Join the discussion

Write your take first — we'll ask for email only when you're ready to publish.

  • Hacker News
  • Sorry, but introspection is not possible in LLMs. They don't have the hardware to have metaphysical dualism and therefore no mind to mentally be aware of.
  • > They don't have the hardware to have metaphysical dualism and therefore no mind to mentally be aware of.

    Somehow this sentence is self-contradictory, but I can't quite lay my finger on the exact spot.

  • I suspect the article is bullshit, but what hardware are you thinking of?
  • They’re just going to define terms so that eventually they discover consciousness and then play dumb long enough for us all to get screwed (see Jacob Tsimerman’s scenario 1 if you prefer your alarmism to come from a prestigious source).
  • Reminds me of that test where they had Grok, Gemini and Claude try to run a radio station.

    Stay in The Manifest !

  • >Submitted on 5 Jan 2026
  • Self-awareness has no direct connection to capabilities. It is entirely dependent upon the conditions that spawn emergent properties.
  • In thinking of directions where LLM's could develop from here, I cant help but think that a models ability to self introspect would immensely improve their utility. The R&D on how to achieve that is beyond me though. How do you train someone how to introspect? Also would it require a continuous learning architecture that doesn't separate training and inference?
  • Well, the algorithms in social media that are rewarding "engagement" (but accidentally ripping society apart by fueling rage) have made the training data — online discussions — a very poor training medium for introspection.

    So, if companies want training data on introspection, they would need to first fix what they broke (trust in each other and society) and have those discussions and realizations made public (by bringing back blogs where people might jot down their "maybe I've been an asshole" realizations) and rewarding other things than "engagement".

  • I think it’s possible this might just require reworking the transformer architecture a bit to give it an inherent awareness of the contents of its attention. Kind of a working memory workspace of sorts.

    I.e. Instead of limiting autoregression to the token level, you introduce a persistent compressed global workspace latent memory vector that is fed back into the self-attention mechanism at every layer or every token step, allowing the network to attend to its own prior attentional states before computing the next token. Obviously that’s going to involve some compression steps.

    Trouble is… I think the architecture there is much simpler a tweak than figuring out how to train it.

    …that’s likely to just destabilize training for not much if any gain at first. You’re probably gonna have to resort to some really clever (and currently missing) tricks to figure out how to train the network to actually use that feature.

  • Previously posted to Anthropic's blog in October. (Maybe not the same version?)

    https://www.anthropic.com/research/introspection

  • > Overall, our results indicate that current language models possess some functional introspective awareness of their own internal states. We stress that in today’s models, this capacity is highly unreliable and context-dependent; however, it may continue to develop with further improvements to model capabilities.
  • Not keen on the anthropomorphisation, it outputs text as we integrate - right at the end of the sausage machine, calling this awareness is like other parts where we use words like "thinking" and "reasoning".
  • If it's context dependent is it really introspection?