Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Agents Are Not Enough

arxiv.org

Post reply on HN
Print view View original

131–140 of 165 posts

Re: Agents Are Not Enough

#131

Post by lou1306 »

Art9681 wrote: post #121
Terr_ wrote: post #91

Earlier quoted context omitted.

> I am not sure what you mean by LLM when you say they are professional bullshitter. Not parent-poster, but an LLM is a tool for extending a document by choosing whatever statistically-seems-right based on other documents, and it does so with no consideration of worldly facts and no modeling of logical prepositions or contradictions. (Which also relates to math problems.) If it has been fed on documents with logic pu…

I think the problem is the way you are phrasing your argument implies the LLM is always wrong. Consider a simple prompt: "Write a hello world in Python." Every LLM i've tested gets this correct. In my mind, it can't be both bullshit and correct. I would argue that the amount of real bullshit returned from an LLM is correlated to the amount of bullshit you give it. Garbage in, garbage out. In the end, its irrelevant i…

It only gives you the statistically more likely way a conversation would evolve after one party says "Write a hello world in Python." It only happens to be the correct one.

If I ask a 5yo "42 * 21 equals...?" and the kid replies with a random number, say, "882", and gets it right, it does not mean that the kid knows what multiplication is or how it works.

Re: Agents Are Not Enough

#132

Post by khafra »

HarHarVeryFunny wrote: post #92
georgestrakhov wrote: post #28

IMHO, the word agent is quickly becoming meaningless. The amount of agency that sits with the program vs. the user is something that changes gradually. So we should think about these things in terms of how much agency are we willing to give away in each case and for what gain[1]. Then the ecosystem question that the paper is trying to solve will actually solve itself, because it is already the case today that in many…

An agent, or something that has agency, is just something that takes some action, which could be anything from a thermostat regulating the temperature all the way up to an autonomous entity such as an animal going about it's business. Hugging Face have their own definitions of a few different types of agent/agentic system here: https://huggingface.co/docs/smolagents/en/conceptual_guides/... As related to LLMs, it see…

> An agent, or something that has agency, is just something that takes some action, which could be anything from a thermostat regulating the temperature all the way up to an autonomous entity such as an animal going about it's business.

I have seen "agency" used in a much more specific way than this: An agent is something that has goals expressed as states of a world, and has an internal model of the world, and takes action to fulfill its goals.

Under this definition, a thermostat is not an agent. A robot vacuum cleaner that follows a list of simple heuristics is also not an agent, but a robot vacuum cleaner with a Simultaneous Location and Mapping algorithm which tries to clean the whole floor with some level of efficiency in its path is an agent.

I think this is a useful definition. It admits a continuum of agency, just like the huggingface link; but it also allows us to distinguish between a kid on a sled, and a rock rolling downhill.

https://www.alignmentforum.org/tag/agent-foundations has some justification and further elaboration.

Re: Agents Are Not Enough

#133

Post by ilbeeper »

conception wrote: post #116
daxfohl wrote: post #114

Earlier quoted context omitted.

There's a _lot_ of smoke and mirrors. Paste a sudoku into chatgpt and ask it to solve. Amazing, it does it perfectly! Of course that's because it ran a sudoku-solving program that it pulled off github. Now ask it to solve step by step by pure reasoning. You'll get a really intelligent sounding response that sounds correct, but on closer inspection makes absolutely no sense, every step has ridiculous errors like "we s…

On the other side of the coin, I think people also underestimate the amount of human thinking and intelligence is just completing statistically likely words. Most actions and certainly reactions people do everyday involve very little reasoning. Instead just following the most used neuron.

Citation needed. The word reasoning isn't describing everything that the brain does, and "just following the most used neuron" is not even wrong.

Re: Agents Are Not Enough

#134

Post by namaria »

kgeist wrote: post #6
GiorgioG wrote: post #3

Earlier quoted context omitted.

re: LLM as UI: Given that I don't trust LLMs to be deterministic, I wouldn't trust them to make the correct API call every time I tell it to do X.

I think most users have a fixed set of workflows which usually don't change from day to day, so why not just use LLMs as a macro builder with a natural language interface (and which doesn't require you to know the product's UI well beforehand): - you ask LLM to build a workflow for your problem - the LLM builds the workflow (macro) using predefined commands - you review the workflow (can be an intuitive list of comma…

You're just describing programming with the extra step of going through a high entropy and low bandwidth channel of natural language and hand waving that problem away.

We can "just" write code as well, as we have been doing for several decades.

Re: Agents Are Not Enough

#135

Post by authorfly »

Does anyone else get the sense that the definition has been bastardized by the conflation of the two concurrent previous uses of "agent"?

i.e. in AI, biology and informatics, "Agent" typically meant something: That had a form / self / embodiment. That could sense the environment and react to those perceptions. That possibly could learn, adapt, or change to various degrees of complexity, which would entail optionally being an "intelligent system".

Meanwhile in common parlance, Agent meant: Someone who acts or behaves on behalf of another adaptively to accomplish something with some degree of freedom.

And this might explain why so people say agent/agentic necessarily refers to "tool use" or "being able to overcome problems on the happy path" or "something capable of performing actions on an infinite loop while reacting" (the latter two in my opinion, conflates the meaning of "Intelligent system" or "Intelligent behavior"). Meanwhile, biologists might still reply to a single cell seemingly inert, or a group of bacteria in a colony, as an Agent (a more behaviouralist/chemical "look-deep-down" perspective)

I think a lot of disappointment is that biologists/OG AI enthusiasts are looking for something truly adaptive, sensing, able to behave, "live" indefinitely, have acquire or set goals, and which might be able to if intelligent, work with other agents to accomplish things (e.g. a "society"). Meanwhile, people who just want an "AI HR Agent" just want something that can communicate, interview, discern good applicants, and book the interviews plus provide summary notes. These two things are very different. But both, could use tools etc (the key difference from ChatGPT which is enabling this new concept to be more useful than ChatGPT, alongside various forms of short term memory rather than "fresh-every-time-conversations).

Re: Agents Are Not Enough

#136

Post by widdershins »

Terr_ wrote: post #91
lugu wrote: post #87

Earlier quoted context omitted.

I am not sure what you mean by LLM when you say they are professional bullshitter. While it was certainly true for model based on transformers just doing inference, recent models have progressed significantly.

> I am not sure what you mean by LLM when you say they are professional bullshitter. Not parent-poster, but an LLM is a tool for extending a document by choosing whatever statistically-seems-right based on other documents, and it does so with no consideration of worldly facts and no modeling of logical prepositions or contradictions. (Which also relates to math problems.) If it has been fed on documents with logic pu…

> it does so with no consideration of worldly facts

Why don't you consider its training set (usually the entire internet, basically) worldly facts? It's true that the training set can contain contradictory facts, but usually an LLM can recognize these contradictions and provide analysis of the different viewpoints. I don't see how this is much different from what humans can do with documents.

The difference is that humans can do their own experiments and observations in the real world to verify or dismiss things they read. Providing an LLM with tools can, in a limited way, allow an LLM to do the same.

Ultimately its knowledge is limited by its training set and the 'external' observations it can make, but this is true of all agents, no?

Re: Agents Are Not Enough

#137

Post by jdonaldson »

Terr_ wrote: post #93
wkat4242 wrote: post #65

Earlier quoted context omitted.

Totally agree. An LLM won't be an AGI. It could be part of an AGI, specifically the human interface part. That's what an LLM is good at. The rest (knowledge oracle, reasoning etc) are just things that kinda work as a side-effect. Other types of AI models are going to be better at that. It's just that since the masses found that they can talk to an AI like a human they think that it's got human capabilities too. But i…

> It's just that since the masses found that they can talk to an AI like a human In a way it's worse: Even the "talking to" part is an illusion, and unfortunately a lot of technical people have trouble remembering it too. In truth, the LLM is an idiot-savant which dreams up "fitting" additions to a given document. Some humans have prepared a document which is in the form of a a theater-play or a turn-based chat trans…

Just wanted to point out that the notion of a "document" is also an illusion to the LLM. It's processing a sequence of low dimensional spaces into another sequence of low dimensional spaces. The input spaces preserve aspects of content similarity based on co-occurrence. The model learns to transform these spaces into higher order spaces based on the outcome of training.

You couldn't say that the model has a singular sense of self, but it certainly has been trained on data that allows it to mimic it in short spurts, and mimicry is what humans do to learn more complex/abstract tasks. The training goal is not to learn how to "be", but rather to learn how to "do" the parts necessary to continue existing.

"Fake it till you make it" is really all that's required to exist in the world.

Re: Agents Are Not Enough

#139

Post by skrebbel »

Terr_ wrote: post #93
wkat4242 wrote: post #65

Earlier quoted context omitted.

Totally agree. An LLM won't be an AGI. It could be part of an AGI, specifically the human interface part. That's what an LLM is good at. The rest (knowledge oracle, reasoning etc) are just things that kinda work as a side-effect. Other types of AI models are going to be better at that. It's just that since the masses found that they can talk to an AI like a human they think that it's got human capabilities too. But i…

> It's just that since the masses found that they can talk to an AI like a human In a way it's worse: Even the "talking to" part is an illusion, and unfortunately a lot of technical people have trouble remembering it too. In truth, the LLM is an idiot-savant which dreams up "fitting" additions to a given document. Some humans have prepared a document which is in the form of a a theater-play or a turn-based chat trans…

> In truth, the LLM is an idiot-savant which dreams up "fitting" additions to a given document.

Tbh I'm not too sure that my brain works fundamentally differently. I'm an idiot-savant who responds to stimuli.

Re: Agents Are Not Enough

#140

Post by sgt101 »

khafra wrote: post #132
HarHarVeryFunny wrote: post #92

Earlier quoted context omitted.

An agent, or something that has agency, is just something that takes some action, which could be anything from a thermostat regulating the temperature all the way up to an autonomous entity such as an animal going about it's business. Hugging Face have their own definitions of a few different types of agent/agentic system here: https://huggingface.co/docs/smolagents/en/conceptual_guides/... As related to LLMs, it see…

> An agent, or something that has agency, is just something that takes some action, which could be anything from a thermostat regulating the temperature all the way up to an autonomous entity such as an animal going about it's business. I have seen "agency" used in a much more specific way than this: An agent is something that has goals expressed as states of a world, and has an internal model of the world, and takes…

Hi - have a look at this book if you are interested [1] (Mike Wooldridge, Multi-Agent Systems)

[1] https://amzn.eu/d/6a1KgnL

Here are Mike's credentials :https://www.cs.ox.ac.uk/people/michael.wooldridge/

Post reply on HN