The AI Bubble Fight: Three Camps, One Proxy War

Illustration: chili pepper mascot illustrating The AI Bubble Fight: Three Camps, One Proxy War

The AI Bubble Fight: Three Camps, One Proxy War

October 2026 is officially the month everyone decided to argue about the AI bubble. But let's be real: this isn't just about P/E ratios or market caps. This week's bubble discourse is a full-blown proxy war, fought on three distinct fronts.

We've got the schadenfreude crew, gleefully sharpening their knives for the tech titans. Then there are the dot-com veterans, twitching with 1999 PTSD. And finally, the genuinely furious, fed up with how this "magic" tech was sold.

Let's break down the battle lines.

Camp 1: The "It's a Con" Crew

First up, we have the folks who think generative AI was a grift from day one. Ed Zitron, for example, didn't mince words in an interview clip on October 13, calling the whole thing "a con."

He says AI was sold for three years as "job-replacing magic," but what we got was "boring cloud software." Worse, it's "unprofitably expensive and unreliable at its core." Companies, he argues, "misleading the entire world" by overstating capabilities and financials. That's a sentence an AI press release would never print.

Turns out, your "magic" AI can't even tell you basic science facts without hallucinating like a bad trip. Just ask Neil deGrasse Tyson. On October 12, he posted an Instagram reel (23K likes, 720 comments) explaining how a friend's AI claimed "carbon" had the highest melting point. Wrong answer, AI. It's tungsten.

Tyson's point? LLMs don't understand science; they just assemble statistically common word sequences. "Carbon" co-occurs with melting-point talk often enough to fool the algorithm. Tungsten, which Edison used for light-bulb filaments because it stays solid while glowing, has an extremely high melting point. Carbon, by contrast, sublimates at atmospheric pressure rather than truly melting. Tyson warns that AI fools people who don't know the subject, urging verification against authoritative sources.

This camp is less about market speculation and more about the fundamental dishonesty of the hype cycle. They're mad about the bait-and-switch.

Camp 2: The Pop-It-Now Cheerleaders

Then there's the crowd that just wants to see the whole thing burn. And they want it now. On October 13, Futurism posted a graphic on Facebook to its 6.3 million followers, quoting Joe Wilkins: "MAINSTREAM NEWS ANCHORS NOW TALKING ABOUT THE AI BUBBLE BURSTING WITHIN MONTHS."

The comments section was a bloodbath of schadenfreude. Many were openly cheering for a bubble burst, specifically hoping it would wipe out the likes of Altman, Thiel, and Musk. Michael Burry, the "Big Short" investor, even weighed in on Substack, saying he's "moving timelines up" because the AI bubble "may burst sooner than later."

This isn't about the tech's utility; it's about the spectacle. It's the digital equivalent of grabbing popcorn and watching billionaires squirm.

Camp 3: The Dot-Com Veterans

Then there’s the crowd who’ve seen this movie before, probably with dial-up modem sound effects. They’re the dot-com veterans, the ones who remember when "web" was the magic word that opened investor wallets. The Betoota Advocate, Australia’s finest satirical export, nailed it on October 12, 2026, with a headline about the Firmus failure, hilariously blaming the investors themselves. The comments section was a goldmine, with 890 likes and 46 comments comparing AI datacenter fundraising to the late-1990s internet boom. One commenter put it plainly: "It's like 1999 again - everything with AI gets heaps of investment like everything with web." Skeptics even dragged up executive Oliver Curtis's criminal conviction, just to really twist the knife.

These veterans aren't saying AI is a total scam, just that the valuations are bonkers. Foreign Policy Magazine, on the same day, dropped an Instagram carousel (420 likes) titled "The Real Danger May Be AI as Dumb as Us." Their point? The risk isn't Skynet, it's AI being exactly as dumb as us, but at the absolute worst possible moment. They even slapped AI object-detection labels on Michelangelo’s Creation of Adam for the visual. It’s a reminder that even if the bubble bursts, the tech itself might still stick around, albeit with a lot less fanfare. As one Futurism commenter put it on October 13, 2026, "stocks ranked but www was transformative nonetheless." The internet survived the crash; maybe AI will too, just with fewer private jets for the founders.

The Wildcard: Fortnite Devs vs "No More Hand-Written Code"

Forget philosophical debates; the real battle is happening in the trenches. IGN, via Bloomberg, reported on October 13, 2026, that Epic is "REPORTEDLY HEAVILY PUSHING DEVS TO USE AI" with a blunt "NO MORE HAND-WRITTEN CODE" directive. This is the kind of sentence an AI press release would never print. The Instagram post, with 6.1 million followers, racked up 16K likes and 590 comments, and it was a bloodbath. Developers were furious, citing layoffs, quality loss, and the erasure of craft. They called the AI output "slop" and some even threatened to quit Fortnite. A brave minority tried to defend AI coding as inevitable, comparing it to Excel, but they were largely drowned out by the digital pitchforks. This isn't about AI's potential; it's about its immediate, messy impact on livelihoods and creative control.

Tyson's Tungsten, or: Why Both Sides Need the Same Story

Neil deGrasse Tyson, ever the voice of reason, gave us the perfect Rorschach test on October 12, 2026. In an Instagram reel via @startalk (23K likes, 720 comments), he recounted how a friend asked an AI for the element with the highest melting point. The AI confidently declared "carbon." Tyson, bless his heart, corrected it to tungsten. His explanation? LLMs just assemble "statistically common word sequences." "Carbon" hangs out with "melting point" talk often enough to win, even though tungsten's melting point is so high Edison used it for lightbulb filaments because it stays solid while glowing. Carbon, meanwhile, just sublimates. Tyson's warning: AI fools people who don't know the subject.

The "con" camp sees this as proof positive of AI's fundamental unreliability – a fancy autocomplete that can't even get basic science right. The dot-com veterans, however, might just shrug. For them, it's a dumb tool, sure, but a tool nonetheless. You wouldn't trust a hammer to do your taxes, and you wouldn't trust an LLM to build a bridge without a human engineer checking its work. It's not about the AI being smart; it's about knowing its limitations and using it anyway.

The discourse

Neil deGrasse Tyson: AI fools those who lack subject knowledge, thanks to its statistical wordplay. Futurism's commentariat: The tech survives, even if the valuations don't. The Betoota Advocate's satirists: AI datacenter fundraising is a rerun of the dot-com bubble, complete with investor blame. Foreign Policy Magazine: The real danger isn't superintelligence, but AI being as dumb as us at the wrong moment. * IGN's Fortnite commentariat: Epic's "NO MORE HAND-WRITTEN CODE" directive is leading to "slop" and developer revolts.

Sources & further reading

The Betoota Advocate on Firmus: https://www.facebook.com/betootaadvocate/posts/pfbid02v21M8XvrSXmtSWA7PzLVUXrFQ8XrGFXv4tBekT7PpAQcHUinA8HZsvhdD9KtyMwBl Foreign Policy on "Dumb AI": https://www.instagram.com/p/DeZdg1wl_Bp/ IGN on Epic's AI push: https://www.instagram.com/p/Deaf9TFjGh4/ Neil deGrasse Tyson on Tungsten: https://www.instagram.com/reel/DeZ1Bcksa3J/ * Futurism on dot-com parallels: https://www.facebook.com/futurism/posts/pfbid02uwy1QGeHvt5qvKgHHJ4CvWZX3y3NkaeZavp9CQAQqngXYUEVVJoVtbDsk2yYWTf6l

Our verdict

So, what's the verdict from spicyautocomplete.com? It's a gold rush, alright, but the gold is often fool's gold, or at best, a useful but profoundly dumb tool. The real danger isn't superintelligence taking over the world; it's super-hype meeting mediocre reality, leading to WorldCom-scale bankruptcies and a whole lot of "slop" that developers are threatening to quit over. AI isn't going to kill us, but it might just bore us to death with its statistically common word sequences, after it's already drained our wallets.