Brief Reflections On "Artificial Intelligence"

Because of my limited knowledge, I have chosen to present these as disconnected notes rather than a more organized essay. However, I have not seen many of these thoughts before, so I hope this short note adds something new to the ongoing discourse.

  1. It’s not, really. Not intelligence, anyway. It doesn’t know truth from falsehood, or right from wrong.
  2. The technologies that are called artificial intelligence, (diffusion models and large language models), are basically very large grammars. They seem to replicate part of brain visual and speech centers, but no other neurological functions.
  3. In brain damaged people, there is a thing called confabulation. Confabulation is what happens when a damaged brain reaches for a memory and finds it’s not there. It just fills in the gaps. And I think that’s pretty much what an LLM does; it seems to be a replication of part of a brain but it has no real memory or logical capacity or ethics. A human author knows to fact check and not to plagiarize; a human artist knows to not to copy too closely. LLMs and diffusion models know neither of these things – only what word or line comes next. They are not minds, but only parts of minds.
  4. “Mundus vult decipi, ergo decipiatur.” People want to believe these are intelligent. They want to believe so badly that they gloss over the flaws. This would not be a problem if the developers of LLMs and diffusion models did not overclaim so broadly, but of course they do – they want to sell them, or at least their output.
  5. At the same time, they give a hint of what real AI might be capable of. LLMs know only language, and that only from typewritten text, but how very much of it they have taken as input! Few humans are as literate.
  6. These technologies open up philosophical topics. Some philosophers have long argued that a part of human cognition is not, in fact, conscious at all, but simply automatic, and the successful construction of these models seems to argue for that view. One also, when one encounters a skillful orator who seems to have no ethics or connections with reality, now has to ask if this is in fact the result of thought, or simply putting together words in pleasing patterns. Uncomfortably, one may also ask how much apparently rational oration is in fact mindless. And, finally, it is now reasonable to speculate that such processes might necessarily be parts of whatever mind might be – that aliens might also have similar cognitive processes.

Comments

Post a Comment

Popular posts from this blog

The General Intelligence of Robots

“Artificial intelligence” researchers mean by the phrase “general intelligence” two things. One is simply a machine learning system that can, like humans, use its knowledge for many different tasks. This is not something that has currently been built; language models produce language, stable diffusion models produce images, and neither seem to have any concepts of the underlying realities these words and images reflect. The second is the “general intelligence,” g , derived from intelligence tests, which is believed by some to be a unifying reality behind intelligence test scores. In a paper on GPT-4 from Microsoft a Wall Street Journal Editorial by Linda Gottfredson, Mainstream Science on Intelligence is cited as providing a definition of general intelligence. It’s an appalling piece, repeating debunked claims about racial differences in intelligence and claiming as scientific consensus hypotheses that are at best debatable and at worst outright false.
Read more

A Grand Unified Theory of Bad New Economy Firms

The wildly successful internet economy firms mostly operate by evading laws. New technology which has allowed firms to move their business to less-regulated jurisdictions and a generation of deregulation and favorable court decisions have allowed these firms to be bad places to work and bad places to do business with. Here’s a catalog of some of their abuses: Evasion of local laws through national operations . Amazon notoriously picks jurisdictions with lax labor laws for its warehouses. Uber simply ignores local laws until penalized. When penalized, they initiate political campaigns to get the laws overridden at the state level. Extortionate business practices . Doordash infamously underpays its workers. Grubhub notoriously overcharges restaurants . Uber has the resurrected long-illegal practice of raising taxi rates when demand is highest. Monopolistic practices . Enforcement in this
Read more

Richard M. Stallman and the Failure of the Free Software Foundation

(Note: since the writing of the following, the Free Software Foundation has tweeted: “The FSF board unanimously condemns misogyny, racism, and other bigotry as well as defamation, intimidation, and unfair attacks on free thought and speech.” They are rejecting the criticism of Stallman as unfair and, in so doing, they aligned themselves with neo-Nazis and white supremacists. They have put themselves beyond redemption and any fair hearing for their ideas. No-one wants to listen to the logic-chopping of rape apologists – it is as bad as listening to Southern racists excusing slavery.)   This is outside of my usual ambit, and it is about cooperative software development rather than market economics, but it seems to belong here more than anywhere. People who follow free and open source software by now know that alleged harasser and rape apologist Richard M. Stallman has, by a secret process, been restored to the board of directors of the Free Software Foundation. Stallman has lon
Read more