Americans are living through a multigenerational project of stamping out solidarity and insisting that we only ever view ourselves as individuals, with no stake in the plights of our neighbors. That's how the US got the most expensive, least effective health care system in the world. And even if you are in the vanishingly tiny minority of Americans who are happy with their health care, you live amongst people who are being killed by the system around you.
The health system is a perfect example of how monopolization drives more monopolization, and how that comes to harm the public and workers. Health consolidation began with pharma mergers, that led to pharma companies gouging hospitals. Hospitals, in turn, engaged in a nonstop orgy of mergers, which created regional monopolies that could resist the pricing power of monopoly pharma – and screw insurers. That kicked off consolidation in insurance, which is why most Americans have a "choice" of between one and three private insurers – and why health workers' monopoly employers have eroded their wages and working conditions.
A new study in American Economic Review: Insights puts some quantitative spine in this tale, tracking the relationship between hospital mergers and skyrocketed health-care prices: https://harris.uchicago.edu/news-events/news/consolidation-hospital-sector-leading-higher-health-care-costs-study-finds?itid=lk_inline_enhanced-template
The researchers investigated 1,164 acute-care hospital mergers, finding that while the FTC only challenged 1% of these, they could – and should – have challenged 20% of them, based on the agency's own criteria for merger scrutiny. The researchers blame the rising costs of hospital care directly on these mergers, and point out that Congress has historically starved the FTC of the budget it needed to investigate these mergers. The annual additional costs to the American people from these mergers exceed the entire annual budget of the FTC.
It's not just hospitals: the entire investor class is hell-bent on spending their way to monopoly. Nowhere is that more true than in AI, where hundreds of billions are being poured into bids to attain permanent dominance through scale. Writing for their excellent AI Snake Oil newsletter, Arvind Narayanan and Sayash Kapoor inject some realism into the AI scale hype: https://www.aisnakeoil.com/p/ai-scaling-myths
Narayanan and Kapoor challenge the idea that throwing more data at large language models will make the better: "With LLMs, we may have a couple of orders of magnitude of scaling left, or we may already be done." They are skeptical that this can be fixed with synthetic data (whose use is limited to "fixing specific gaps and making domain-specific improvements"). They also point out that if returns from data slow, then adding more compute or making bigger models might also be throttled.
They reserve their most skeptical take for "AGI" – the idea that LLMs are going to achieve consciousness. This is a fundamentally unserious idea, one that they unpack in detail in their forthcoming book: https://press.princeton.edu/books/hardcover/9780691249131/ai-snake-oil
One thing I'm hoping for from the book is some analysis of the material usefulness of AI hype – what purpose does the hype serve? I mean, obviously, hype is useful if you're looking to suck up investor capital, or flip an investment to a greater fool. But there's a specific character to AI hype: namely, the claim that AI will displace labor, which is really a claim that a bet on AI is a bet on the increasing wealth of capital at labor's expense.
In other words, AI is a bet on oligarchy. In America, that's a pretty safe bet, and the odds just got even better, thanks to a string of brutal Supreme Court decisions that legalized bribery, banned most regulatory enforcement, and made being alive and unhoused into a crime (Poor Laws 2.0): https://prospect.org/justice/2024-06-29-whos-gonna-check-supreme-court-chevron-separation-powers/