Silicon Valley's Consent Problem Is About to Get Ugly

---

Meta just killed encrypted messaging on Instagram. Not because users demanded it. Not because the technology failed. Because Ofcom—Britain’s telecom regulator—is charging Meta fees it considers unreasonable, and the company decided to strip privacy rather than pay up or fight fairly.

That’s the lede. That’s the scandal. Everything else flows from there.

Photo by Brett Sayles / Pexels

The Privacy Reversal That Says Everything

Instagram had end-to-end encryption on DMs. You knew this. Meta knew this. Then one day it didn’t, and the company framed it as a choice. But flip the actual timeline: Meta’s battling UK regulators over fees it says are “disproportionate,” and suddenly privacy becomes optional. The causality isn’t subtle.

Here’s what makes this different from typical corporate privacy erosion: Meta isn’t claiming this makes their product better. They’re not selling users on convenience or new features. They’re removing a protection—one they’d already implemented—because a regulator made their business model slightly more expensive. It’s punishment masquerading as product development.

The Ofcom fight matters because it’s the first real test of whether tech giants can be regulated at scale. Meta’s bringing a High Court challenge. Ofcom says it’ll defend its position. This isn’t some toothless fine that gets paid and forgotten. This is a jurisdiction saying: you operate here, you pay proportional fees, and no, you don’t get to undermine public policy by stripping user protections in retaliation.

My read: Meta’s playing a dangerous game. They’re betting that killing a privacy feature will be less painful than accepting regulatory costs. But they’re telegraphing something uglier—that privacy is a negotiating chip, not a right. Other regulators are watching. They’ll remember this.

Photo by nappy / Pexels

The Musk Apparatus Eating Its Own Young

Shivon Zilis was on OpenAI’s board. While advising the company, she also began a relationship with Elon Musk. They now have four children together. She’s now being cast in a trial as Musk’s inside source at OpenAI—the very company he’s suing through his lawyers and his proxies.

This isn’t a scandal because office romance exists. It’s a scandal because the power structure is completely inverted from how we pretend governance works. Zilis wasn’t some junior employee navigating complicated feelings. She was positioned to influence OpenAI’s direction while maintaining a relationship with someone who had explicit financial incentives in the company’s decisions and trajectory.

The trial detail is almost beside the point. What matters is that Musk is now so deeply embedded in multiple AI companies—through lawsuits, through funding, through SpaceX’s new Terafab chip factory—that conflict of interest isn’t a bug in his strategy. It’s the feature.

SpaceX investing $55 billion into semiconductor manufacturing called Terafab? That’s not diversification. That’s vertical integration of AI infrastructure. Musk owns the rockets (SpaceX), he’s fighting the AI company (OpenAI), he’s trying to acquire Twitter (X), and now he’s manufacturing the chips everyone will need. It’s a closed loop. The Zilis situation is just the human embodiment of how tangled this web has become.

Here’s where I genuinely don’t know what happens next: Does this end in criminal charges? Regulatory action? Or does Musk’s wealth and profile simply immunize him from consequences that would destroy anyone else? I think we’re about to find out.

The Heat-Trapping Molecule Nobody’s Talking About

Somewhere between Instagram’s privacy collapse and Musk’s empire-building, there’s a molecule that could actually matter.

Scientists figured out how to use molecules that absorb heat—inspired by how sunburn works—to store thermal energy. This isn’t a headline-grabbing breakthrough. There’s no billionaire attached. No lawsuit. No congressional hearing scheduled. But read it twice: a new technology for storing heat could actually help decarbonize heating systems.

Heating accounts for roughly half of Europe’s energy consumption. It’s massive. It’s unsexy. It doesn’t get VC funding or Twitter discourse. But a molecule that captures and releases heat on demand? That’s infrastructure. That’s the boring, necessary work that actually reduces emissions.

I mention this because the contrast is instructive. We’re obsessed with Musk suing Altman while the actual tools for solving climate problems sit in academic journals. The AI companies are spending fortunes on compute and lawsuits. Meta’s fighting regulators over fees. And researchers found a molecule that works.

When AI Search Beats Google (and Doesn’t)

Google’s new AI search isn’t good at celebrity gossip. It’s excellent at picking groceries and detecting scams.

This is weirdly specific, and I’m telling you about it because it reveals something most people miss: AI search isn’t a universal replacement for Google. It’s better at some tasks and worse at others. You want AI to help you choose vegetables at the market? Go. You want to know if a website’s a scam? Use it. You want celebrity news? Stick with traditional search.

The implication: the “AI will replace search” narrative was always too broad. What’s actually happening is task-specific displacement. Google’s moat was universal search dominance. If AI only wins certain categories, that moat survives.

Anthropic’s CEO saying the company could grow 80 times this year tells you something different: the compute requirements are exploding. More users, more tasks, exponentially more GPU-hours needed. The ceiling isn’t the technology. It’s the chips, the power, the infrastructure. Which brings us back to Musk’s Terafab and the frantic race to own the semiconductor layer.

Photo by Markus Spiske / Pexels

The Robot Monk and What We’re Actually Asking Technology to Do

South Korea got a robot Buddhist monk. It vows not to overcharge. This is either a joke or a prophecy, and I genuinely can’t tell which.

But it crystallizes something: we’re so convinced that technology solves every problem that we’re now trying to solve spiritual and ethical questions by automating them. A robot that monks can trust not to overcharge is technically impressive and philosophically bewildering. It suggests we’ve given up on humans being trustworthy, so we’re building systems that physically can’t betray us.

It’s not connected to the other stories directly, but it’s the same impulse: when human institutions fail (privacy, fairness, earnings calls), we outsource to technology. Instagram removes encryption because regulators are annoying. OpenAI hires someone in a complicated relationship with Musk. AI search sometimes lies about celebrities. And now a robot takes monastic vows.

We’re not fixing the underlying problems. We’re automating away the friction points.

What I’m Watching

• Ofcom vs. Meta High Court decision timeline. If Meta wins, regulatory oversight of Big Tech just got a lot weaker. If Meta loses, every tech company’s operating costs in Europe jump. Watch for the ruling date and what it signals about whether regulators can actually enforce consequences.

• Anthropic’s capital requirements through Q3 2024. Dario Amodei said 80x growth requires exponentially more compute. Watch the announcement of their next funding round’s size. If it exceeds $10 billion, we’re in a chip-supply war that makes the semiconductor shortage look quaint.

• Terafab’s first actual chip production. SpaceX announced the factory. When does it actually produce semiconductors? When does the first batch ship? This is the moment Musk stops talking and starts competing with TSMC. If it ships by late 2024, his vertical integration strategy works. If it slips past 2025, it doesn’t.

• The Musk v. Altman trial verdict and regulatory response. This isn’t just about who gets sued. Watch whether any government body opens an investigation into the Zilis-Musk-OpenAI dynamic. That’s the real test of whether conflict of interest rules apply to billionaires.