The Riddler's wacky scheme in "Batman Forever" (1995) - sell vast quantities of a consumer device that entertains its users but makes them dumber, while extracting vast amounts of information from them and using that to artificially boost his organization's intelligence - sure seems pretty silly and alarmist in retrospect now after 30 years of the egalitarian, wealth-levelling World Wide Web, doesn't it.


Very much this.

I wonder if neural networks are gonna have another 1980s style collapse in a few years when people work out that they aren't a route to Artificial General Intelligence and an infinite spiral of exponential smartness.




tldr: Both "AI" as a general field of study in computing, and neural networks as the star of the "connectionist" subfield of AI have been around since the 1950s, and interest in both seems to boom and then bust in a recurring circle of life, like stock markets. Each time AI is cried up as utopia being finally here and then eventually reality sets in and there's a backlash.

We're right about at peak of the AI boom cycle now.


I mean if you count Isaac Asimov's fictional "positronic brains" then the idea of self-programming neural networks has been around since 1939.

But ours in 2021 are a little smarter because something something recursion and layers, but also mostly we just brute-force massive amounts of data and CPU power into them. Whether any actual insight is coming out the other end of the pipe is an open question. They're good at classifying images, but they were good at that in the 1980s too.

@natecull I toy with the AI in my shopping app's suggestions all the time. I'll pick a theme like "pictures with a lot of blue in them" and see if the AI will spot that pattern and start showing me suggestions that feature the colour blue.

They never do.

It's almost as if there's no intelligence whatsoever behind them.

Sign in to participate in the conversation

The original server operated by the Mastodon gGmbH non-profit