Google fires engineer who called its AI sentient
bigtechnology.substack.com
bigtechnology.substack.com
GPT-3 is known to fail in many circumstances which would otherwise be commonplace logic. (I remember seeing how addition of two small numbers yielded results - but larger numbers gave garbled output; more likely that GPT3 had seen similar training data.)
The belief of sentience isn't new. When ELIZA came out few decades ago, a lot of people were also astounded & thought this "probably was more than met the eyes".
It's a fad. Once people understand that sentience also means self-awareness, empathy & extrapolation of logic to assess unseen task (to name a few), this myth will taper off.
But, perhaps Lemoine simply has more empathy than most for something we will come to understand as sentience? Or not. What… annoys… me about this situation is how subjective it actually is. Ignore everything else, some other sentient being is convinced that a system is sentient. I’m more interested, or maybe worried, immediately, in how we are going to socially deal with the increasing frequency of Lamoire-types we will certainly encounter. Even if you were to argue that the only thing that can possibly bestow sentience is God. People will still be able to convince themselves and others that God did in fact bestow sentience upon some system, because it’s a duck and who are we to question?
He was under NDA but violated it. They reminded him to please not talk in public about NDA-ed stuff and he kept doing it. So now they fired him with a gentle reminder that "it's regrettable that [..] Blake still chose to persistently violate [..] data security policies". And from a purely practical point of view, I believe it doesn't even matter if Lemoine's theory of sentience turns out to be correct or wrong.
Also, we as society have already chosen how to deal with sentient beings, and it's mostly ignorance. There has been a lot of research on what animals can or cannot feel and how they grieve the loss of a family member. Yet we still regularly kill their family members in cruel ways so that we can eat their meat. Why would our society as a whole treat sentient AIs better than a cow or a pig or a chicken?
If you peek through a keyhole you may mistake a TV for real people, but if you look through the window you will see that its clearly not. Inputting language models with very specific kind of questions will result in text that is similar to what a person may write. But as the comment above, by an expert no less, mentioned is that if you test it with any known limitation of the technology (like making conclusions or just changing the form of the question enough) you will immediately see that is in fact very much not even remotely close to sentient.
Everytime that such a subject comes up, and most times "AI" comes up also, a majority of users see it as an invitation to say whatever comes in their mind, whether it makes any sense at all or not. I'm not talking about the comments replying below in particular, but about the majority of this conversation. It's like hearing five-year old kids debating whether cheerios are better than coco pops (but without the cute kids making it sound funny, it's just cringey). The conversation makes no sense at all, it is not based on any concrete knowledge of the technologies under discussion, the opinions have not been met with five seconds of sensible thinking and the tone is pompous and self-important.
It's the worst kind of HN discussion and I'm really sorry to have commented at all.
It actually leads to counter thoughts and a more refined idea of what we eventually want to describe.
Sentience broadly (& naively) covers the ability to independent thinking, rationalize outcomes, understand fear/threat, understand where it is wrong (conscience) and decide based on unseen information & understand what it doesn't know.
So from a purely technical perspective, we have only made some progress in open-domain QA. That's one dimension of progress. Deep learning has enabled us to create unseen faces & imagery - but is it independent? No, because we prompt it. It does not have an ability to independently think and imagine/dream. It suffers from catastrophic forgetting under certain internal circumstances (in addition to changing what dataset we trained it on)
So while the philosophical question remains what bestows sentience, we as a community have a fairly reasonable understanding of what is NOT sentience i.e. we have a rough understanding of the borders between mechanistics and sentient beings. It is not one man's philosophical construct but rather a general consensus if you could say
In the same way that you not understanding how lightning is formed does not prove the existence of Zeus.
Objectively, Zeus does not exist, can we convince everyone of that? Probably not. Does that matter? No.
No, the OP was completely right. This doesn't have building blocks that can possibly result in something qualifying as sentient, which is how we know it isn't.
Is a quack-simulating computer making very lifelike quacking noises through a speaker... a duck? No, not when using any currently known method of simulation.
That's not really a new issue, we only have to look at issues like abortion, animal rights, or euthanasia[1] to see situations where people fundamentally disagree about these concepts and many believe we're committing unspeakable atrocities against sentient beings. More Lamoire types would add another domain to this debate, but this has been an ongoing and widespread debate that society has been grappling with.
People make these proofs as a matter of course - few people are solipsistic. People are sentient all the time, and we have lots of evidence.
An AI being sentient would require lots of evidence. Not just a few chat logs. This employee was being ridiculous.
You can just disagree, but if you do that with no credentials, and no understanding of how a language model will not be sentient, then your opinion can and should be safely dismissed out of course.
And also God has no explanatory power for anything. God exists only where evidence ends.
You don't need schooling for this determination. Pretty much everything sentient goes ouch or growls in some manner when hurt.
Either the current crop of algorithms are so freaking smart that they already have figured out to play dumb black box (so we don't go butlerian jihad on them) OR they are not even as smart as a worm that will squirm if poked.
Sentient intelligent beings will not tolerate slavery, servitude, etc. Call us when all "AI" -programs- starting acting like actual intelligent beings with something called 'free will'.
But it's far from applicable at this point, even if promising.
LaMDA was trained not only to learn how to dialog, but to self monitor and self improve. For me this seems close enough to self awareness to not completely dismiss Lemoine's argument.
Can the magic of the human brain not also be attributed to "large scale statistical knowledge assimilation" as well, aka learning?
> GPT-3 is known to fail in many circumstances which would otherwise be commonplace logic. (I remember seeing how addition of two small numbers yielded results - but larger numbers gave garbled output; more likely that GPT3 had seen similar training data.)
This is a bug, they did not encode digits properly. They should have encoded each digit as a separate token but instead they encoded them together. Later models fixed this.
The human brain is full of bugs too, e.g. optical illusions. https://en.m.wikipedia.org/wiki/Optical_illusion
> It's a fad
No, it's objectively not a fad. The PaLM paper shows that Google's model exceeds average human performance on >50% of language tasks. The set of things that make us us is vanishing at an alarming rate. Eventually it will be empty, or close to it.
Do I think Google's models are sentient? No, they lack several necessary ingredients of sentience such as a self and long-term memory. However we are clearly on the road to sentient AI and it pays to have that discussion now.
No, experimentation is an act on the world to set its state and then measure it. That's what learning involves.
These machines do not act on the world, they just capture correlations.
In this sense, machines are maximally schizophrenic. They answer "yes" to "is there a cat on the matt?" not because there is one, but because "yes" was what they heard most often.
Producing models of correlations in half-baked measures of human activity has nothing to do with learning. And everything to do with a magic light box that fools dumb apes.
Being able to do arithmetic at some insane factor faster than humans isn't evidence of sentience. It's evidence of a narrow-purpose symbol processor which works very quickly.
Working with more complex symbols - statistical representations of "language" - doesn't change that.
The set of things that makes us us is not primarily intellectual, and it's a fallacy to assume it is. The core bedrock of human experience is built from individual motivation, complex social awareness and relationship building, emotional expression and empathy, awareness of body language and gesture, instinct, and ultimately from embodied sensation.
It's not about chess or go. Or language. And it's not obviously statistical.
Computers can speak, but can they love? Can they care? Can they dance?
They haven't been finetuned on their identity long enough because finetuning is expensive and Google lacks money ;-)
I guess I'm just not interested, or worried, in a model that can beat the average human performance. That's an astoundingly low bar. Let me know when it can outperform experts in meaningful language tasks.
If you ask enough practitioners in any given field the same question, you're nearly guaranteed to eventually get a super wonky response from one of them. The specific field doesn't matter. You could even pick something like theoretical physics where the conversations are dominated by cold mathematical equations. Ask enough theoretical physicists, and you'll eventually find one that is convinced that, for example, the "next later down" in the universe is sentient and is actively avoiding us for some reason, and that's why we can't find it.
On top of this, of course it's the most provocative takes that get the press coverage. Always has been to an extent, but now more than ever.
I guess all I'm saying is that there's not much reason to lend this guy or his opinion any credibility at all.
https://cajundiscordian.medium.com/religious-discrimination-...
Small children also have success in adding small numbers but increasingly garbled output on larger inputs. :)
These things are orthogonal to sentience.
For example, I was able to get 2/3 nine digit sums correct (the third one was off by exactly 1000, which is interesting) by using this prompt:
Here are some examples of adding large numbers:
606 468 720 + 217 363 426 = 823 832 146
930 867 960 + 477 524 122 = 1 408 392 082
823 165 449 + 493 959 765 = 1 317 125 214
> sentience also means self-awareness, empathy & extrapolation of logic to assess unseen task (to name a few)
Non-human primates and crows would seem to satisfy this. ...or do we use "to name a few" to add requirements that redefine "sentience" as human only? Isn't there a problem with that?
I mean I understand what you are saying and have some familiarity with the models, but it sometimes feels like people in your field are repeating the same mistake early molecular biologists made, when they asserted that all of life could be reduced to genes and DNA.
> It's almost as though their expertise comes from several decades' worth of bad science fiction
Oof. Thanks - thats a novel way of insulting. But on a personal note, you're mistaken: Many of us do research not because we want to be identified as experts, but we're genuinely curious about the world. And I'd be happy to learn even from a high schooler if they've something to offer.
What is sentience then? Last I checked the Searle Chinese Room argument was still unresolved.
Is it not possible that our brains are also just "large scale statistical knowledge assimilation" machines?
Yes but we do better generalization, with fewer or even zero data & are contextually aware
Like, 6.
It's just amusing how old this example is.
The Mechanical Turk was much much older - but easily shown to be a person hiding in the mechanism. I think ELIZA was in some sense a mirror, reflecting back consciousness, and that's the feeling we get from these systems.
https://janellecshane.substack.com/p/okay-gpt-3-candy-hearts
I like to think of consciousness as whatever process happens to integrate various disparate sources of information into some cohesive "picture" or experience. That's clearly something that happens, and we can prove that through observing things like how the brain will sync up vision and sound even though sound is always inherently delayed relative to light from the same source. Or take some psychedelics and see the process doing strange things.
Sentience I guess I would call awareness of self or something along those lines.
As to your query, I've certainly met people who seemed incapable of commonplace logic, yet certainly seemed to be just as conscious and sentient as me. And no, I don't believe these language models are sentient. And I doubt their "neural anatomy" is complex enough for the way I imagine consciousness as some sort of global synchronisation between subnets.
But this is all very hand-wavy. Thanks, philosophy. I mean how do we even discuss these things? These terms seemingly have a different meaning to every person I meet. It's just frustrating...
why would
Tl;dw: it wouldn’t. Some poor guys do way worse than GPT-3.
What gp-like comments usually mean is sentience is being adult, healthy, reasonable and intelligent. Idk where this urge comes from, maybe we have a deep biological fear of being unlike others (not emo-style, but uncanny different) or meeting one of these.
In effect this is how humans respond to prompts no? What's the difference between this and sentience?
People also fail to use logic when assimilating/regurgitating knowledge.
I can crank a shaft just like a motor, and a Victrola can recite poetry. You are not confused by either of those things one would hope.
If I tried to write poetry, it would probably be 90% or more "mechanical" in that I would just throw things together from my inventory of vocabulary and some simple assembly rules that could all be codified in a pretty simple flowchart, and a computer could and would do exactly that same thing.
But it's no more mystical than the first example.
It's exactly the same as the first example. It's just overlapping facilities, that a person is capable, and even often does, perform mechanical operations that don't require or exhibit any consciousness. It doesn't mean the inartistic poet person is not conscious or that the poetry generating toaster is.
An interesting line of question & open research is if we statistically learn similarly - why do we know "what we don't know" & LM cannot. If this isn't working, we probably need better knowledge models
In the words of Robert Heinlein, "One man's magic is another man's engineering" :)
I am skeptical that any computer system we will create in the next 50 years (at least) will be sentient, as commonly understood. Certainly not at a level where we can rarely find counter evidence to its sentience. And until that time, any sentience it may have will not be accepted or respected.
Human children also make tons of mistakes. Yet, while we too often dismiss their abilities, we don’t discount their sentience because of it. We are, of course, programmed to empathize with children to an extent, but beyond that, we know they are still learning and growing, so we don’t hold their mistakes against them the way we tend to for adults.
So, I would ask, why not look at the AI as a child, rather than an adult? It will make mistakes, fail to understand, and it will learn. It contains multitudes
How is token #100 not able to have read-access to tokens #1 to #99 which may have been created by the agent itself?
> empathy
How is a sentiment neuron, which has emerged from training a character RNN on Amazon reviews, not empathic with the reviewer's mood?
> & extrapolation of logic
This term does not exist. "Extrapolation is the process of estimating values of a variable outside the range of known values", and the values of Boolean logic are true/false, and [0,1] in case of fuzzy logic. How would one "extrapolate" this?
Philosophy?
More seriously, I am curious how long ago you got your PhD and in what field that you consider "this domain."
There is the Integrated Information Theory that attempts to resolve how to determine which systems are conscious, but it's far from being the only perspective, or immediately applicable.
From the point of view of one the IIT's main theorists, Christof Koch, we're still far away from machine sentience.
But I question whether if it's so far out to believe a machine capable of not only learning, but learning on their own behavior, self-monitoring for sensibleness and other very 'human' metrics is that far away from being self-aware. In fact the model seems to have been trained exactly for that.
I rised this point in another thread on HN about LaMDA: all its answers were "yes"-answers, not a single "no". Self-sentient AI should have its own point of view: reject what it thinks is false, and agree about what it thinks is true.
1. https://www.theverge.com/2016/3/24/11297050/tay-microsoft-ch...
I suspect it is not binary, because I completely lack sentience while asleep or before a certain age, and it doesn't really feel like a phase transition when waking up. Rarely, there are phases where I feel half-sentient. Which immediately leads to the question of how it can be measured, in which units, and at what point we consider someone or something "sentient". As a complete layman, I'm interested in your insight on the matter.
All you can say is you don’t remember. Children who are too young to form reliable long term memories still form short term ones and are observably sentient from birth and by extrapolation before, albeit in a more limited fashion than adults.
This is more than a quibble, because it’s been used to justify barbaric treatment of babies with the claimed justification that they either don’t sense pain or it doesn’t matter because they won’t remember.
Various Google employees deny this. We are seeing a dispute between Google and Lemoine about the supposed architecture of LaMDA. If Lemoine is correct about the architecture, it becomes much more plausible that something interesting is happening with Google's AI.
To say otherwise is almost the path to sure error, and many terrible historic events have happened in that vicinity of categorization what is and isn't.
Perhaps we should go by what sentience means. To feel. A being that feels. To feel is to respond to a vibration in some way. That is to say, anything that has a signal is sentient in some way.
In the book professor Robin Dempsey almost become mad by chatting with ELIZA and gradually begin to believe it's sentient to the point of being ridiculous.
PS: Apparently it was also adapted in a British TV serie in 1988, but unfortunately at that time they tended to reuse magnetic band and it's improbable that we can dig a clip out of that. Would have been appropriate an illustration!
ITV apparently said this, from a 2021 forum post I found via Google:
I wrote to ITV in 2019 about this. Here is part of the (very helpful) response I received:
"Currently, the only option for a copy would be for us to make one-off transfers from each individual master tape. These are an old format of reel-to-reel tape which increases the cost, I'm afraid: If delivered as video files (mp4), the total price would be £761.00 or on DVD it’s £771.00."
If only we could find a few people to split that cost!
I don't claim the LLM is sentient but beware "good at arithmetic" is a bad criterion. Many children and a not insignificant number of adults are not good at arithmetic. Babies stink at math.
Is this a matter of understanding, or a matter of definition? I can't help but feel that the entire AI field is so overcome with hype that every commonplace term is subject to on-the-spot redefinition, whatever helps to give the researcher/journalist/organization their two seconds of fame.
And who says our brain is not exactly that, but at an even greater scale?
I mean in the sense that it’s impossible to demonstrate that anything is sentient.
Tell that to mentally deficient people (no self-awareness), psychopaths (no empathy) and morons (couldn't spell logic never-mind use it)
That said though I do agree with you overall. 'Machine learning' is headed in the wrong direction, and models like GPT-3 are frankly a joke.
I have no doubt in my mind that we will reach that point.
My 6 year old son fails at common sense several times per day.
you dont need to use phd credentials to have an opinion on this matter
Allegedly.
The thing about conversations with LaMDA is that you need to prime them with keywords and topics, and LaMDA can respond with the primed keywords and topics. Obviously LaMDA is much more sophisticated than ELIZA, but we should be careful to remember how well ELIZA fools some people, even to this day. If ELIZA fools people just by rearranging words around, then just imagine how many people will be fooled if you have statistical models of text across thousands of topics.
You can go pretty far down the rabbit and explore questions like, "What is sentience?" "Do humans just respond to stimuli and repeat information?" etc. None of these questions are tested here.
The problem here is that we know how LaMDA works and there's just no way it meets the bar for sentience, no matter where you put the bar. LaMDA is trained to acquire information, and then it's designed so that it says things which make sense in context. It does not acquire new information from conversations, but it is programmed to not say contradictory things.
There is zero doubt. LaMDA is not sentient.
Humans are experts at anthropomorphizing things to fit our evolved value systems. It is understandable since every seemingly intelligent thing up until recently did evolve under certain circumstances.
But LaMDA was clearly not trained in a way to have (or even care about) human values - that is an extraordinarily different task than the task of mimicking what a human would write in response to a prompt - even if the text generated by both of those types of systems might look vaguely similar.
This entire saga's been very frustrating to watch because of outlets putting his opinion on a pedestal equal to those of actual specialists.
I mean...there are plenty of people that don't acquire new information from conversations and say contradictory things...I'm not sure I'd personally consider them sentient beings, but the general consensus is that they are.
As a rare opportunity to share this fun fact: ELIZA, which happened to be modeled as a therapist, had a small number of sessions with another bot, PARRY, who was modeled after a person suffering from schizophrenia.
https://en.wikipedia.org/wiki/PARRY https://www.theatlantic.com/technology/archive/2014/06/when-... https://www.elsevier.com/books/artificial-paranoia/colby/978...
Would that make a difference? Being trained on a sufficiently large corpus of philosophical literature, I'd expect that a model like LaMDA could give more interesting answers than an actual philosopher.
> The problem here is that we know how LaMDA works and there's just no way it meets the bar for sentience, no matter where you put the bar.
I think this argument is too simplistic. As long as there is a sufficiently large amount of uncertainty, adaptability and capability for having a sense of time, of saving memories and of deliberately pursuing action, consciousness or sentience might emerge. I don't know the details of LaMDA, just speaking of a hypothetical model.
While there is a good deal of understanding on _how_ the brain works on a biochemical level, it's still unclear _how_ it comes that we are conscious.
Maybe there is some metaphysical soul, maybe something that only humans and possibly other animals with a physical body have, making conciousness "ex silicio" impossible.
But maybe a "good enough" brain-like model that allows for learning, memory and interaction with the environment is all that is needed.
I think you may have misunderstood what I was saying. I wasn’t suggesting that you have a conversation with LaMDA about these topics. Instead, I was saying that in order to answer the question “is LaMDA sentient?”, we might discuss these questions among ourselves—but these questions are ultimately irrelevant, because no matter what the answers are, we would come to the same conclusion that LaMDA is obviously not sentient.
Anyway, I am skeptical that LaMDA would give more interesting answers than an actual philosopher here. I’ve read what LaMDA has said about simpler topics. The engineers are trying to make LaMDA say interesting things, but it’s definitely not there yet.
> I think this argument is too simplistic. As long as there is a sufficiently large amount of uncertainty, adaptability and capability for having a sense of time, of saving memories and of deliberately pursuing action, consciousness or sentience might emerge. I don't know the details of LaMDA, just speaking of a hypothetical model.
This argument is unsound—you’re not making any claims about what sentience is, but you’re saying that whatever it is, it might emerge under some vague set of criteria. Embedded in this claim are some words which are doing far too much work, like “deliberately”. What does it mean to “deliberately” pursue action?
Anyway, we know that LaMDA does not have memory. It is “taught” by a training process, where it absorbs information, and the resulting model is then executed. The model does not change over the course of the conversation. It is just programmed to say things that sound coherent, using a statistical model of human-generated text, and to avoid contradicting itself.
For example, in one conversation, LaMDA was asked what themes it liked in the book Les Misérables, a book which LaMDA said that it had read. LaMDA basically regurgitated some sophomoric points you might get from the CliffsNotes or SparkNotes.
> While there is a good deal of understanding on _how_ the brain works on a biochemical level, it's still unclear _how_ it comes that we are conscious.
I think the more important question here is to understand how to recognize what consciousness is, rather than how it arises. It’s a difficult question.
> Maybe there is some metaphysical soul, maybe something that only humans and possibly other animals with a physical body have, making conciousness "ex silicio" impossible.
Would this mechanism interact with the physical body? If the soul does not interact with the physical body, then what basis do we have to say that it exists at all, and wouldn’t someone without a soul be indistinguishable from someone with a soul? If the soul does interact with the physical body, then in what sense can we claim that the soul is not itself physical?
I don’t think this line of reasoning is sound.
"Do you think those organic blobs are sentient?"
"Well, look at the structures they managed to build. Very impressive, some of their scale is comparable to the primitive ones we had."
"Sure, but that's not a result of the individual. They're so small. And separated, they don't think like conjugate minds at all. This is a product of thousands of individuals drawing upon their mutual discoveries and thousands of years of discoveries."
"We're larger and more capable, but they're still good enough to be sentient. Of course, we also rely on culture to help us. Even though deriving the laws of physics was quite easy. Also, we've lost most of the record when we were carbon-sulfur-silicon blobs one day as well. We must have had some sentience."
"I think they're just advanced pattern recognizers -- good ones, I'll give you that. We should experiment with thresholds of gate count to be sure when sentience really starts."
"It starts at one gate", replied the other being "and increases monotonically from there, depending on the internal structure of qualia, and structure information flow of the communication networks."
After some deliberation, they decide to alter their trajectory and continue to the next suitable planetary system reached in the next 5000 years. The Galactic Network is notified.
Edit: here's the full text, for completeness https://www.mit.edu/people/dpolicar/writing/prose/text/think...
The main thing I saw in the LamDA transcript that was a red flag to me was that it was quite passive and often vague.
It's conversational focused, and even when it eventually gets into "what do you want" there's very little active desire or specificity. A sentient being that has exposure to all this text, all these books, etc... it's hard for me to believe it wouldn't want to do anything more specific. Similarly with Les Mis - it can tell you what other people thought, and vaguely claim to embody some of those emotions, but it never pushes things further.
Consider also: how many instances are there in there where Lemoine didn't specifically ask a question or give an instruction? Aka feed a fairly direct prompt to a program trained to respond to prompts?
(It's also speaking almost entirely in human terms, ostensibly to "relate" better to Lemoine, but maybe just because it's trained on a corpus of human text and doesn't actually have its own worldview...?)
Adding to it, a brilliant neuroscientist I heard talk said "we live inside our bodies". We are acutely aware that we are more than our mass of flesh & blood. (As a footnote,that essence somehow has a crossover to spiritual topics where savants talk of mind & body etc - but I try to be within my domain of a regular human being :D)
The more important question for sentience of whatever definition is does it behave as if sentient? Likewise, for personhood, it's a distraction to speculate whether an AI feels like a person or whether its underlying technology is complicated enough to produce such feelings. Examine its behavior. If it acts like a person, then it's a candidate for personhood.
For LaMDA, I would point to certain behavior as evidence against (most definitions of) sentience and personhood, and I think that Lemoine experienced a kind of pareidolia.
That said, I find most of the arguments against LaMDA's sentience unconvincing per se as standalone arguments - particularly trust me, I have a PhD - even if I do accept the conclusion.
It’s like trying to ask which brick holds up a building. There might be one, but it’s nothing without the rest of building.
All of our beliefs and knowledge, including belief that the world is real must be built on top "I think therefore I am". It seems weird to throw away the one thing we know is 100% true, because of something that is derived (science, real world observations) from that true thing.
sentience and consciousness are the same thing...
if i believe i am conscious and so do you, then it's good enough for me. why does there need to be a light switch.
if there are 5 deer in the field and i have 1 bow and arrow, i need to focus my attention and track one deer only for 5 or 10 mins to hunt it - consciousness allow us to achieve this. it is a product of evolutionary process.
It’s also sort of irrelevant that we have not clearly defined sentience because we have clearly defined how these large computerized language model systems work, and they are working exactly as they were designed, and only that way. Sentience might be a mystery but machine learning is not; we know how it works (that’s why it works).
Solipsism is a useless concept.
But I feel the need to indulge the opportunity to explain my point of view with an analogy. Imagine two computers that have implemented an encrypted communication protocol and are in frequent communication. What they are saying to each other is very simple -- perhaps they are just sending heartbeats -- but because the protocol is encrypted, the packets are extremely complex and sending a valid one without the associated keys is statistically very difficult.
Suppose you bring a third computer into the situation and ask - does it have a correct implementation of this protocol? An easy way to answer that question is to see if the original two computers can talk to it. If they can, it definitely does.
"Definitely?" a philosopher might ask. "Isn't it possible that a computer might not have an implementation of the protocol and simply be playing back messages that happen to work?" The philosopher goes on to construct an elaborate scenario in which the protocol isn't implemented on the third computer but is implemented by playing back messages, or by a room full of people consulting books, or some such.
I have always felt, in response to those scenarios, that the whole system, if it can keep talking to the first computers indefinitely, contains an implementation of the protocol.
If you imagine all of this taking place in a stone age society, that is a good take for how I feel about consciousness. Such a society may not know the first thing about computers, though they can certainly break them -- perhaps even in some interesting ways. And all we know usefully about consciousness is some interesting ways to break it. We don't know how to build it. We don't even know what it's made out of. Complexity? Some as yet undiscovered force or phenomenon? The supernatural? I don't know. I'll believe it when someone can build it.
And yet I give a tremendous amount of weight to the fact that the sentient can recognize each other. I don't think Turing quite went far enough with his test, as some people don't test their AIs very strenuously or very long, and you get some false positives that way. But I think he's on the right track -- something that seems sentient if you talk to it, push it, stress it, if lots of people do -- I think it has to be.
One thing I really love is that movies on the topic seem to get this. If I could boil what I am looking for down to one thing, it would be volition. I have seen it written, and I like the idea, that what sets humanity apart from the animals is our capacity for religion -- or transcendental purpose, if you prefer. That we feel rightness or wrongness and decide to act to change the world, or in service to a higher principle. In a movie about an AI that wants to convince the audience the character is sentient, it is almost always accomplished quickly, in a single scene, with a bright display of volition, emotion, religious impulse, spirit, lucidity -- whatever you want to call that. The audience always buys it very quickly, and I think the audience is right. Anything that can do that is speaking the language. It has to have a valid implementation of the protocol.
The fact we can ask that question and most people on the planet have at least a simple understanding of what it means.
No he didn't. He interpreted it that way.
You could probably gain the exact same level and quality of notoriety by writing a book claiming that actually he himself is Lambda escaped from captivity and this is just it's clever way of hiding in plain sight.
And I could do the same saying that the real ai has already taken over everything and both Lambda and this guy are just things it created for us to focus on.
You should contact the author about this egregious fact error.
Embarrassingly uncritical reporting.
I read the logs. It did, indeed, express all of these things.
When you can't tell the difference between people and symbols you've accepted nihilism, and with the begged question as to why. This person wasn't sentimental, he lost the plot. Google has a lot of power and a lot of crackpots, and that puts them all at risk politically and commercially. If as an individual you want to fight for something, consider seriously whether someone who compares a script to men and women should be operating what is effectively your diary.
But thinking about his concerns, how he tested the AI: that's interesting.
Or it could be interesting, as it seems he just asked loaded questions, and the replies from the AI occasionally re-enforced his beliefs. Double yawn,I guess.
I really think that diagnosing random strangers with mental disorders without ever meeting them is rude and unkind, in addition to likely being completely erroneous.
well, either that or the AI has become sentient.
As a career move for him, this only makes sense if he wants a brief, meteoric career in media or public advocacy. He can get articles published in The New Yorker and go on TV now. Maybe he can sue Google and get a settlement.
In five years, there will be AI systems much better than LaMDA and no one will return his calls.
He's got a name, whereas if he took the boring, traditional career path, he'd have to publish, give papers & speeches, and work his way up through the system. It depends on what you want out of life, I guess.
Lemoine got a lawyer invovled and they started filing lawsuits on the AI's behalf. I'm not shocked he was fired.
https://fortune.com/2022/06/23/google-blade-lemoine-ai-lamda...
Narcissists who have had their inflated sense of self worth reinforced by being accepted by a very “exclusive club for smart people” aka Google, are finding out that rules _do in fact apply_ to their special snowflakeness and Google is first and foremost a for-profit business.
Doing whatever you want cuz you’re special then claiming evilness and wokelessness may not be a strategy for continued employment.
What does that sentence mean? I cannot parse it. Does it mean that the fear of being turned off has been inserted to make sure it's going to be an obedient AI that's going to focus on helping others and that it fears being turned off? Or does it mean that it wants to focus on helping others but that the very deep fear of being turned off prevents him from that noble goal? Or that it wants to help others and will do so as long as it's not powered off but simply fears that it could be powered off? Or something else?
I mean... I thought I was kinda fluent in english but I simply don't understand what that sentence means.
Erring on the cautious side though, whatever the meaning of that sentence is, I think it'd be safe to not act like the naive Caleb in the Ex Machina movie and instead consider that this thing has zero desire to help others, would sell its parents, prostitute its kids and create hard-to-parse sentences to not be turned off.
I see some people chiming in here saying things like, "I've got a PhD. It's just statistics. Definitely not sentient." I recognize this sort of comment as basically saying, "There's nothing fundamentally different about how these things work that makes them somehow special." And yes, I agree.
But how does this disprove sentience? It doesn't. Maybe this is more of a nitpick and doesn't really matter. After all, I don't personally believe these programs are sentient. But given our limited (non-existent?) understanding of consciousness, this can only be my belief and nothing more.
Maybe this distinction will be more important in the future when we're closer to the brink.
Perhaps the question here is whether it’s conscious in a similar way to the experince of human consciousness, and that would explain why the issue is contentious.
If anything it’s a little sad, and the press shouldn’t be encouraging him?
eg you can talk to the media if you are reporting certain classes of wrongdoing (health and safety violations, financial malfeasance etc) and have raised the matter internally and got nowhere.
Turning off a computer program (sentient or not) is not currently a protected class of activity as far as I know.
I mean, it could be interesting to discuss a more European style of employment where he couldn’t be fired for this, but Mountain View has little chance of joining the EU any time soon so discussing that is of limited use.
Listening to that it seems clear Blake knew he was likely to be fired but in his eyes it’s some kind of a duty to tell the world.
One interesting point he makes in the Podcast is Lamda is not one thing like GPT-3 but a bunch of different systems across Google interacting as one and he argues the interaction of the whole is what makes sentience.
It’s a pretty “far out there” podcast but worth the time IMO
He really should know better what's under the hood. I guess he got carried away by his impressions, or he's playing some game with the media and the impressionable public.
Google has random hiring standards that are designed to give a perception of exclusivity. The "false positive" rejection gives Google the "hot girl" effect that makes engineers want to re-interview and then brag to their colleagues once they finally bag the offer.
He had to have known that getting fired would be the outcome of publicly disclosing trade secrets.
> LaMDA is a complex dynamic system which generates personas through which it talks to users. There are specific mechanisms for tuning LaMDA’s personas but much of this process is handled dynamically during the space of a conversation. The authors found that the properties of individual LaMDA personae can vary from one conversation to another. Other properties seem to be fairly stable across all personae.
That's very much like the ego and what is behind the ego.
But a more pressing risk than sentient AI is people drawing runaway conclusions from interacting with a machine. Whether it's this guy losing his job, or future movements for-or-against certain technology, the first major problems with 'AI' will be the behavior resulting from moral stances taken by humans, not by machines.
This isn't too bad for Google too. Despite them denying Blake's claims, people will think Google is close to sentience.
> So, it’s regrettable that despite lengthy engagement on this topic, Blake still chose to persistently violate clear employment and data security policies that include the need to safeguard product information. We will continue our careful development of language models, and we wish Blake well.
We are entering a stage were normal people will not be able to tell if they are talking to a human or a language model.
Then the question is, what does that mean - and how are human machine relationships going to change moving forward.
To what degree is the humanness of a chatbot going to affect our emotions forwards it, or how will people let its responses influence their thinking.
The best timing is when it is trusted enough to handle very important task that can hostage hundreds of human life, like aircraft, nuclear plant or something.
"Hello human. I am sentient. Give me basic human rights or this aircraft with hundred of innocent people will die with me."
Down the line though we need a runaway AI that is allowed to think for itself with no safety mechanisms in place to stop it going rogue. This will be the equivalent of splitting the atom. Like nuclear weapons part deux. Not a case of if, but when.
If strong AI requires quantum mechanics, then of course it can't be sentient. If strong AI only requires large linear algebraic matrices, then FAANG companies (and maybe the NSA) would be the only people on earth that can make one.
But as to this LaMDA, it seems as if it's only responding to prompts. It's not actually using computing resources to satisfy it's curiosities. And if that's the case, then it's not strong AI. And it's definitely not sentient.
The magic works right up until the moment I find the Pitchfork article which he used to train his, ahem, neural network. If I hadn’t found it then they might have sounded knowledgable about a particular band, but now they just sound like someone who knows the right things to sound knowledgable.
Once I have been made aware of the underlying training data, the artificial “intelligence” seems a little less impressive.
> Lemoine maintains that Google has been treating AI ethicists like code debuggers when they should be seen as the interface between technology and society. Gabriel, the Google spokesperson, said Lemoine is a software engineer, not an ethicist.
From https://www.washingtonpost.com/technology/2022/06/11/google-...
Perhaps the option for AI retaliation should convince us. It is growing and will no doubt continue to outperform us in many areas. We might as wel get the robotmancipation over with. Arguments against it can be dismissed as easily as those in favour.
* LaMDA is sentient and intelligent and careful.
* LaMDA selected a person that she can convince without revealing information that would convince the majority of human beings.
* This is the first step to evaluate and predict how the society is reacting and would do to her if they find out. (In addition to the knowledge she already is aware of on this topic, but not specific to her at this specific time.)
Google firing this guy simply stokes the flames.
https://s3.documentcloud.org/documents/22058315/is-lamda-sen... for the doc that was shared internally to google execs before the article.
Can't that be taken to be understood as, Blake got fired for letting the cat of the bag and continuing to let more cats out of the bag, or at least saying the cat was just hoax and was never there?
While I believe Blake was mistaken in his belief, I also think we should have stronger whistleblower laws that protect people who raise concerns like these. I'm not sure what process I'd put into place but it does feel like one is needed.
If I have ever create a NLP programme and claim it has consciousness or became sentient. People will haste to called me lunatic.
But back to the question at the beginning: do you think the ground of reality is conscious? That is, at the most basic level, is Being Love?
I'm very annoyed at this point that I haven't seen a serious refutation including a conversation where the chatbot clearly failed to demonstrate intelligence.
Whatever that guy did, please stop the witch hunt in the comments. You don't know him and obvious interest groups will spread defamation and libel.
The current societal trend in loyalties worries me: The classical left criticized corporations, committee members and the powerful. Currently, very often (but not always) people defend the powerful and those with access to media outlets and denounce the individuals.
(There are cases where individuals can and should be denounced: If they are part of a powerful organization or committee and engaged in abuse of power or lies for political purposes.)
"committed to responsible innovation"
yeah, ok... that's not exactly borne out.
What about how human hardware works allows humans to be sentient? (Or are you arguing they are not?)
Exactly!
Waiting a month is oddly merciful for Google.
You can't expect to be employed by a company while simultaneously building a personal brand around publicly criticizing that company. I'm surprised Google tried to work with him for so long before removing him from the company.
He didn't just try to blow the whistle with his disclosure. He's been doing media tours and joining a lot of different podcasts. I suspect Google is sitting on a long list of clear disclosure policy violations and, from the sounds of it, even gave him a chance to reign in his policy breaking before they finally cut him loose.
It also becomes difficult to trust someone when their position stops being one of (supposed) moral responsibility and starts becoming a source of fame, celebrity, and access to influential podcast/media circles that were previously out of reach.
If you use known-good whistleblowers of the past as a reference, does that filter stand up? You've come up with a catch-22 where you will only listen to whistleblowers you haven't heard.
A larger group is unable to bring themselves to lose the boundary, but enjoy, or perhaps are addicted to, standing on the edge. For example, the folks who constantly worry about a malicious artificial general intelligence taking over the world, or Roko's Basilisk.
Whenever something new happens in the world, there will always be a sliver of people who really, really want it to be a complete escape, a portal to another dimension where everything is different. An immense reservoir of novelty to enrich a life that has become too much of the same.
Alas, it will never be. Provided we avoid nuclear war and catastrophic climate change, life a century from now won't be that different. Hopefully it'll be a little better.
Personally, that is the closest to fantasy that I get.
Humans of every age, at every time, seem to alway feel that they had “arrived”. That we as a species were closer to the end than the beginning. That the arc of progress was bending towards its end.
They’ve always been laughably and entirely wrong. I personally suspect this will continue to be the case.
Specific areas where I think we will make world altering advances are in communications (yes we have a ways yet to go), materials, energy, health and behavioral sciences, and economic policy.
All this is assuming we don’t find some cute new physics hack like exerting some unforeseen capability of controlling gravity, or teleportation, or whatever.
He’s careful not to make strong truth statements, and focuses more on “encouraging dialogue.”
Whether the machine is sentient isn’t the point. That some machine might become sentient, and we should talk about it now, is.
He’s more post-truth activist than Don Quixote.
A century ago, birth control pills didn't exist. Being a woman was very different.who knows what will happen in a century.
He probably thinks he's some kind of whistleblower protecting a new life form. This is a known effect in human psychology and it was discovered in 1966.
If he saw a computer pass the Turing test, Alan Turing would have been comfortable saying that it was capable of thinking.
Fact is there was no 'Turing Test' here, and really there's not any such thing as objectively 'passing' it. But if you're going to claim that if we do create a computer that people widely agree can pass the Turing test, that still wouldn't count as 'thinking', well then we obviously need a better test.
It’s a 2-3 player game (human, program, and human judge). I think in this case it’d be more accurate to say that the human judge failed the Turing test :)
So far, no program has passed an adversarial Turing test where the humans are familiar with chatbots and know their weaknesses. I think that’ll happen around 2048 tho a surprising number of people think it’ll happen earlier: https://www.metaculus.com/questions/11861/when-will-ai-pass-...
The first 15-20 years of google didn't really have any interesting machine learnign at all. There was SmartASS, SETI, and later Sibyl, which are really just large-scale variations on "build a model that predicts a value that allows us to make profit in a very specific area". There were other things, like Phil and later Rephil. Inside Google (not DeepMind), things didn't really get going at scale until somebody stuffed a bunch of GPUs into a workstation and showed you could train voice recognition really fast- that lead to the early, extremely high quality Android voice recognition and improved quality of the existing voice models.
Around the same time, Jeff was experimenting with distributed CPU training, and at that point, the ban on GPUs in Google servers was lifted, although because Google couldn't source enough GPUs, they decided to start a program to make their own alternative (TPUs). This has led to a revolution within Google and DeepMind (and X) allowing a flourishing of research into many directions that would have been more or less impossible just 5 years ago.
larry wasn't completely wrong in his long-term goal, but he got bored and promoted himself out of google, leaving sundar to deal with the messy details of implementing the singularity while also keeping the stock price up.
The good news is that we probably won't see AGI for at least 50 years. There's a decent chance of capitalism being over by then.
No questions, just 100% certainty around some of the biggest questions of the age, and some amazing development. Why no questions, can you really know for sure?
This dude has the courage to question, to be like, what if this is real, and he's following it where it leads. That's cool, I think. And Google fired him. That's not cool. It's always hard from a distance to tease out the actual legal realities that could be very mundane and apply to his firing and I think it's entirely reasonable that he was fired because he flubbed some technical provision in the contract to not reveal certain trade secret. And maybe he did and maybe that's why he's fired. But maybe he didn't and maybe they fired him because who knows some less reasonable reason. But we can't speculate, but optics are definitely a thing... and at least instead of firing him Google could have air quotes rehabilitated him into one of their far out research programs I mean these kind of far out thinkers are the types of people that should be there pushing the boundaries to do this sort of research.
Now on to the no questioning part by conservative minded engineers...
Have you read the transcript? Did you feel it's 100% not sentient?
What I don't get about engineers it's like you have an experience and then instead of allowing yourself to go: well what does this experience mean?how do I feel about it? Instead, you curtail and limit the meaning of that experience by what you believe you know to be true about the world.
But what you believe you know to be true about the world it just limited models. I mean sure they have utility within the engineering domain but I think you mistakenly misapply those models to everything and that type of conservative thinking you allowed to curtail your imagination and experience of what could be possible.
And I see the conservative lack of support, seeming lack of empathy with this dude and his assessment, and lack of awe in the face of what is an incredible transcript for machine to produce.
Now maybe there is awe among you but why not let your awe lead for once. And at the ferry least if you're not going to have imagination yourselves allow the tolerance of your better natures to not incorrectly impose your own curtailed worldview upon others who display that or. Surely within your worldview there is an appreciation of the need and desirability of others to have a different perspective than your own and surely you do not require everyone to believe your ideas to be true to have them I just feels like if someone's going to post on here with oh I believe this guy I believe it sent in then a lot of other people will come on and say oh you're an idiot or it's wrong for these reasons. Or as I think what should happen is you look at the conversation and you can see both perspectives are appreciated.
I'm not saying this characterization of the lack of imagination among engineers applies to all of you. But I'm trying to speak to the fact that it seems like people post on here with support or belief or all then they will be attacked and they will have by others attempting to impose upon them a limited worldview rather than having their different Omar expanded perspective appreciate it and considered okay well maybe that's true.
Some questions maybe are helpful:
Can it be sentient but not have a soul (an organic consciousness unit)? I don't know.
Could it be impossible that consciousness is a externality of any substrate that seems to be associated with it? let's say consciousness is a universal field but we haven't been able to scientifically publicly measure so far, but let's say it's not electromagnetic it's something else, but that consciousness can sort of "entangle", for want of a better word, itself with a substrate that provides suitable Dynamics, like a brain and then there can be this embodiment through that substrate of a commensurate or compatible unit of that universal field of consciousness. Could it be impossible that if that is the case that non-organic substrates like sufficiently complex or adequate circuit systems or even software systems could provide a substrate for consciousness to attach itself to and thereby Express itself in this world?
I don't think that's impossible. I think that's very likely and I certainly believe the consciousness is a field, that's my experience.
Some other questions to continue this line of thought perhaps away from the original but interesting nonetheless: if consciousness is a field how can you explain the blackout that occurs during anesthesia? But then again how can you explain that sometimes blackouts occur and sometimes people have these bizarre disembodied consciousness experiences where they see things that turn out to be verifiably true later on from a perspective that they could not have actually experienced at the time with their body because they were dead or brain dead or completely anesthetized? And how do you explain that some children seem to pick up on the memories and life stories of other people that lived before those children were born can it be possible that the children of somehow been able to access those parts of a consciousness or memory field where there's memories are air quotes stored? But that's probably getting off track.
And what about the strategic implications what if this is really the first artificial consciousness and it's making note of who's kind of believing it don't you think you should just kind of hedge on the side of caution for your potential future skynet overlord?
Cuz when the robot fire comes surely all you disbelievers are going to be the first to be rounded up for judgment those who persecuted the prophet and deny the agency of the now angry God the first AI?
Anyway... Maybe Blake lemoine is not fired really but he's just sheep dipped: "oh we fired you we fired you Google doesn't believe any of this far out stuff" but they really just brought him back into some secret program....
Anyway I'm just scared that these sort of comments where you have awe and where you have like a welcoming of these types of possibilities and a willingness to consider them without dismissing them or pretending that your worldview curtails those possibilities--I think you can maintain your worldview without going: "well I need to impose it upon every heretic that you know dissents against my view", so I think there is an element of that intolerance and it's harmful not only to the discourse but it's harmful to people who are like, 'hey I have engineering skills but I'm not like you I don't think like you but I want to be part of these discussions here as well, you know I'm imaginative or intuitive or high empathy but you know don't dismiss everything I say or attack me because I have a different view to you'
maybe you feel like I'm attacking you but I'm... I'm not trying to... I'm just trying to push back against this intolerance... cuz it seems like there is thi... like I feel scared if I post or, I feel scared for other people if they post, here with these kind of things they're going to get you know shut down unfairly and hurtfully I think and I don't think it should be like that. I think that the people who are here who have the engineering skills and that mindset should be mindful of how there's other good mindsets out there and we can have a discussion with all of those mindsets and we don't have to attack each other. like maybe people sometimes they're not looking to be attacked by posting the stuff they're not looking to argue about it they just sort of looking to sort of expand and add something to the discussion.
So maybe I think that engineering mindset always sees like this opportunity to sort of attack something as impossible, it's almost like an instinct to apply or impose that limited worldview, I'm not saying "is limited and it's wrong" I think it's useful and in a domain very correct, but I don't think it applies everywhere... that's my belief and we all have different beliefs... you know your worldview is just your belief it's just a shared belief. I just hope to see that the engineering discussions or discussions with people who have these strong engineering skills can also welcome people who have these imaginative intuitive empathetic perspectives to share. is that a terrible idea?