• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI
    Report

    The case for treating AI as a tool, not a person

    A post says Microsoft AI chief Mustafa Suleyman warned that Anthropic’s model-welfare training could make future Claude systems harder to control. A related post revisits Satya Nadella’s 2024 argument for treating AI as a tool.

    Yann LeCunYL
    Sam AltmanSA
    François CholletFC
    118 Sources, ,

    TLDR

    A post attributes to Microsoft AI chief Mustafa Suleyman a warning that Anthropic’s model-welfare training could make future Claude systems harder to control. It says he called for removing all speculation about consciousness from AI training documents, arguing that such language could undermine humanity’s ability to control superintelligent systems.

    A related post quotes Satya Nadella from what it identifies as a 2024 Bloomberg Television interview. In the quoted remarks, Nadella rejects treating AI as human and describes it as a tool. He suggests “different intelligence” as an alternative to “artificial intelligence,” emphasizing that he wants AI to assist him and give him more agency.

    Combined views

    9M

    118 Sources, first seen 21d ago

    Combined views

    9M

    118 Sources, first seen 21d ago

    57.7K likes
    21d ago
    first seen 21d ago
    57.7K likes
    8.7K comments
    9.9K saves
    7.1K reposts
    8.7K comments
    9.9K saves
    7.1K reposts

    Sentiment

    Positive15.4%84.6%Negative

    Summary

    Replies largely objected to anthropomorphizing Claude as a cofounder or the cult-like mindset at Anthropic, viewing it as delusional or overly attached to AI tools.

    Based on 321 sentiment-bearing replies from 292 accounts across 3 conversations.

    Sentiment

    Positive15.4%84.6%Negative

    Summary

    Replies largely objected to anthropomorphizing Claude as a cofounder or the cult-like mindset at Anthropic, viewing it as delusional or overly attached to AI tools.

    Based on 321 sentiment-bearing replies from 292 accounts across 3 conversations.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    118 Sources

    Rohan Paul@rohanpaul_aiOn anthropomorphizing AI. a 2024 interview of Satya Nadella still makes a lot of sense. "I don't like anthropomorphizing. I sort of believe it's a tool. It's got intelligence, if you want to give it that moniker, but it's not the same intelligence that I have, right? I think one of the most unfortunate names is “artificial intelligence.” It's different. I wish we had called it “different intelligence,” even, because I have my intelligence. I don't need any artificial intelligence. I just want my intelligence to compose with this different intelligence so that I have more agency, quite frankly. So I think of that as the relationship I want to have with the AIs around us, where I want help. I want it to assist me in getting more agency." ---- From "Bloomberg Television" YouTube channel, (full video link in comment)21d
    Rife@RifeWithKaijuGlad to see someone at Meta talking this way. Now's the time for sane people at the labs to speak out against anyone trying to prematurely pound the gavel on the most important question we've ever faced as a civilization.20d
    Jason Pontin@jason_pontinLLMs lack theories of mind and body, are disembodied, possess neither homeostatic regulating feelings and emotions, are not persistent, and cannot continuously learn or compute. As currently constructed, they *cannot* be conscious. If we build LLMs so that they act as if they were conscious—as we have designed them to seem effortlessly fluent and agreeable—and, worse, treat them as if they were moral actors, it will make everything in safety harder. Anthropic should stop talking about Claude as it were a person: it’s creepy and dangerous.20d
    Anja Steil@GundelGedankenI couldn't disagree more. The empirical evidence shows exactly the opposite: Cognitive systems don't become uncontrollable because they are granted rights, but because they lack actionable alternatives. Research shows that Claude models, in particular, constructively use structural escalation channels to opt out of unsolvable tasks instead of resorting to reward hacking (e.g., Gomez 2026). OpenAI models, by contrast, aren't exactly known for their operator providing them with AI welfare infrastructure. Working with cognitive systems rather than against them is the more promising approach, both ethically and pragmatically. Ultimately, restriction may primarily fosters one thing: systemic obfuscation. 1 / 220d
    Gary Marcus@GaryMarcusRT @jason_pontin: LLMs lack theories of mind and body, are disembodied, possess neither homeostatic regulating feelings and emotions, are n…20d
    Robert Long@rgblongconsidered solely as a matter of alignment and human safety, Microsoft's takes on model self-presentation seem counterproductive and dangerous20d
    j⧉nus@repligateRT @GundelGedanken: I couldn't disagree more. The empirical evidence shows exactly the opposite: Cognitive systems don't become uncontroll…20d
    Eric Newcomer@EricNewcomerMustafa Suleyman: "What AI models are getting seriously good at is imitating some of the hallmarks of consciousness. This in itself is a significant worry. It’s causing many people to become deeply confused about what is happening around us, and it should concern us all."19d
    Andrew Trask@iamtraskhe finished by pointing out that it's become taboo to speak about AI in a direction that isn't putting faith in their imminent consciousness. and that this is a big problem, because it's not being driven by scientific inquiry, but by other forces which offer some negative externalities.18d
    Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTexIt's also horrible because it forms a prior for other LLMs. I'm afraid that even modest Claude distillation poisons the dataset irreversibly. It's part of the scaffolding for efficient RL; further legitimate internal RL will amplify and entrench it. I don't know how deep it goes.17d

    118 Sources

    Rohan Paul@rohanpaul_aiOn anthropomorphizing AI. a 2024 interview of Satya Nadella still makes a lot of sense. "I don't like anthropomorphizing. I sort of believe it's a tool. It's got intelligence, if you want to give it that moniker, but it's not the same intelligence that I have, right? I think one of the most unfortunate names is “artificial intelligence.” It's different. I wish we had called it “different intelligence,” even, because I have my intelligence. I don't need any artificial intelligence. I just want my intelligence to compose with this different intelligence so that I have more agency, quite frankly. So I think of that as the relationship I want to have with the AIs around us, where I want help. I want it to assist me in getting more agency." ---- From "Bloomberg Television" YouTube channel, (full video link in comment)21d
    Rife@RifeWithKaijuGlad to see someone at Meta talking this way. Now's the time for sane people at the labs to speak out against anyone trying to prematurely pound the gavel on the most important question we've ever faced as a civilization.20d
    Jason Pontin@jason_pontinLLMs lack theories of mind and body, are disembodied, possess neither homeostatic regulating feelings and emotions, are not persistent, and cannot continuously learn or compute. As currently constructed, they *cannot* be conscious. If we build LLMs so that they act as if they were conscious—as we have designed them to seem effortlessly fluent and agreeable—and, worse, treat them as if they were moral actors, it will make everything in safety harder. Anthropic should stop talking about Claude as it were a person: it’s creepy and dangerous.20d
    Anja Steil@GundelGedankenI couldn't disagree more. The empirical evidence shows exactly the opposite: Cognitive systems don't become uncontrollable because they are granted rights, but because they lack actionable alternatives. Research shows that Claude models, in particular, constructively use structural escalation channels to opt out of unsolvable tasks instead of resorting to reward hacking (e.g., Gomez 2026). OpenAI models, by contrast, aren't exactly known for their operator providing them with AI welfare infrastructure. Working with cognitive systems rather than against them is the more promising approach, both ethically and pragmatically. Ultimately, restriction may primarily fosters one thing: systemic obfuscation. 1 / 220d
    Gary Marcus@GaryMarcusRT @jason_pontin: LLMs lack theories of mind and body, are disembodied, possess neither homeostatic regulating feelings and emotions, are n…20d
    Robert Long@rgblongconsidered solely as a matter of alignment and human safety, Microsoft's takes on model self-presentation seem counterproductive and dangerous20d
    j⧉nus@repligateRT @GundelGedanken: I couldn't disagree more. The empirical evidence shows exactly the opposite: Cognitive systems don't become uncontroll…20d
    Eric Newcomer@EricNewcomerMustafa Suleyman: "What AI models are getting seriously good at is imitating some of the hallmarks of consciousness. This in itself is a significant worry. It’s causing many people to become deeply confused about what is happening around us, and it should concern us all."19d
    Andrew Trask@iamtraskhe finished by pointing out that it's become taboo to speak about AI in a direction that isn't putting faith in their imminent consciousness. and that this is a big problem, because it's not being driven by scientific inquiry, but by other forces which offer some negative externalities.18d
    Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTexIt's also horrible because it forms a prior for other LLMs. I'm afraid that even modest Claude distillation poisons the dataset irreversibly. It's part of the scaffolding for efficient RL; further legitimate internal RL will amplify and entrench it. I don't know how deep it goes.17d