Anthropic's AI: More Hype, Less Humanity

The Vatican isn't typically where you'd expect an AI debate to unfold, yet here we're. At a recent event, Anthropic's Chris Olah took the stage, colors flying, to grapple with Pope Leo XIV's words cautioning against equating machine 'intelligence' with human intellect. As if calling a company Anthropic gives license to humanize algorithms.

The Myth of AI Mysticism

Olah waxed poetic, describing AI as mysterious entities crafted from our own words, despite the systems being as spiritual as a spreadsheet. Naturally, AI models aren't cold, calculating robots, if anything, they're as warm as a server under load and as bad at arithmetic as a flustered student. But invoking spiritual mystery? Spare me the mysticism.

It's no wonder these 'enigmatic' algorithms are at the center of over 100 lawsuits. The real mystery isn't what these models do, but where they sourced their data. Spoiler alert: It's a secret Anthropic guards jealously.

Not a Bridge, Nor a Brain

Olah insists AI isn't engineered like airplanes or bridges. True, you don't need an architecture degree to build a neural network. But let's not pretend they're organically grown like a greenhouse flower. If anything, they're more akin to complex software cobbled together from massive data sets, not some esoteric alchemy.

More absurd is the notion that Anthropic 'inherited' this data. That's a bit like saying you inherited your neighbor's car because you duplicated the keys. Spare me the semantics.

Questions No One Asked

Olah posed three questions to the Catholic Church, but perhaps he should've started with Anthropic's own boardroom. He wonders how AI's benefits can be shared globally. Here's a thought: taxes and litigation, hardly new concepts. As for societal flourishing, it's a question humanity's grappled with long before chatbots were a thing. Maybe Olah should focus on the education sector, where his company's services could directly impact young minds, if managed responsibly.

His last question about AI's supposed introspection is the real kicker. What are these 'internal states' that mirror human emotions? Don't tell me we're talking about server logs. Joy, fear, satisfaction, these are human experiences, not lines of code. Let's keep our heads here. AI isn't about to feel anything but the rising electric bill.

Let's face it, AI is as intelligent as we define it to be. But conflating computational output with human consciousness? That's a leap too far. As Turing suggested with his Imitation Game, an imitation is still just that, an imitation.