The IQ in the US started declining since the start of the Internet:
https://www.popularmechanics.com/science/a43469569/american-...
"Leading up to the 1990s, IQ scores were consistently going up, but in recent years, that trend seems to have flipped. The reasons for both the increase and the decline are sill [sic!] very much up for debate."
The Internet is relatively benign compared to cribbing directly from an AI. At least you still read articles, RFCs, search for books etc.
As someone who grew up reading encyclopedias, LLMs are the most interesting invention ever. If Wikipedia had released the first chat AI we'd be heralding a new age of knowledge and democratic access and human achievement.
It just so happens unimaginative programmers built the first iteration so they decided to automate their own jobs. And here we are, programmers, worrying about the dangers of it all not one bit aware of the irony.
As someone who grew up reading encyclopedias, I find LLM:s profoundly hard to find a use for besides crude translations of mainly formal documents and severely unreliable transcriptions.
I like structured information and LLM:s output deliberately unstructured data that I then have to vet and sift out and structure information from. Sure, they can fake structure to some extent, I sometimes get XML or JSON that I want, but it's not really either of those and also common that they inject subtle, runny shit into the output that take longer to clean out than it would have to write a scraper against some structured data source.
I get that some people don't like reading documentation or talking to other people as much as having a fake conversation, or that their editors now suggest longer additions to their code, but for me it's like hanging out with my kids except the LLM is absolutely inhuman, disgustingly subservient and doesn't learn. I much prefer having interns and other juniors around that will also take time to correct but actually learn and grow from it.
As search engines I dislike them. When I ask for a subset of some data I want to be sure that the result is exhaustive without having to beg for it or make threats. Index and pattern matching can be understood, and come with guarantees that I don't just get some average or fleeting subset of a subset. If it's structured I can easily add another interactive filter that renders immediately. They're also too slow for the kind of non-exhaustive text search you might use e.g. Manticore or some vector database for, things like product recommendations where you only want fifteen results and it's fine if they're a little wonky.
Yeah doesn’t matter what you prefer. New hardware will boot strap models and eliminate the layers of syntax sugar devs use to write and ship software.
Hardware makers aren’t living some honorific quest to provide for SWEs. They see a path to claim more of the tech economy by eliminating as many SWE jobs as possible. They’re gonna try to capitalize on it.
Bwahaha. This is about as likely (in practice) as the whole ‘speech to text software means you’ll never need to type again’ fad.
Before you jump to conclusions, you should make a reasonable claim that IQ is still a reasonable measure for an individual's intellectual abilities in this context.
One could very much say that people's IQ is bound to decline if schooling decided to prioritize other skills.
You would also have to look into the impact of factors unrelated to the internet, like the evolution of schooling and its funding.
Pretty good chance that this is the impact of a generation of lead poisoned children growing up with stunted brains.
IQ scores may be declining, but it's far from certain that the thing they're trying to measure (g, or general intelligence) have actually declined.
https://open.substack.com/pub/cremieux/p/the-demise-of-the-f...
That's an article apparently from a white nationalist, Jordan Lasker, a collaborator of Emil Kirkegaard's. For a fun, mathematical take (by Cosma Shalizi) on what statistics tells us about "g" itself:
That's an article apparently from a white nationalist, Jordan Lasker, a collaborator of Emil Kirkegaard's.
http://bactra.org/weblog/523.html
The author objects to treating 'g' as a causal variable, because it doesn't help us understand how the mind works. He doesn't deny that 'g' is useful as a predictive variable.
I highly recommend reading the whole piece.
I will! Weekend starts soon!
The Borsboom and Glymour papers he links to are worth a skim too. It's a really dense (in a good way!) piece. Also shook up the way I think about other psych findings (the "big 5" in particular).