• Welcome to the Internet Infidels Discussion Board.

Artificial intelligence paradigm shift

It's been few years since they claimed that computer speech recognition is better than human
Better than which human?

Because average human speech recognition can be pretty shit, too.

It is clearly better than about 100% of other 5 year olds.
 
No, you just didn't actually make a valid argument. Some short years ago, CG appeared in animation, looking like sloppy shit,
Are you really making comparison between CG and AI?
AI is a literal 5 year old.
And it is still mostly unsubstantiated hype.
Seriously. It's been few years since they claimed that computer speech recognition is better than human. And yet, when you try to use it it clearly not better than human.

If barbos is using English speech recognition to understand a thick Russian accent, then it probably does do poorly. However, the trick to speech recognition is a program that relies on context to guess the most likely words being pronounced, not just to take the acoustic input at face value. People tend to produce speech with a consistent accent, but speech understanding requires the ability to handle a wide variety of phonetic variation in the speech signal, not to mention morphological and syntactic variation. That is, one can understand accents that one cannot accurately reproduce. So the development of AI, especially including LLMs, has vastly improved the ability of computer programs to transcribe speech. They are pretty good these days at realtime closed captioning of speech in recorded media (for example, on Youtube). Combined with AI-assisted translation, you can even get some reasonably good translations of foreign videos, depending on the language being translated into English. I've been studying machine translation techniques for a long time, and I am really impressed with the quality of translation I am seeing now. I still see a lot of errors, but the translations are serviceable even when there are a lot of errors.
 
context to guess the most likely words being pronounced, not just to take the acoustic input at face value
Humans have to do as much, as well. Have you ever tried to follow a conversation in Japanese? They roll out the pronouns and context once at the beginning and if you miss it, forget it, or can't match it up at any point, the whole conversation melts into apparent gibberish for the listener.

"Context heavy" I heard it called once?
 
Several of my coworkers have started using AI in a way that can be summarised as, "outsourcing the reason you have a job".

A couple of programmers use ChatGPT etc. to generate big pieces of code, including complex business logic, and then post it for review without understanding it. This naturally has negative consequences for the programmers, because they get found out when reviewers see their mistakes and ask them to explain it, or when they can't fix the bugs in their plagiarised code. It's like copy-pasting from Stack Overflow, but with extra brainrot.

One programmer said to me, "I put the code into ChatGPT and it told me to change [X] to [Y]. I tested it and it works. Is this solution OK?"
While I appreciated their candour, I was not impressed that they did not understand the solution.

A couple of programmers uses ChatGPT etc. to write documentation for code. This is blatantly obvious because ChatGPT has excellent English while the programmers do not, but the programmers make it even more obvious by just pasting big blocks of irrelevant fluff. Their AI-generated written communication looks better on a superficial level, but is far less effective than their own messy but purposeful writing.

The thing is, these programmers get hired precisely because they are people who have deep knowledge about software systems and have the soft skills to make them effective in a cross-functional team. And as with a lot of well-paying knowledge work, we have to keep our knowledge fresh and our skills sharp if we want job security. By delegating to AI, they are basically signalling to their employer that they are interchangeable with any junior programmer who can write English well enough to prompt a chat bot.

ChatGPT, Claude, Copilot etc. are suggestion engines. They provide suggestions which may or may not be correct. These virtual assistant have no idea what's they're doing yet present every answer with 100% confidence.The user, especially if they are a knowledge worker, must validate these suggestions before using them. They are very useful assistants when used correctly; they are brain-eating prions when not.
Exactly
 
Better than which human?
average apparently..
Because average human speech recognition can be pretty shit, too.
Average is average, what you are talking about is random.
It is clearly better than about 100% of other 5 year olds.
Why do you keep making this ridiculous comparison?
Because AI systems can and do and are being constantly tuned and developed and it's absolutely idiotic to look at the acceleration from "can barely step together a sentence" to "can code better than Barbos" in 5 years. This implies a velocity, and your view is looking at the state of the earth and saying that it will not continue advancing as such.

Even if it were decelerating (it's not, yet, apparently), it's still got a much higher current velocity of change than any living human, without any significant or apparent need for major internal architectural changes, beyond things which expand the usefulness and depth of their contexts.
 
Because AI systems can and do and are being constantly tuned and developed and it's absolutely idiotic to look at the acceleration from "can barely step together a sentence" to "can code better than Barbos" in 5 years. This implies a velocity, and your view is looking at the state of the earth and saying that it will not continue advancing as such.
But can it really code at all?
Or it simply create an impression of coding by basically taking code from the internet which seems appropriate as a solution?


It can not code at any useful level, it's good for searching for useful snippets of code. It does not think any sense of the word.
It's just very complicated search engine trained to make appearance of intelligence.

You bought into this hype where they say AI is better at driving, better at speech recognition. When in reality it's not even close. It's useful as a search engine in fields where large amount of info need to be operated on. It can generate better than nothing subtitles. It can translate simple news text. But that does not imply thinking even at 5 year old level.
 
Last edited:
It was said that computers tends to be good at things humans are bad at and vice versa.
AI is an attempt to make computers good at what humans are good and so far, it does not look as good as they promised.
 
This is just another "lookup table" or "collage" hand wave.
Prove me wrong.
How much info a typical 5 year old is fed ? 5 books and a bunch of cartoons maybe?
How much info ChatGPT consumed? Last time I heard they said that they ran out of available data for training.

And yet. 5 year old is better at pretty much all human tasks than ChatGPT.
 
Last edited:
I tried to use one of these cheap ChatGPTs to ask some programming question. It gave me useless answer and then I simply googled it and found out where ChatGPT got its answer from ..... it was stack overflow, pretty much verbatim.
Its answers are useless, they are very generic and its "understanding" of the problem is supercilious and search-engine-like.
 
Take for example SpaceX and their mind boggling reusable rocket landing. No human will be able to do that manually.
But computer program (not very sophisticated) can do that without problem. It's just simple Newtonian physics.
But pretty much any human can learn to drive a car in real world settings, whereas computers simply can not do that at all.
It can do it only where there is no uncertainty on the road. Piece of trash on the road and it won't be able to distinguish it from a human, the end. Search engines are great but they do not think.
 
Last edited:
"Context heavy" I heard it called once?
Google speech uses it a lot sometimes overdoing to ridiculous level.
They have to, in order to hide their poor ability to recognize separate words spoken in a perfect midwestern american english. To be fair, their understanding of russian is perfect. but russian phonetics is very simple and very redundant, no need to use context.
 
This is just another "lookup table" or "collage" hand wave.
Prove me wrong.
How much info a typical 5 year old is fed ? 5 books and a bunch of cartoons maybe?
How much info ChatGPT consumed? Last time I heard they said that they ran out of available data for training.

And yet. 5 year old is better at pretty much all human tasks than ChatGPT.
The impossibility of the claim has been widely discussed across the internet. We have no need to reproduce the discussion here other than that such claims are in fact a wide source of mockery for the people who make them elsewhere.

I'm not going to deal with you playing the quisling.
 
This is just another "lookup table" or "collage" hand wave.
Prove me wrong.
How much info a typical 5 year old is fed ? 5 books and a bunch of cartoons maybe?
How much info ChatGPT consumed? Last time I heard they said that they ran out of available data for training.

And yet. 5 year old is better at pretty much all human tasks than ChatGPT.
The impossibility of the claim has been widely discussed across the internet. We have no need to reproduce the discussion here other than that such claims are in fact a wide source of mockery for the people who make them elsewhere.

I'm not going to deal with you playing the quisling.
I see, resorting to Reductio ad Hitlerum must have been part of your AI training.
 
Last edited:
context to guess the most likely words being pronounced, not just to take the acoustic input at face value
Humans have to do as much, as well. Have you ever tried to follow a conversation in Japanese? They roll out the pronouns and context once at the beginning and if you miss it, forget it, or can't match it up at any point, the whole conversation melts into apparent gibberish for the listener.

"Context heavy" I heard it called once?

Japanese is a verb-final language. I speak quite a few languages like that, including several (like German) that are sort of in between. I am familiar with the type of clitic pronouns that you are referring to, but you can mess around with their placement without having such a devastating effect on listeners. There is a term "context-dependent" when speaking about syntactic systems, but it is limited primarily to syntactic theories that tend to focus on isolated expressions rather than extended discourse.

In normal conversation, there are other contextual modalities--vision, touch, smell, etc.--that can affect the interpretation of a spoken string of words. A theory of linguistic discourse has to take a broader perspective on the nature of linguistic context. It is grounded in shared experiences that are not part of the linguistic text. LLMs are exclusively grounded in text, because those programs are developed independently of sensory experiences that are grounded in bodies. They have no need of conscious awareness of their surroundings. The situation is different in the field of robotics, where one is working with programs that control moving bodies that interact with chaotic environments. In that field, there is considerably more interest in the nature of philosophical concepts such as consciousness, awareness, and volition.
 
shared experiences that are not part of the linguistic text
I find that this view takes an overly narrow view of linguistics.

Essentially, each of the components of some experience is itself some phoneme that sums to the overall vector of some communication, and some of it can be baked into intrinsic.

When someone says "I didn't really say..." I can say "no, you did say; these actions and inflection form a statement as clear as any verbal word."

Sometimes the words are difficult to see, and are mediated with other than "sound"? But all things amount to language imputed meaning either indirectly or directly.
 
A few months ago there was this:
Along the way, OpenAI is watching through the ChatGPT interface, and the company is reportedly coming down hard on any attempts to probe o1's reasoning, even among the merely curious.
One X user reported (confirmed by others, including Scale AI prompt engineer Riley Goodside) that they received a warning email if they used the term "reasoning trace" in conversation with o1. Others say the warning is triggered simply by asking ChatGPT about the model's "reasoning" at all.
But now there's the open source Chinese AI DeepSeek which allows you to see its reasoning process (DeepThink R1).
Apparently DeepSeek V3 is the best or close to the best in benchmarks/metrics.
It also costs a lot less to use and train compared to US AIs.
BTW here is an example of its reasoning (R1 enabled)
what's the most humorous chess opening move?
Thought for 53 seconds

Okay, so I need to figure out what the most humorous chess opening move is. Hmm, let's start by recalling some chess openings. The first move in chess is usually either 1.e4 or 1.d4, sometimes 1.c4 or 1.Nf3. Those are the main ones. But the question is about humorous ones. So maybe something unconventional or tricky?

I remember hearing about the Grob's Attack, which is 1.g4. That's a pretty aggressive and unusual opening. The idea is to push the g-pawn early, which can surprise the opponent. But is it considered humorous? Maybe because it's so rare and looks a bit silly to just push that pawn forward.

Then there's the Barnes Opening, 1.f3. That's kind of a weak move because it doesn't control the center and blocks the knight. I think it's considered one of the worst opening moves. But maybe people joke about it because it's so bad. There's also the Durkin's Opening, 1.Na3, which is called the Sodium Attack because Na is the chemical symbol for sodium. That's a pun, so that's a bit humorous.

Wait, there's also the Fool's Mate, which is the fastest possible checkmate. It's not an opening move per se, but if someone tries to set it up, they might do something like 1.f3 and then 2.g4, leading to mate in two moves. But that's more of a trap than an opening.

The Bongcloud Attack is another one. That's when you move the king's pawn forward and then bring the king out, like 1.e4 e5 2.Ke2. It's a joke opening because moving the king early is terrible, but some people do it for laughs. Maybe that's considered humorous. It's become a meme in online chess communities.

Another possibility is the Hippopotamus Defense, where you develop all your pieces to the third rank and create a solid but passive structure. It's not a specific opening move, but the setup is kind of funny because it looks like a hippo.

Or the Orangutan Opening, 1.b4. The story goes that a chess master played this move after seeing an orangutan at the zoo, hence the name. The move itself is offbeat and leads to unusual positions, so the name adds some humor.

So, putting this together, the most humorous might be the Bongcloud because it's intentionally bad and a meme. Alternatively, the Sodium Attack because of the pun. The Grob's Attack and Barnes Opening are also contenders because they're so unconventional. Maybe the answer is the Bongcloud Attack since it's widely recognized in the chess community as a joke. But the user is asking for the opening move, not necessarily a sequence. The Bongcloud starts with 1.e4, then 2.Ke2. So the first move is still normal. Maybe the first move itself isn't humorous, but the follow-up is. So perhaps the most humorous opening move is 1.g4 (Grob's) or 1.b4 (Orangutan) because they are unusual and have funny names or stories. The Sodium Attack (1.Na3) is a pun, so that's a good candidate too.

Wait, also, there's 1.h4, which is called the Desprez Opening. It's similar to Grob's but on the h-file. It's not very effective and looks funny. So maybe that's another option. But which of these is considered the most humorous? I think the Bongcloud is more of a meme in recent times, especially with streamers playing it for laughs. But strictly the first move, maybe Grob's or Sodium Attack. Alternatively, the Fool's Mate attempt with 1.f3, but that's more about the trap. I need to check which of these openings are commonly cited as humorous.

I think the Sodium Attack (1.Na3) is a good one because of the pun. Also, 1.a3, the Anderssen's Opening, is sometimes called the "Creepy Crawly" opening because it just moves the a-pawn one square. But maybe that's stretching it.

In online forums and chess circles, the Bongcloud is definitely a joke opening. But again, the first move is 1.e4, which is standard. The humor comes in the second move. So if we're strictly talking about the first move, maybe 1.Na3 (Sodium) or 1.g4 (Grob's) are the funniest because they're so odd. The Orangutan (1.b4) has a funny story. So maybe the answer is the Orangutan Opening because of the name's origin. Alternatively, the Bongcloud is a modern meme.

I think the Bongcloud is more widely recognized as humorous in today's context, especially with Hikaru Nakamura and others popularizing it as a joke. But the initial move is 1.e4, which is normal. The humor is in the follow-up. So perhaps the most humorous single opening move is 1.g4 (Grob's Attack) because it's so unexpected and looks ridiculous. Alternatively, 1.h4 is equally funny.

Another thought: The "Monkey's Bum" opening, which I think is 1.f3, but that might not be an official name. Or the "Duras Gambit" where white plays 1.e4 and then 2.Ke2, but again, that's two moves.

I think the consensus might be that the Bongcloud is the most humorous, even though it's two moves. But the question is about the opening move. If it's just the first move, then maybe 1.g4 or 1.b4. Alternatively, 1.Nh3, the "Pickering Van", which is another odd knight move.

In conclusion, I think the most humorous chess opening move is either the Grob's Attack (1.g4) for its shock value and rarity, the Sodium Attack (1.Na3) for the pun, or the Orangutan (1.b4) for the story. The Bongcloud is funny but requires two moves. Since the question is about the move itself, not the sequence, I'd go with either 1.g4 or 1.b4. The Sodium Attack is a good pun, which adds humor. Maybe the answer is the Sodium Attack (1.Na3) because of the chemical symbol joke.
The most humorous chess opening move is often considered to be the Sodium Attack (1.Na3), primarily due to its punny name derived from the chemical symbol for sodium (Na). This move, which develops the knight to the edge of the board, is strategically dubious but earns laughs for its clever wordplay.

Other contenders include:

Grob’s Attack (1.g4): Known as the "Spike" or "Ahlhausen Opening," it’s humorously audacious, flouting opening principles by advancing the g-pawn early.

Orangutan Opening (1.b4): Named after a chess master’s zoo visit, its quirky backstory and unorthodox pawn push add whimsy.

Bongcloud Attack (1.e4 followed by 2.Ke2): A modern meme favorite, though the humor peaks on the second move when the king strolls out prematurely.

For pure first-move hilarity, 1.Na3 edges out with its nerd-chic chemistry joke, while 1.g4 and 1.b4 delight in their sheer absurdity. The Bongcloud, though iconic, requires a follow-up move to fully land the punchline. 🏁♟️
The open source DeepSeek can also be run locally on your computer.
 
Last edited:
A few months ago there was this:
Along the way, OpenAI is watching through the ChatGPT interface, and the company is reportedly coming down hard on any attempts to probe o1's reasoning, even among the merely curious.
One X user reported (confirmed by others, including Scale AI prompt engineer Riley Goodside) that they received a warning email if they used the term "reasoning trace" in conversation with o1. Others say the warning is triggered simply by asking ChatGPT about the model's "reasoning" at all.
But now there's the open source Chinese AI DeepSeek which allows you to see its reasoning process (DeepThink R1).
Apparently DeepSeek V3 is the best or close to the best in benchmarks/metrics.
It also costs a lot less to use and train compared to US AIs.
BTW here is an example of its reasoning (R1 enabled)
what's the most humorous chess opening move?
Thought for 53 seconds

Okay, so I need to figure out what the most humorous chess opening move is. Hmm, let's start by recalling some chess openings. The first move in chess is usually either 1.e4 or 1.d4, sometimes 1.c4 or 1.Nf3. Those are the main ones. But the question is about humorous ones. So maybe something unconventional or tricky?

I remember hearing about the Grob's Attack, which is 1.g4. That's a pretty aggressive and unusual opening. The idea is to push the g-pawn early, which can surprise the opponent. But is it considered humorous? Maybe because it's so rare and looks a bit silly to just push that pawn forward.

Then there's the Barnes Opening, 1.f3. That's kind of a weak move because it doesn't control the center and blocks the knight. I think it's considered one of the worst opening moves. But maybe people joke about it because it's so bad. There's also the Durkin's Opening, 1.Na3, which is called the Sodium Attack because Na is the chemical symbol for sodium. That's a pun, so that's a bit humorous.

Wait, there's also the Fool's Mate, which is the fastest possible checkmate. It's not an opening move per se, but if someone tries to set it up, they might do something like 1.f3 and then 2.g4, leading to mate in two moves. But that's more of a trap than an opening.

The Bongcloud Attack is another one. That's when you move the king's pawn forward and then bring the king out, like 1.e4 e5 2.Ke2. It's a joke opening because moving the king early is terrible, but some people do it for laughs. Maybe that's considered humorous. It's become a meme in online chess communities.

Another possibility is the Hippopotamus Defense, where you develop all your pieces to the third rank and create a solid but passive structure. It's not a specific opening move, but the setup is kind of funny because it looks like a hippo.

Or the Orangutan Opening, 1.b4. The story goes that a chess master played this move after seeing an orangutan at the zoo, hence the name. The move itself is offbeat and leads to unusual positions, so the name adds some humor.

So, putting this together, the most humorous might be the Bongcloud because it's intentionally bad and a meme. Alternatively, the Sodium Attack because of the pun. The Grob's Attack and Barnes Opening are also contenders because they're so unconventional. Maybe the answer is the Bongcloud Attack since it's widely recognized in the chess community as a joke. But the user is asking for the opening move, not necessarily a sequence. The Bongcloud starts with 1.e4, then 2.Ke2. So the first move is still normal. Maybe the first move itself isn't humorous, but the follow-up is. So perhaps the most humorous opening move is 1.g4 (Grob's) or 1.b4 (Orangutan) because they are unusual and have funny names or stories. The Sodium Attack (1.Na3) is a pun, so that's a good candidate too.

Wait, also, there's 1.h4, which is called the Desprez Opening. It's similar to Grob's but on the h-file. It's not very effective and looks funny. So maybe that's another option. But which of these is considered the most humorous? I think the Bongcloud is more of a meme in recent times, especially with streamers playing it for laughs. But strictly the first move, maybe Grob's or Sodium Attack. Alternatively, the Fool's Mate attempt with 1.f3, but that's more about the trap. I need to check which of these openings are commonly cited as humorous.

I think the Sodium Attack (1.Na3) is a good one because of the pun. Also, 1.a3, the Anderssen's Opening, is sometimes called the "Creepy Crawly" opening because it just moves the a-pawn one square. But maybe that's stretching it.

In online forums and chess circles, the Bongcloud is definitely a joke opening. But again, the first move is 1.e4, which is standard. The humor comes in the second move. So if we're strictly talking about the first move, maybe 1.Na3 (Sodium) or 1.g4 (Grob's) are the funniest because they're so odd. The Orangutan (1.b4) has a funny story. So maybe the answer is the Orangutan Opening because of the name's origin. Alternatively, the Bongcloud is a modern meme.

I think the Bongcloud is more widely recognized as humorous in today's context, especially with Hikaru Nakamura and others popularizing it as a joke. But the initial move is 1.e4, which is normal. The humor is in the follow-up. So perhaps the most humorous single opening move is 1.g4 (Grob's Attack) because it's so unexpected and looks ridiculous. Alternatively, 1.h4 is equally funny.

Another thought: The "Monkey's Bum" opening, which I think is 1.f3, but that might not be an official name. Or the "Duras Gambit" where white plays 1.e4 and then 2.Ke2, but again, that's two moves.

I think the consensus might be that the Bongcloud is the most humorous, even though it's two moves. But the question is about the opening move. If it's just the first move, then maybe 1.g4 or 1.b4. Alternatively, 1.Nh3, the "Pickering Van", which is another odd knight move.

In conclusion, I think the most humorous chess opening move is either the Grob's Attack (1.g4) for its shock value and rarity, the Sodium Attack (1.Na3) for the pun, or the Orangutan (1.b4) for the story. The Bongcloud is funny but requires two moves. Since the question is about the move itself, not the sequence, I'd go with either 1.g4 or 1.b4. The Sodium Attack is a good pun, which adds humor. Maybe the answer is the Sodium Attack (1.Na3) because of the chemical symbol joke.
The most humorous chess opening move is often considered to be the Sodium Attack (1.Na3), primarily due to its punny name derived from the chemical symbol for sodium (Na). This move, which develops the knight to the edge of the board, is strategically dubious but earns laughs for its clever wordplay.

Other contenders include:

Grob’s Attack (1.g4): Known as the "Spike" or "Ahlhausen Opening," it’s humorously audacious, flouting opening principles by advancing the g-pawn early.

Orangutan Opening (1.b4): Named after a chess master’s zoo visit, its quirky backstory and unorthodox pawn push add whimsy.

Bongcloud Attack (1.e4 followed by 2.Ke2): A modern meme favorite, though the humor peaks on the second move when the king strolls out prematurely.

For pure first-move hilarity, 1.Na3 edges out with its nerd-chic chemistry joke, while 1.g4 and 1.b4 delight in their sheer absurdity. The Bongcloud, though iconic, requires a follow-up move to fully land the punchline. 🏁♟️
The open source DeepSeek can also be run locally on your computer.
Yeah. Honestly this is the reason I find the 'collage' and 'database' claims so ridiculous.

DeepSeek is offline and 2 gigabytes and the Internet is many, many terabytes of information.

Even Wikipedia, compressed, is bigger than a DeepSeek model, and Wikipedia is hardly comprehensive.

This kind of 'knowledgebase' function is only possible if the system is leveraging general, rather than 'tabled' understandings.
 
Yeah. Honestly this is the reason I find the 'collage' and 'database' claims so ridiculous.

DeepSeek is offline and 2 gigabytes and the Internet is many, many terabytes of information.

Even Wikipedia, compressed, is bigger than a DeepSeek model, and Wikipedia is hardly comprehensive.

This kind of 'knowledgebase' function is only possible if the system is leveraging general, rather than 'tabled' understandings.
There is a local version that is about 2 Gb but I think the version that did really well in the benchmarks is a lot bigger than that.
 
Back
Top Bottom