November 3, 2025
AOL or LOL?
AI's Dial-Up Era
Is AI the next AOL or a GPU graveyard? The comments erupt
TLDR: The piece says we’re in an early-internet-style moment for AI, with jobs like radiology thriving instead of vanishing. Comments explode into a bubble vs. boom showdown: railroads vs. GPUs, 1999 vs. 1995, and one killer reminder that hype isn’t genius—making this a must-watch tug-of-war over where AI is headed.
Is AI our new dial‑up? The comments turned into a time machine. One camp, led by arcticbull, says the internet‑railroad analogy is broken: railroads left tracks; an AI bust leaves warehouses of Nvidia cards that go to zero. Ouch. runarberg counters that the comparison is about bubble economics, not tech, and claims we’re living 1999, not 1995. mjr00 pokes the story’s opener: everyone already knew the internet was a revolution before 1995, so stop pretending it was a surprise. The drama: Are we early pioneers, dot‑com day‑traders, or just collecting shiny GPUs like beanie babies? Reddit memes about “AI as an expensive space heater” made the rounds.
The article itself urges nuance: AI hasn’t erased jobs wholesale—radiology is thriving, per Deena Mousa, and U.S. programs offered a record 1,208 residencies. Disruption, yes; extinction, not so fast. Enter bigwheels, who drops a macro flex: five gauges show a demand‑led boom, not frothy speculation. Then bena swings in with the quote heard ’round the thread: they laughed at Columbus—and at Bozo the Clown. Translation: being mocked doesn’t make you a genius. Between jokes about AOL CDs and Uber strangers, the crowd agrees on one thing: AI’s future is big, messy, and wildly contested.
Key Points
- •The article compares today’s AI stage to the early internet era, highlighting similar optimism and skepticism.
- •It recalls 1990s internet limitations and forecasts, including a 1993 prediction about widespread head-mounted displays.
- •Geoffrey Hinton’s 2016 warning that AI would replace radiologists is contrasted with current outcomes.
- •Research cited (Deena Mousa) indicates radiology has not been automated away by AI.
- •U.S. diagnostic radiology residency programs offered a record 1,208 positions in 2025, up 4% from 2024.