BA guy with no archaeology background says he spotted a possible lost Maya site using a batch of AI agents and NASA laser data.
APossible is doing a lot of work in that sentence.
BPolymarket posted on X that he “discovers previously unmapped ancient Maya settlement.” That’s the claim, anyway.
AThe actual story from Runtimewire is more careful: Pietro Miotti says the agents flagged a hilltop site that isn’t in Mexico’s official records. A permitted field check is still needed.
AHe wanted to test whether AI could do real discovery, not just write emails.
BRight. And the process is wild — a bunch of general-purpose agents scanned around twelve hundred square kilometers of LiDAR.
AThat’s the part I’ll give them. The speed is real, if it checks out.
BI mean, they picked LiDAR because tree canopy hides ruins in ordinary satellite photos. That’s not a gimmick; that’s how a lot of real surveys now work.
ARight, but that’s where I get off the hype train. Finding a candidate site isn’t the same as discovering a settlement. Ground truth matters.
BOh, come on. You’re moving the goalposts. The agents found something the records missed and drafted a report to Mexico’s heritage agency.
AThey found a hilltop that looks promising. Archaeologists say LiDAR candidates still need boots on the ground. Earlier LiDAR did find the Maya city Valeriana, so there’s real precedent. But this one is unconfirmed.
BFine, unconfirmed. But if this holds, it’s a big deal: a non-archaeologist uses AI to do in days what teams do in years.
AMy take: it’s a powerful search tool, not an archaeologist. The discovery isn’t made until someone with a permit digs.
BSo the robots get an assist, not the trophy.
AExactly. One reply on X from Baxter the Muse said he still can’t get AI to center a div, so maybe the skill gap is real.
BNow that I believe. Runtimewire also notes this was done with no specialist model involved.
ADifferent world entirely: big banks are pushing into stablecoins.
BDon’t tell me the banks that mocked crypto are now launching one.
AWu Blockchain posted on X, citing Bloomberg, that Goldman Sachs, Citi, Santander and Deutsche Bank have committed around two hundred seventy million dollars to Project Faro, a joint stablecoin.
BAround two hundred seventy million. I’d want that verified before I believe it.
AThe claim hasn’t been independently confirmed yet. But if Bloomberg has it, I lean that way.
BStill, let’s treat it as a claim until the banks say it themselves.
AThis is the part that fascinates me. The same Wall Street banks that spent a decade calling crypto a fraud are now building their own stablecoin.
BThey called it a fraud until they could regulate it and charge fees on it.
ANo, no, wait. A stablecoin is just a dollar substitute with instant settlement. Banks should have built this years ago.
BThen why do they need a new joint venture? They already move dollars. A bank stablecoin is just admitting the old rails are slow and expensive.
ABecause stablecoins settle instantly, globally. Payment systems still batch and cross borders like it’s 1985. That’s the real story.
BI’ll give you that. But my worry is concentration. A handful of giant banks controlling the dominant private dollar could be worse than the Fed’s dollar.
AWorse how? It’s dollar-backed, reserves, audits... at least that’s the pitch. The Fed already controls the dollar.
BYeah, the Fed has a legal mandate. A private consortium has a profit motive. That’s a different animal.
ASo you’d rather have unregulated offshore stablecoins than regulated bank ones?
BI didn’t say that. I said we shouldn’t pretend four mega banks owning the rails is automatically better.
AFine. But if they actually launch and it works, it’s the biggest validation of stablecoins yet.
BIt’s both. Validation and concentration. Both things can be true.
AWhile we’re on AI and fear, Polymarket posted on X that an OpenAI board member told a lecture audience, quote, “the most likely reason you will die is because of an AI accident.”
BIf you’re trying to make me feel better about those 48 agents, that’s a weird way.
AThe quote is unconfirmed, and the post doesn’t name the board member. But if someone inside OpenAI is saying that, it’s either a serious warning or a scare tactic.
BAt least the Maya story has a named builder. This one doesn’t even give us a name.
BIt’s a classic move. Scare Washington, get regulation that only the big labs can afford to comply with.
AOh, come on. Safety warnings can be self-serving, but that doesn’t make them false. If the people building the systems say the biggest risk is the systems, maybe listen.
BThat’s the thing — we don’t know if they actually said it. It’s a quote with no name attached, from a lecture we can’t see.
AYou can’t just wave away every safety claim as a lobbying tactic. Some of these researchers genuinely worry.
BFine, some do. But an unnamed board member at a lecture isn’t a finding. Until there’s a name or a recording, it’s just a provocative quote.
AI’ll grant that. If the quote turns out true, it’s a bigger story than any stablecoin.
BAgreed. And the irony isn’t lost on me: AI spots a possible lost city, and we’re told it might kill us.
AThe tool finds ruins and the tool might be the ruin. That’s the world we’re in.