July 31, 2026
Ctrl+Alt+Delete Your Thesis
"the very foundation of modern academia has been blown to bits"
Prof says AI can fake a PhD now, while commenters ask: was the old system overrated anyway
TLDR: A Berkeley professor says AI can now produce a thesis good enough to pass in many cases, putting the old PhD system under pressure. Commenters are split between alarm and mockery, with many arguing the bigger scandal is that academic gatekeeping may have mattered less than everyone pretended.
A math professor from UC Berkeley dropped the academic equivalent of a fire alarm: according to him, modern artificial intelligence can now churn out a passable PhD thesis “with the push of a button,” and the old foundations of higher education are basically toast. He walks through the long journey from dusty library stacks to Google Scholar to today’s AI tools, arguing that what once took years of slow reading, memorizing, and grinding can now be imitated in record time. But in the comments, the real show begins: instead of pure panic, a lot of readers responded with a shrug, a side-eye, or a full-on “and maybe that’s fine?”
That’s where the drama kicks in. One camp says the professor is exaggerating, because no machine can replace the years of training that shape a real expert. Another camp is much less sentimental, bluntly pointing out that the professor himself admits most PhD theses were already terrible, painful to read, and mostly ceremonial. So if AI can mass-produce something nobody was reading anyway, some commenters are asking the spicy question: what exactly has been lost? A few took the long-view, comparing today’s AI freakout to past moral panics over search engines and question-and-answer sites like Stack Overflow. The vibe is part doom, part roast, part dark comedy: maybe academia is collapsing… or maybe the comments section just discovered it had been running on paperwork, gatekeeping, and vibes for years
Key Points
- •The article quotes Tony Feng of UC Berkeley saying that AI now allows a passable PhD thesis to be produced with minimal effort.
- •The author describes how research moved from physical libraries and paper journals to searchable PDFs and Google Scholar, making literature review much faster.
- •The article states that annual PhD production increased from about 36,000 to over 58,000 in the United States between 1990 and 2024, and to roughly 88,000 per year in China from only a few thousand in the early 1990s.
- •It says research workflows have also been eased by code sharing, data sharing, drones, and robots used for data collection and experimentation.
- •The article argues that agentic AI and the low failure rate of oral defenses have undermined traditional assumptions about the value and evaluation of PhD theses.