One more question @balazskegl
: if the LLM was trained to optimise a reward, then it could be argued that it has agency. This is what happens in LLM post training. I guess the counter argument would be that such rewards are external. Yet, the Schrodinger argument is that the
ETHICS
-
LLM Agency and Reward Optimization: External vs Internal Motivation
By
–
-
Do LLMs possess intrinsic intentional stance like humans
By
–
So I guess my question is: Is your position that humans had an intrinsic intentional stance but that LLMs don't?
-
LLM Intentionality and the Herding Method for Neural Networks
By
–
Thanks Balazs. I posted the herding method. It's certainly pretty good for normalised distributions. LLM folks should probably be trying it. But I agree your argument goes beyond this. It is about whether the net has an intentional stance. Daniel Dennett offered some thoughts
-
India AI Safety Institute Opens Call Strategic Collaborations
By
–
Call for Strategic Collaborations | IndiaAI Safety Institute (AISI) .
@GoI_MeitY under the IndiaAI Mission, invites academic and research institutions, startups, industry leaders, legal entities, autonomous bodies, and government agencies to submit proposals under the Call for -
How LLMs Organize Concepts Differently from Human Brains
By
–
Do LLMs Think Like Humans? New Study Says Not Quite This paper dives into a fascinating question:
Do Large Language Models organize concepts the way humans do?
Spoiler: They don’t—at least not yet. Humans use semantic compression—they categorize by balancing expressive -
Preserving dignity while achieving professional success
By
–
White pill: this is a false dichotomy. You don’t have to be either. You can get through life with both your dignity preserved and being able to accomplish a lot.
-
Herding Algorithms and Human Intelligence in AI Systems
By
–
One more argument: We could easily replace the sampling of the multinomial by a dynamical system, eg by herding https://
icml.cc/Conferences/20
09/papers/447.pdf
… I’m not sure randomness is needed. More on the opinion side of things: I don’t think humans are very far from being parrots. We may yet -
Language as Tool for Communication and Self-Reflection
By
–
Digesting these arguments with great interest. Thanks everyone. Here are some quick sketchy reflections to expand our debate: 1. Language is a tool for communication but it’s more than that. We turned the tool on ourselves to self reflect and think. Of course we still need
-
Research Impact and Scientific Communication Beyond Disciplines
By
–
"Broader impact" means "what would be the impact of your research, beyond the local neighborhood of your academic discipline." A perfectly reasonable question to ask.
"Outreach" means "will you be able to communicate your research beyond your circle of expert colleagues"
Nothing -
Jailbroken AI Models and Cybersecurity Threats to Online Services
By
–
I worry about agentic tool use. If you manage to jailbreak a mainstream model you can already get it to tell you exactly how to break 90% of the world’s websites/online services (esp. govt services) but most people lack the agency and energy to actually do it and debug issues