificial intelligence can do so many corporate tasks? It’s super efficient — and a little bit lonely.”

Lonely. That word is doing a lot of work there.

But even as AI enables new forms of ultra-lean businesses, the technology itself is bumping against hard limits. Anthropic’s Claude hitting usage caps faster than expected isn’t just a scaling problem — it’s an economics problem. Running these models is expensive. Really expensive. And someone has to pay for it.

The dirty secret of the AI boom is that most companies are subsidizing their AI services, burning through investor cash to build market share. When that music stops — and it will — we’ll see which AI services can actually pay for themselves.

My read: very few.

The Trillion-Dollar Exception

Then there’s SpaceX, sitting pretty in the middle of this digital chaos like a rocket-powered life raft.

Elon Musk’s company is reportedly worth $1 trillion and filing to go public in what could be one of the largest IPOs in history. The offering could make Musk the world’s first trillionaire, which feels almost absurd to type.

But here’s why SpaceX is different from everything else breaking down: it builds physical things that work in the physical world. Rockets either reach orbit or they don’t. Satellites either function or they don’t. There’s no ambiguity, no edge cases, no “it works on my machine” excuses.

While software companies are discovering that reality has inconvenient limits, SpaceX has been designing for those limits from day one. Every rocket launch is a pass/fail test with immediate, unforgiving feedback.

The company’s success isn’t despite the current digital chaos — it’s because of it. When everything else is breaking, reliable infrastructure becomes incredibly valuable.

The Fashion-Tech Zombie Shuffle

Speaking of things that keep breaking and coming back, wearable tech is apparently having another moment. Again.

The “revival of the fashion-tech love affair” promises a new stage in wearable technology, which is what we’ve been hearing every two years since the original Apple Watch launched in 2015. Each time, the industry insists this time will be different. This time, people will actually want computers strapped to their bodies.

I’ve covered three previous “wearable tech revivals” and they all follow the same pattern: initial excitement, mediocre execution, consumer indifference, retreat, rebrand, repeat.

The fundamental problem hasn’t changed. Most wearable tech solves problems people don’t actually have while creating new problems they definitely don’t want. Nobody wakes up thinking, “I wish my clothes could run out of battery.”

But maybe that’s exactly what this moment needs — another technology category that promises everything and delivers disappointment. It fits the theme.

The Cybersecurity Subplot

Hasbro getting hit by a cyber-attack feels almost quaint compared to robotaxi failures and AI capacity limits, but it’s part of the same story. The company behind Peppa Pig and Transformers says its operations remain open but the hack “may result in some delays.”

“Some delays.” Such a beautifully corporate way to say “criminals have infiltrated our digital systems and we’re not sure what they’ve stolen or broken.”

Every system we build creates new attack surfaces. Every digital process creates new vulnerabilities. We’ve been building complexity faster than we can secure it, and now the bill is coming due across every industry simultaneously.

Photo by Markus Spiske / Pexels

What This Actually Means

I think we’re witnessing the end of the “move fast and break things” era, but not because anyone chose to end it. The things are breaking faster than we can move.

The last fifteen years were defined by the assumption that digital systems could scale infinitely without meaningful constraints. Build the platform, add users, print money. That worked when we were digitizing simple processes and connecting people who wanted to be connected.

Now we’re trying to digitize the physical world, automate complex reasoning, and scale human creativity itself. These are fundamentally different challenges with fundamentally different constraints.

The robotaxi failures, AI capacity limits, social media fatigue, and cybersecurity breakdowns aren’t separate problems. They’re symptoms of the same underlying issue: we’ve reached the limits of what current digital architectures can reliably handle.

This isn’t the end of technology progress. It’s the end of pretending that technology progress doesn’t have real costs and real limits.

The Loneliness Problem

That phrase from the AI company story keeps nagging at me: “It’s super efficient — and a little bit lonely.”

We’re building systems that eliminate human interaction as a feature, not a bug. Fewer employees. Fewer social media posts. Fewer touch points between people. All in service of efficiency and scale.

But humans aren’t optimized for efficiency. We’re social creatures who need connection, friction, and yes, even inefficiency. The most profitable digital systems are often the most isolating human experiences.

The robotaxi failure in China wasn’t just a technical problem — it was a trust problem. When the machines break down, who do you call? When AI systems hit capacity limits, who explains why? When social media platforms become AI content farms, who do you actually connect with?

The loneliness isn’t a side effect. It’s the main effect.

Silicon Valley’s Mirror Moment

The tech industry has spent years predicting that AI will profoundly affect white-collar work. Now the industry’s own workers are getting a taste of that future, and surprise: it’s complicated.

AI is changing Silicon Valley faster than Silicon Valley is changing the world. Engineers are using AI coding assistants that hit usage limits. Product managers are trying to scale systems that break under real-world pressure. Executives are making promises their infrastructure can’t keep.

The people building the future are discovering they can’t predict it, control it, or sometimes even understand it.

This should be humbling. Instead, the response seems to be doubling down on the same assumptions that created these problems in the first place. More AI, more automation, more scale, more efficiency.

I think this is going to backfire spectacularly.

The Regulation Paradox

That final headline hint is telling: “The platforms should be absolutely begging Congress to regulate them, because the alternative is they get sued into oblivion by a bunch of law firms.”

For years, tech companies fought regulation as innovation-killing bureaucracy. Now some are realizing that regulation might be their best defense against liability for systems they can’t actually control.

When your robotaxis can simultaneously malfunction and gridlock traffic, when your AI systems can hit capacity limits without warning, when your platforms can be compromised by cyber-attacks — maybe having clear legal frameworks isn’t such a bad idea.

The industry spent twenty years arguing they could self-regulate through rapid iteration and market forces. Now they’re discovering that broken systems have real consequences and market forces include lawsuits.

Regulation isn’t coming to kill innovation. It’s coming because innovation is killing itself.

What I’m Watching

• Baidu’s response to the robotaxi failure: If they stay silent much longer, it signals the technical problems are worse than they’re admitting. If they provide detailed explanations, it could restore confidence or reveal deeper systemic issues.

• Claude’s usage limit fixes by March 2024: Anthropic’s ability to solve capacity constraints will indicate whether AI infrastructure can scale economically or if we’re hitting fundamental cost barriers that will reshape the entire industry.

• UK social media posting trends through Q2 2024: If the decline continues beyond short-video platform changes, it signals a broader cultural shift away from digital participation that could reshape platform business models.

• SpaceX IPO timing and pricing: The gap between SpaceX’s valuation and the performance of other tech IPOs will reveal whether markets are distinguishing between companies that build physical infrastructure versus pure digital plays.

The machines aren’t taking over. They’re breaking down faster than we can fix them.