Dennis Hackethal’s Blog
My blog about philosophy, coding, and anything else that interests me.
Tweets
An archive of my tweets and retweets through . They may be formatted slightly differently than on Twitter. API access has since gotten prohibitively expensive – I don't know whether or when I'll be able to update this archive.
But in case I will, you can subscribe via RSS – without a Twitter account. Rationale
Okay. When you use the term “intelligence,” are you referring to the sophistication of those pre-installed apps, the creativity algorithm itself (regardless of how sophisticated the knowledge it produces may be), the sophistication of the knowledge it produces, or something else?
The difference is that people are creative. We are not born with knowledge of how to communicate, or how to recognize faces, or how to build space shuttles, etc. We learn this after birth by creating the knowledge ourselves.
Yes, rules for simple communication can be determined entirely genetically. So the example of crows recognizing faces over generations need not be evidence of intelligence.
You'll see, but can you answer my question?
So we can agree that the arrogance doesn't have any effect on a claim's truth value?
Regarding your other point, do you then think that the spaceshuttles people have built were genetically encoded? If they were, how come it took us so long to build them?
What do you think we do know about the human mind?
I know nothing about you or your background knowledge, so I’ll just say this: some people do know. Not all of humanity thinks or knows the same things.
I can expand on what we know about the topic if you’re interested. Up to you.
He later sends the Spartans a poem "in commemoration of our wrestling competition."
Am I the only one seeing this?
Plato tells the others about how he's been hanging out with some Spartans, and how they "were up all night, wrestling by candlelight." He then "smiles indulgently in recollection."
They say homosexuality was much more common in antiquity than it is today.
Chapter 10 of "The Beginning of Infinity," called "A Dream of Socrates," contains a fictional dialog between Socrates and some of his associates, including Plato.
🧵👇
Sure thing. I didn't use a library for that. Server-side filter functions based on topics should do the trick. That's not to say it might not be easier to use a library!
Which story gets them more clicks? A doomsday one or an optimistic one?
So it take it your answer to my question is, "it's stored in the beaver's genes."
Where do you think that knowledge comes from?
Do you think that, if a statement is true, if it is made arrogantly, that that changes it to being false?
Exactly. And can you think of a way communication could be done non-creatively?
Can you think of a way to explain this that doesn't involve Lamarckism?
Ah, good point, so let's take perhaps the most complex of artefacts some animals create, such as beaver dams and spider webs. Where do you think the knowledge of how to build beaver dams resides in the beaver, and where does it come from?
Do you think a newborn baby can do as much as a newborn dolphin?
Humans know about the properties of uranium from explanations (which they created themselves), not from rewards or punishment (i.e., positive or negative reinforcement).
In the case of humans, those ideas originate from their creativity. But in the case of pigeons, these behaviors are just variations on inborn ideas, slightly changed according to a reinforcement-"learning" algorithm.
Depending on what you mean by "capacity," that may amount to a belief in the supernatural; see David Deutsch's concept of "universal explainers," explained in his book "The Beginning of Infinity."
No, datascript has a listen! fn to listen for any transaction made on the client, which transaction I then send to the server, which transacts it against the database, and relays it to (some or all) other clients, who run the transaction (only) locally in their turn.
@HeuristicAndy @ToKTeacher
I think the getting dizzy/high is a side effect and not what the pigs are seeking. The reward may just be some genetically-determined response that says "the nutrients contained in this food are good, eat more of them."
Kitt, are you interested in starting a thread in the 4S group and discussing it?
The main reason people doubt the achievability of AGI is failure to understand the implications of computational universality.
The main reason those who claim to work on AGI don't actually is failure to understand that it requires epistemological work.
I think you're still thinking in terms of the knowledge crows, monkeys, and dolphins contain, not in terms of the knowledge they create, and you're then considering the fact that they contain sophisticated knowledge as evidence of their intelligence. Am I wrong about that?
There are uncivilized tribes who live in the wild and they have spears, fire, high-level language, they build roofs and huts, albeit primitive ones. Have you ever seen a dolphin, crow, or cow do any of that?
And is there no way to possibly explain that, too, through genetically determined behaviors?
I reiterate the difference between "smart" and "intelligence" a lot with people but usually they keep giving me more examples of smarts and claim they're evidence of intelligence, which they're not.
Which can be explained through the use of an inborn facial-recognition algorithm, which is admittedly impressive and sophisticated, but doesn't require intelligence.
Not sure which part you're referring to, but either way: ironically, it's only because we are humans that we can even have these thoughts.
Regarding illiteracy, you should have more faith in your abilities, and your ability to improve them.
You're interested in discussing a complex problem, and that alone shows you're capable.
Having a society is cool but some animals have tribes and primitive "societies" too. And in any case, intelligence is an attribute of individuals, not societies.
In "The Beginning of Infinity," Deutsch explains clearly and lucidly what makes people different (and infinitely special). In short, it's that we're what he calls "universal explainers." Highly recommended!
Again, we'd want some explanation in favor of animal intelligence so powerful that what I'm saying couldn't be right.
But I think the better explanation is, that crow has no idea what cars are, didn't want to save the hog, was just pecking at it for food, which happened to make the hog move, and that all this happened on a road with a car approaching is incidental.
...because that means they must have created that knowledge during their lifetime (modulo reinforcement "learning," which doesn't count).
Agreed, Twitter is terrible at facilitating discussions.
What you're saying is true if the crows contain knowledge about cars despite never having inherited that knowledge genetically...
- It's not so much that the words are wrong—doesn't matter so much what words we use—but what we know them to mean. And in their meaning they don't fit. Also, I don't think you're a retard, nor am I trying to "prove [myself] right," which, as I have pointed out, is impossible.
- That conclusion is one that, interestingly, some people actually draw—with devastating results! Intelligence is possible, it's just that it doesn't involve "acquiring" knowledge passively, but actively creating it, even when you are the "recipient" in a conversation.
- Even Google makes mistakes.
- I recommend Popper's book "Objective Knowledge" and David Deutsch's "The Beginning of Infinity."
Yes, knowledge encoded in genes is biological (compare Popper).
Yes, humans are animals—but a very special and different kind because we are creative.
And it seems you ignored my other comment:
Dunno if there are any out there, but it's easy to do: since all your transactions are just data, you can send the EDN encoding those transactions to the server and back and run them both on the server and on the client (with the requisite authorizations).
@HeuristicAndy @ToKTeacher
E.g. when pigs eat fermented grapes it's not because "they enjoy being drunk" but because they are genetically programmed to eat grapes and seek reward, and if eating grapes triggers a reward in them, they will do it again and more so, whether it "buzzes" them or not.
- Blame Google for what?
- You can't check out what?
- People are intelligent.
- This is related. Also, when did I change the subject before?
We’d want an explanation that that could not have been the case.
In any case, what I was getting at is that we have to explain animal behavior in terms of biological evolution and selective pressures, and cars are too recent a phenomenon to play a role in that.
And even if they did, the crow may as well just be picking the hog for food.
If they know what roads are given that that knowledge is not encoded genetically, that could be evidence they’re intelligent. Not proof (because no amount of evidence proves any conjecture to be true, see Popper).
“Acquiring” knowledge is impossible because it requires induction/Lamarckism, which is impossible (see Popper, Deutsch).
@HeuristicAndy @ToKTeacher
A mind in the sense of some software running on the animal’s brain, sure, but not necessarily in the sense of a creative mind.
Also, those particular behaviors you mention may be better explained as unintended consequences of a reinforcement-“learning” algorithm.
@n_iccolo @ModelsofMind @CNC3P0 @SamHarrisOrg
Well, the thought experiment is a bit flawed, because such a catastrophe would only happen if all people decided to take the risk and interact, whereas those who don’t stay home and don’t die.
In any case, the deadlier a disease, the easier persuasion should be.
No, if anything, these birds are smart, not intelligent. Smart means they contain sophisticated knowledge, intelligence means they can create knowledge. And cars are way too recent a phenomenon to have genetically encoded knowledge of in genes.
Yeah, I could see some groups of inner-species animals having some limited culture based on very simple memes that are spread non-creatively.
Hmm I don’t recall, and I think it’s rare for strongly held beliefs and in the middle of a conversation. I’m guessing strongly held beliefs usually take time to change.
Now if only they used their brilliant minds to work on the field's lost goal: understanding how the human mind works.
Bad epistemology, sir. Data don’t support theories.
Presence of sophisticated knowledge != intelligence.
Presence of sophisticated knowledge != intelligence.
Note also how in many of these videos what we don’t see is that the filmed results are preceded by a loooong training phase (akin to the training phases in narrow AI) in which humans reinforce good and punish bad behavior.
Reinforcement “learning” != creative learning. Note how she praises the dog. The error correction happens in the owner, that in the dog. The owner merely passes it on to the dog.
Mimicking behavior does not require intelligence.
He’s trying to see if the cone is edible. No intelligence required for that.
I enjoy the rare occasions when people openly and readily change their minds. t.co/WdSICI43FE
Reminds me of Cameron Hanes’ line “Nobody cares, work harder.”
Yup exactly, that’s what I meant with the original tweet.
Exactly. That realization is what’s missing in the whole “are animals intelligent” debate. Presence of knowledge != intelligence.
Once again, the behavior in this video can be explained entirely in terms of inborn, genetically inherited knowledge. No intelligence on the part of the worm required.
@ModelsofMind @CNC3P0 @SamHarrisOrg
And the fact that people disagree about what should be done implies there is a problem to be solved, and, thereby, knowledge to be created. Which is better than just steamrolling over those who disagree.
@ModelsofMind @CNC3P0 @SamHarrisOrg
Preferences are ideas, too. And ignoring them can lead to coercion.
What I was trying to say is that people need to be persuaded if we want them to do something.
Just caught myself writing: "the fact that... shows..." Then I corrected it to: "the reason that... is..."
The first is inductive, authoritarian, and not an explanation, but a statement. The second is not inductive, leaves room for having made a mistake, and is an explanation.
The bird is pecking at the hog, maybe because the hog smells like food. The bird probably has no idea what a road is or what cars are or that they can kill animals.
It’s pecking at it. It’s probably just trying to get food out of it. It has no idea what forces are.
Not everyone agrees on what needs to be done. Nor should children be told what to do. They're full-fledged people, too.
Plus they have a universally applicable schema so that they can fit any data model. They also support a powerful query language (datalog).
I think they make real-time sync easier than other approaches. Oh, and they don't make you code in strings—db queries are data—which is a major plus for me.
Datomic in the backend and datascript in the frontend because it makes real-time syncing a walk in the park.
...and doesn’t realize it’s stuck and keeps repeating the same action over and over.
I don’t think rationality is selected by biological evolution, it that’s what you mean. People have to learn how to be rational (and are never done learning that).
"We need another lockdown." How about not telling people what to do?
Though of course, that is still granting agency on the part of the elephant, which it doesn't have. I'm just trying to draw attention to the fact that its "motivation" need not have been to save the human.
Another possible caption: "Ooh, looks like it's a human, I've had fun playing with them before and they give me treats. I'll go play with that one again, maybe I'll get treats again this time!"
Correction:
"Any time we're impressed by what non-human animals do, it's simply because we forget that they have genetically inherited sophisticated knowledge, which was created by biological evolution, not by those animals."
Common error: mistaking the sophistication of an animal's knowledge with the animal being intelligent.
For the former, the animal need only contain sophisticated knowledge—never mind where that knowledge came from—for the latter, it must have created the knowledge itself. twitter.com/neiltyson/stat…
Not sure either labels strong/weak or mind virus apply to either of those theories.
But generally speaking, if most people have Newton's idea of gravity, it's because his meme is better at spreading than Einstein's.
From here, it's just one more step to meeting a necessary condition for AGI: making it possible for the program to correct its own errors.
Applied to programming, it means we shouldn't judge a program by how well it solves a given problem, but by how easy it makes it to detect and correct errors it already contains.
Similarly, in science, we do not judge institutions by how much they produce or entrench good theories, but by how easy they make it to criticize and replace bad theories (ibid).
How do we get from Popper to AGI? 🧵👇
Popper's criterion of democracy entails that we shouldn't judge a political system by whether it produces good policies or leaders, but by how easy it makes it "to remove bad ones that are already there." — The Beginning of Infinity, ch. 9