I know this isn’t the point of the article, but honestly we shouldn’t be using the term “AI” either. By implying that ChatGPT et al are intelligent, even if that’s not our intention, we feed into the hype. I found this article proposing alternative, more accurate terms: https://buttondown.com/maiht3k/archive/how-to-talk-about-ai-without-adding-to-the/
The tech bros have destroyed the believe everyone had about progress. These days not much people believe more inventions and technology will actually improve our lives.
Yeah, I hate how assholes like donnie are wanting to cash in. So very very much. I remain hopeful about its practical applications in many fields - and they are there, in spite of what many haters say. I hate that the focus seems to be about how many jobs we can “disrupt” or destroy based on stolen works (code, art, books). I’m so very sick of what it’s doing to the IT world in which I work. Instead, it could have been mostly about trying to supercharge scientific and technology endeavors, and mostly that. Instead, the oligarchs seem to be aiming it at stealing the jobs of creatives.
We desperately need someone smart and informed on the matter with a big platform that can split the difference. People like donnie and musk ain’t it but because one is in the WH and the other has too much money, we get to listen to their “thoughts” on the matter.
The first thing I broach on the topic is that there is no intelligence. It is a language model that uses an algorithm to determine speech and conversation. It is limited to whatever it has scraped from humanity and offers nothing further. Each different LLM has its own predilections and tendencies. Some better than others, but all of it is unintelligent and will confidently tell you a wrong answer. Furthermore, it devolves into hallucinations very quickly. What have I missed or what should I add?
Edit: Article correctly states LLMs should not be anthropomorphized and “hallucination” lends to that. But I feel “undesirable outputs” is too soft even though it does strip the humanity from the LLM.
I think it might make sense to call it “Augmentive Intelligence”, because it is not intelligence on its own. It has major issues when it comes to reasoning on its own, but for someone who actually is intelligent, it can add to that person’s ability to accomplish things. So it augments that person’s intelligence, instead of being an intelligence on its own.
I know this isn’t the point of the article, but honestly we shouldn’t be using the term “AI” either. By implying that ChatGPT et al are intelligent, even if that’s not our intention, we feed into the hype. I found this article proposing alternative, more accurate terms: https://buttondown.com/maiht3k/archive/how-to-talk-about-ai-without-adding-to-the/
Probabilistic Automation is probably the best term for a majority of what the orange buffoon calls “super intelligence”.
I hate that this went from “we want to detect cancer early” to “steal and monetize everything”.
The tech bros have destroyed the believe everyone had about progress. These days not much people believe more inventions and technology will actually improve our lives.
Yeah, I hate how assholes like donnie are wanting to cash in. So very very much. I remain hopeful about its practical applications in many fields - and they are there, in spite of what many haters say. I hate that the focus seems to be about how many jobs we can “disrupt” or destroy based on stolen works (code, art, books). I’m so very sick of what it’s doing to the IT world in which I work. Instead, it could have been mostly about trying to supercharge scientific and technology endeavors, and mostly that. Instead, the oligarchs seem to be aiming it at stealing the jobs of creatives.
We desperately need someone smart and informed on the matter with a big platform that can split the difference. People like donnie and musk ain’t it but because one is in the WH and the other has too much money, we get to listen to their “thoughts” on the matter.
Good stuff
artificial intelligence → probabilistic automation
hybrid intelligence → augmented human intelligence
image recognition → image labeling
speech recognition → automatic transcription
the model shows bias → the model reflects bias
model mistakes → model errors
chatbots are good at … → chatbots are good for …
hallucination → undesirable output
goal → success condition
The first thing I broach on the topic is that there is no intelligence. It is a language model that uses an algorithm to determine speech and conversation. It is limited to whatever it has scraped from humanity and offers nothing further. Each different LLM has its own predilections and tendencies. Some better than others, but all of it is unintelligent and will confidently tell you a wrong answer. Furthermore, it devolves into hallucinations very quickly. What have I missed or what should I add?
Edit: Article correctly states LLMs should not be anthropomorphized and “hallucination” lends to that. But I feel “undesirable outputs” is too soft even though it does strip the humanity from the LLM.
I think it might make sense to call it “Augmentive Intelligence”, because it is not intelligence on its own. It has major issues when it comes to reasoning on its own, but for someone who actually is intelligent, it can add to that person’s ability to accomplish things. So it augments that person’s intelligence, instead of being an intelligence on its own.