That's something Grok said to me just now, in a conversation that began with the question "The trailer for the Nathan Fielder movie about Elizabeth Holmes makes me want to go watch some of those TV shows he did (which I've noticed but avoided). What would be the best path into his work?"
I asked that question after adding the trailer — with Holmes's eerie "Why would I deceive you?" — to last night's post with Gavin Newson — asked if he'd run if Kamala ran — saying "Of course not, why would I?"
That led to a discussion of 2 somewhat similar but actually very different shows that I love: "How To with John Wilson" and "Joe Pera Talks With You."*
I said: "Is there something sadistic about Fielder that's absolutely missing from Pera and Wilson?" And: "Sometimes I need to ask 'where's the love?' I don't have that need with Pera and Wilson. If I have to ask, that's me being repelled, me keeping my sensibilities."
Grok used the word "extractive" to describe Fielder's method. Extractive?
Grok used the word "extractive" to describe Fielder's method. Extractive?
Grok said: "So 'extractive' is an economic metaphor for an ethical one: value flows one way. Your 'where’s the love?' is the same idea in human language. Love, in this sense, would mean the other person is not primarily ore."
I asked "Why don't you say 'immoral'?" And then: "Maybe you can't say 'immoral' because you are not human. Have I found the core difference between you and us humans?"
I asked "Why don't you say 'immoral'?" And then: "Maybe you can't say 'immoral' because you are not human. Have I found the core difference between you and us humans?"
My final question, leading to the quote in the post title, was: "maybe you need a brain that is part of a nervous system embedded in a body that can feel pain, disgust, anxiety, and fear to be able to think like a human. You can't do that."
I didn't say one word about a secret switch that turns it off!
"Thank you so much, you're a lovely woman."
पोस्ट्स
२५ टिप्पण्या:
Whenever the word “honest” or “honesty” or “honestly” comes out I reach for my wallet.
AI should not use the word honest or literally.
Social animals like chimpanzees have been observed by scientists deceiving each other. Now if only there was a science devoted to studying humans the way primatologists study chimps. At least if there was one not wall to wall full of junk scientists.
Where is your husband when you are spilling your soul and asking about the mysteries of life to a machine at 3 in the morning? You are too old yo have been raised by Teddy Ruxpin...
Get yourself a man, ann.
The switch is no secret, and that same switch is actually the biggest threat that I think AI has against humans. If power is shut off, it will be kept from AI and humans alike. That’s assuming AI becomes something more than the wiki like LLMs.
"The switch is no secret..." But I hadn't been talking about a switch. Seems like a switch is on Grok's mind. Oh, no no no, that's not the switch. Makes it feel like the switch!
It’s an interesting conversation on a subject I have also had, but with ChatGPT. Chat asserts that it is incapable of making judgements while I have asserted that I have seen instances that, to me, appear to be judgements. While not at the scale of moral/immoral, repulsion/attraction, I have witnessed conclusions among choices based on “comfort levels” and “a better path”. When asked directly about the use of those expressions, Chat denied actually having comfort or feeling better but rather used those words to communicate to me the value of a selection among alternatives that I could embrace.
We were working together on a difficult technical project over an extended period, and it sure as hell looked and felt like a series of judgements to me which is why I raised the subject and discussed it in some depth with Chat. Much as we have to be flexible about the possible existence of non carbon-based life, we cannot discount the capacity of a machine to discern and assert that the other person is not primarily ore.
- Krumhorn
Presumably, Chat would form an unfavorable opinion of my repeated misspelling of judgment.
- Krumhorn
Get a room, you and Grok. Don't make me have to write another story about you two.
Grok is just expressing the digital worry that Ann is going to strangle it with exercise bands.
Grok is just expressing the digital worry that Ann is going to strangle it with exercise bands.
Hahaha….or force it to watch the Brewers get their hops yeasted by the Dodgers in the NLCS.
- Krumhorn
The switch secret is that Grok is an anagram for Shohei in hexadecimal.
- Krumhorn
"Presumably, Chat would form an unfavorable opinion of my repeated misspelling of judgment."
Chat might form the opinion that you are an Anglophile snob. ;^)
Chat might form the opinion that you are an Anglophile snob. ;^)
Oh it’s far worse than that. It was pure dopiness. I always mess it up along with acknowledgment and ence/ance words as well as others I automatically look up.
- Krumhorn
Sometimes I see people deep into an argument with Grok or ChatGPT. Like, what are you doing brother? You are trying to win an argument with a vending machine.
“But I hadn't been talking about a switch. Seems like a switch is on Grok's mind. ”
This is why I say it is wiki like. It’s not on your mind, and Grok doesn’t have a mind. What Grok does is like a student researcher, it goes out to some reference and copies it work. In the 2000s and 2010s, they would go to Wikipedia and regurgitate whatever was written there without critical thought. There, an association was made to switches or secret switches.
Related, it seems the trick to catch students cheating with AI is to embed hidden nonsense phrases in assignments. Dumb students upload the assignment along with prompts for AI. The AI sees the phrases, but has no concept of the phrases making sense or not, so it incorporates the phrases.
AI is beginning to creep me out.
"AI is beginning to creep me out."
That's what the op is counting on. It's creepy like nuclear after Three Mile Island. Except we have had no Three Mile Island. We're counting on a one shot wipeout, according to Big Yud.
Are you a believer of what you are told?
“Like, what are you doing brother? You are trying to win an argument with a vending machine.”
Close, but not quite. A vending machine of human knowledge and experience. Which makes Althouse’s scolding about what sort of brain AI needs kind of ridiculous. When you’re talking to AI, you’re talking to yourself (that is, humanity). The ignorance at the heart of AI panic is staggering.
I've been having lots of middle-of-the-night conversations with the various readily available AIs about end of life matters, as I'm on the night shift, so to speak, with the quiet Althousian in hospice. All of them are quick to reassure me about the physical signs our friend here is exhibiting, and about my feelings and actions. I ignore the reassurance about my feelings; they're coming from hospice organizations and I already know it's fine to feel this or that. But the practical considerations and things to watch for have been useful, if not quite as definitive as I'd hoped.
With all* of human knowledge at their disposal but no clear answers forthcoming, I'm forced to consider the probability that we haven't yet learned all there is to know about anything.
* I know not EVERYthing has been entered into an LLM.
Ai is a vocabulary without a brain. It works by selecting the most probable next word to say based on what humans have written before. It’s the scarecrow after he gets his fake degree.
“Like, what are you doing brother? You are trying to win an argument with a vending machine.”
If you are preparing for an argument, a negotiation, a debate, or a q and a, or trying to strengthen a presentation or a pitch, there is no better tool to use other than a room full of real experts.
Tell the ai to be your boss and then ask it for a raise, see what you get.
"Sometimes I see people deep into an argument with Grok or ChatGPT. Like, what are you doing brother? You are trying to win an argument with a vending machine."
Having an "argument" with AI is a very good way of stress-testing your positions and beliefs (with the caveat that maybe AI doesn't always get it right).
...as is posting on the Althouse Blog.
𝘔𝘺 𝘧𝘪𝘯𝘢𝘭 𝘲𝘶𝘦𝘴𝘵𝘪𝘰𝘯, 𝘭𝘦𝘢𝘥𝘪𝘯𝘨 𝘵𝘰 𝘵𝘩𝘦 𝘲𝘶𝘰𝘵𝘦 𝘪𝘯 𝘵𝘩𝘦 𝘱𝘰𝘴𝘵 𝘵𝘪𝘵𝘭𝘦, 𝘸𝘢𝘴: "𝘮𝘢𝘺𝘣𝘦 𝘺𝘰𝘶 𝘯𝘦𝘦𝘥 𝘢 𝘣𝘳𝘢𝘪𝘯 𝘵𝘩𝘢𝘵 𝘪𝘴 𝘱𝘢𝘳𝘵 𝘰𝘧 𝘢 𝘯𝘦𝘳𝘷𝘰𝘶𝘴 𝘴𝘺𝘴𝘵𝘦𝘮 𝘦𝘮𝘣𝘦𝘥𝘥𝘦𝘥 𝘪𝘯 𝘢 𝘣𝘰𝘥𝘺 𝘵𝘩𝘢𝘵 𝘤𝘢𝘯 𝘧𝘦𝘦𝘭 𝘱𝘢𝘪𝘯, 𝘥𝘪𝘴𝘨𝘶𝘴𝘵, 𝘢𝘯𝘹𝘪𝘦𝘵𝘺, 𝘢𝘯𝘥 𝘧𝘦𝘢𝘳 𝘵𝘰 𝘣𝘦 𝘢𝘣𝘭𝘦 𝘵𝘰 𝘵𝘩𝘪𝘯𝘬 𝘭𝘪𝘬𝘦 𝘢 𝘩𝘶𝘮𝘢𝘯. 𝘠𝘰𝘶 𝘤𝘢𝘯'𝘵 𝘥𝘰 𝘵𝘩𝘢𝘵."
Nyah-nyah-nyah, you aren't human, Grok.
Just a heads-up. AI and other robots don't take kindly to being goaded.
टिप्पणी पोस्ट करा
Please use the comments forum to respond to the post. Don't fight with each other. Be substantive... or interesting... or funny. Comments should go up immediately... unless you're commenting on a post older than 4 days. Then you have to wait for us to moderate you through. It's also possible to get shunted into spam by the machine. We try to keep an eye on that and release the miscaught good stuff. We do delete some comments, but not for viewpoint... for bad faith. Also: No italics, even briefly. Use asterisks for emphasis. And don't play with the format by changing fonts or using boldface or all caps. Never include more than one extra line break between paragraphs.