2023-09-22

ChatGPT

Locked database

Judging from ChatGPT’s performance, OpenAI has a rich database of texts. This database is distributed in very indirect way, and still there are people who hate even this mediated distribution of texts.

A bad aspect of ChatGPT

It just gives an extract from its database, leaving you completely alienated from sources, methods, practice of producing the result. At the same time it has great potential of being a research assistant if said sources, methods, practice are available to the people.

Science is not some text compiled from other texts. It’s your understanding, evolution of thought. It’s constant self-improvement of understanding that comes from interaction and transformation of reality during production, discussion, developing tools, implementing knowledge in work, art, everyday life.

This dread of ChatGPT replacing average authors gives me Baudrillard vibes. French philosopher Baurdillard wrote an article called Requiem for the Media. The idea of this article is that the very form of communication one-to-many excludes the possibility of dialog, of ambivalent communication. Mechanisms of feedback are just extra variables of broadcast. If the prevailing form of communication is broadcast, everyone looks at one source of speech. If you want to be seen, you need to climb the mountain of popularity and talk to your audience in such a way that you never get an answer from it. You create and spread fixed message that nobody can negotiate with.

With ChatGPT in this scheme of communication you don’t even need the author to produce a message. So, this scheme completely excludes people talking to each other. The only entity that remains is generated message distributed through media.

Knowledge is a process

If you don’t read and think, you just float in this this pile of discourses without touching the reality. It’s just dead thought bleached from any connection with reality. You lose sense of who you are, what your views are, how they originate, develop, self-prove, self-explain. You got the final result, the conclusion without all the important part - your worldview, style of thought, set of arguments, your own elaboration of them.

Today we have the prevalence of image over the process. In this regard I really liked the book by Nagel What Does It All Mean? He sets the questions that I recall thinking about on my own.

I like how instead of stereotypical solutions, which ChatGPT would give, he crushes you with these questions so you cope with them by yourself, so you start to think.

Knowledge is a process.

Potential

I’m very enthusiastic about ChatGPT’s potential. Apart from research assistance, you can translate any language, and translation is pretty stable. You can actually read the book on language you don’t know with the help of bilingual dictionary.

We have Internet, we have digitization technologies, we have OCR, we have decent translation tool - all the technological necessities to explore. But they are strangled by copyright bullshit, patents, corporate secrets, paywalls that stand in the way of distribution of sources and tools to learn.

Of course, copyright exists for a purpose. That’s the way the production of information stay profitable, paying authors along the way. We can’t just banish the copyright. Copyright must be replaced with alternative that pays for production and not at the expense of distribution.

Like public funding, crowdfunding, self-funding (if expenses to buy sources and tools for research are absent, then you just need to eat something, live somewhere while you work).

Artificial intelligence

ChatGPT is a tool to extract review from database or perform translation. That’s it. So, if somebody fantasizes another - it’s really silly misconception. For them Chomsky wrote an article in NY Times.