>you'd never be able to know for sure that thats what you've done.
Words mean what they're defined to mean. Talking about "general intelligence" without a clear definition is just woo, muddy thinking that achieves nothing. A fundamental tenet of the scientific method is that only testable claims are meaningful claims.
The people who go around saying "LLMs aren't intelligent" while refusing to define exactly what they mean by intelligence (and hence not making a meaningful/testable claim) add nothing to the conversation.
I'll happily say that LLMs aren't intelligent, and I'll give you a testable version of it.
An LLM cannot be placed in a simulated universe, with an internally consistent physics system of which it knows nothing, and go from its initial state to a world-spanning civilization that understands and exploits a significant amount of the physics available to it.
I know that is true because if you place an LLM in such a universe, it's just a gigantic matrix of numbers that doesn't do anything. It's no more or less intelligent than the number 3 I just wrote on a piece of paper.
You can go further than that and provide the LLM with the ability to request sensory input from its universe and it's still not intelligent because it won't do that, it will just be a gigantic matrix of numbers that doesn't do anything.
To make it do anything in that universe you would have to provide it with intrinsic motivations and a continuous run loop, but that's not really enough because it's still a static system.
To really bootstrap it into intelligence you'd need to have it start with a very basic set of motivations that it's allowed to modify, and show that it can take that starting condition and grow beyond them.
You will almost immediately run into the problem that LLMs can't learn beyond their context window, because they're not intelligent. Every time they run a "thought" they have to be reminded of every piece of information they previously read/wrote since their training data was fixed in a matrix.
I don't mean to downplay the incredible human achievement of reaching a point in computing where we can take the sum total of human knowledge and process it into a set of probabilities that can regurgitate the most likely response to a given input, but it's not intelligence. Us going from flint tools to semiconductors, vaccines and spaceships, is intelligence. The current architectures of LLMs are fundamentally incapable of that sort of thing. They're a useful substitute for intelligence in a growing number of situations, but they don't fundamentally solve problems, they just produce whatever their matrix determines is the most probable response to a given input.
You could ask Gemini 2.5 to do that today and it's well within its capabilities, just as long as you also let it write and run unit tests, as a human developer would.
AGI isn't ASI; it's not supposed to be smarter than humans. The people who say AGI is far away are unscientific woo-mongers, because they never give a concrete, empirically measurable definition of AGI. The closest we have is Humanity's Last Exam, which LLMs are already well on the path to acing.
Consider this: Being born/trained in 1900 if that were possible and given a year to adapt to the world of 2025, how well would an LLM do on any test? Compare that to how a 15 years old human in the same situation would do.
I'd expect it to be generalised, where we (and everything else we've ever met) are specialised. Our intelligence is shaped by our biology and our environment; the limitations on our thinking are themselves concepts the best of us can barely glimpse. Some kind of intelligence that inherently transcends its substrate.
What that would look like, how it would think, the kind of mental considerations it would have, I do not know. I do suspect that declaring something that thinks like us would have "general intelligence" to be a symptom of our limited thinking.
2B models by themselves aren't so useful, but it's very interesting as a proof of concept, because the same technique used to train a 200B model could produce one that's much more efficient (cheaper and more environmentally friendly) than existing 200B models, especially with specialised hardware support.
That's how the current system works too: newly created money goes through the banks, and so the ones who spend it first (and benefit from seigniorage) are the financial industry and its customers: corporations borrowing to invest in productive activity. The key thing to note is that this system (and Ford's) represents a system of continuous wealth transfer from the working class (who hold most of their money in cash or low interest savings accounts, and suffer the most from inflation) to the owners of the means of production.
>A very asymmetric setup exposing the underbelly of free speech cultures
As opposed to non-free-speech cultures like Russia and China where people have absolutely no say in whatever their leaders do? Because that's inevitably what happens when you give people in power the power to restrict speech: they restrict any speech critical of them. We're even seeing this in developed democracies like Germany where a journalist was recently fined for posting a meme online of a politician holding a sign saying "I hate free speech".
It's important to be precise because everything is not the same. In the German case the ruling was not because someone posted a critical meme, but because it was not entirely obvious the picture was edited (as in: you and I can immediately see the photo was edited, but some people will not recognize the edit). I do not agree with the ruling, but as a citizen I am happy that in Germany we still care if claims are true or not (and try to prevent people from lying).
Does this happiness that some people care whether claims are true or not overrule the arrest and deportation of peaceful protestors, and people in general based on social media posts, or do you also feel happy about that?
Isn't that just relying on the stupidity of someone who may not exist?
Like every single year people make the same dumb joke "Republicans vote on Tuesday and Democrats vote on Wednesday" leading to prosecution for misinformation when they cannot prove anyone actually tried voting on the wrong day because of the meme
Once you start talking millions of people someone will make that or any other mistake.
The US has a higher threshold, but it’s clear those standards mean many people are duped by “obvious” lies. It’s kind of an arbitrary line, but ignoring the dumb feels like a mistake to me when dumb people are active in society, still vote, etc.
Whataboutism does not change anything about it being a weak spot. I’m only saying the free speech west can’t use the same tactic against these kind of adversaries because they’re insulated against them.
Free speech including paid speech isn’t really a knock on free speech.
Someone can be persuaded by an argument they heard once, but can’t per persuaded by an argument they never hear. Thus blocking speech by preventing any kind of speech including paid speech is problematic.
They are noise generators with a goal of raising the noise floor above the pain threshold, in essence they’re using free speech to shut down free speech itself.
I’d say the same about social media. However IMO the value of free speech isn’t in having a clear message to directly improve things, the value is being able to steal ideas from anyone. “Obamacare” was originally a Republican idea, but once an idea is out there anyone can take it.
Ideas don’t need to win on day one, if it takes 30 years that’s still plenty useful.
Kids who grow up watching commercials start distrusting them. Free speech is not about any one issue but all topics. In many ways curating so people see the kinds of things they agree with is vastly more harmful than propaganda.
1.The majority of people are not intelligent. Source is polls on whether there was wide spread election fraud
2. Politicans want money and power. They have no issues lying or manipulating people to get it
3. In a country like Russia the government can counter any information with widespread arrests and fear.
4. In a country with free speech there is little to no recourse.
Meaning that Russia, China, etc can use misinformation against us and we can't do anything. On the other hand we can try the same but they can simply use authoritarian tactics to supress it.
5. Trump has shown that the threshold for lying was set artificially low by past politicians. His success while lying about events that are easily disproven multiple times is evidence for all future politicians to lie.
“The number of people overall who believe the election was fraudulent has hovered around 35% since November 2020, but this percentage has not increased significantly as the claim purports.” https://www.politifact.com/factchecks/2022/feb/02/viral-imag...
Which is different than asking if voting fraud changed the outcome and more importantly different than asking which side benefited. Someone who calls targeted overly aggressive culling of voter registrations fraud has something of a point, even if that’s a long way from stuffing ballots or meaningful changes in results.
I don't think taking all political affiliations into account makes sense. Let me use another poll that had a similar outcome of your poll for all political affiliations:
A 2023 poll found that 71% of Republicans believe the election was illegitimate. [1]. The exact question in the poll was "Thinking about the results of the 2020 presidential election, do you think that Joe Biden legitimately won enough votes to win the presidency, or not? Do you think there's been solid evidence of that, or is that your suspicion only?"
All - Note legitimate Solid + suspicious = 38%
Republican - Not legitimate: solid evidence - 41% suspicious only - 30%
1. Democrats or liberals (poll allowed for either) who didn't vote for Trump or dislike him are going to say the election was legitimate regardless of evidence and outcomes of investigations. This is why I only use what Republican voters think (about 2020) as an indicator of public stupidity *
2. This poll was in 2023, after court cases and numerous state investigations/recounts. Therefore saying it's "suspicious" is as stupid as saying there is "solid evidence".
If you have a suspicion a crime occurred, then multiple investigations find nothing or show the evidence your suspicious were based were fake, and you don't change your view that's stupid.
> Which is different than asking if voting fraud changed the outcome...
That's what Trump and many of the key players on his side claimed.
> Someone who calls targeted overly aggressive culling of voter registrations fraud has something of a point..
No, they don't. They are misusing the term "fraud" in an election situation (a.k.a "election fraud) [2]. Voter/Election fraud is clearly defined by the US government [3]. Voter suppression through a legal action isn't fraud. You can claim that it's "wrong" or "immoral" but not fraud.
The difference is clear if you look at something as either an opinion or fact. An opinion is not falsifiable.
"Widespread election fraud is why Trump lost the 2020 election" - This either happened or it didn't. It's not an opinion/judgement. [4]
"Aggressive culling of registrations caused a candidate to win/lose" - Since culling of registrations legally happens [5] whether or not it's aggressive is a judgement because "aggressiveness" is subjective.
> even if that’s a long way
It's not on the same scale because one is a crime. I think I need more to understand why you want to merge different accusations of fraud or suppression when discussing different elections.
[2] Wikipedia's article on Election fraud describes it better.
"Electoral fraud, sometimes referred to as election manipulation, voter fraud, or vote rigging, involves illegal interference with the process of an election, either by increasing the vote share of a favored candidate, depressing the vote share of rival candidates, or both. It differs from but often goes hand-in-hand with voter suppression. "
[4] You can say "I believe X happened" which is an opinion however this is a judgement that needs a factual base. If the evidence is fake, doesn't exist, or you were lied and you are aware of this, then you're lying about the basis for your opinion which invalidates it (imo)
[5] I'm assuming you meant legal culling
* There's similar high numbers for Democrats talking about Trump's win in 2016 though most polls ask about Russian interference helping him, which is a judgement not a lie since this did happen, but it could also be an indicator. The 2020 situation was just much more obvious because the claim by Trump is of cheating NOT influence. The lie is that Trump was directly involved and to a high degree but blah blah complicated.
> Therefore saying it's "suspicious" is as stupid as saying there is "solid evidence".
Hardly, I find quantum mechanics suspect without having a better option. I’m not saying there’s any kind of conspiracy or anything and sure it fits the experiments we have done. Yet, I suspect most people who actually learn the details have similar reactions it doesn’t fit our experience. Sadly the universe doesn’t care it it seems consistent to us.
There’s a deep cultural divide in the US to the point where people have trouble remembering how close support is for each party. Because politics is so regional it’s easy for each side to overestimate how popular that side is. Imagine living in a county where 80% are voting for one side and almost all roadside posters are supporting one candidate. Suddenly the other side winning just doesn’t fit everyday experience.
When either side wins a huge number of people will find it suspicious, that’s just how our heuristics and pattern matching work. A historian looking back on 2020 and 2024 isn’t going to find the election results odd because wider forces definitely favored the winning side in those elections, but people today don’t have that separation. Thinking there’s widespread and obvious fraud is different.
Everyone doesn’t fit those criteria. Motivated reasoning exists with and without propaganda. The specific words used may end up mimicking “a message,” but you can find millions of disgruntled people after any election.
There’s a great deal of talk around how much social media etc changed the landscape but American politics looks basically the same before and after Facebook.
>American politics looks basically the same before and after Facebook.
Trump and MAGA Republicans lie more openly than traditional Republicans. American politics are not the same.
Here's an easy exercise.
This is a post from MGT. Show me anything even close to this insane from an elected person in high office (house, senate, president) in the last 50 years
That’s fairly mild, but it’s not so easy to link to 20+ year old clips. There’s some real bangers of homophobic rants in your time frame, but you may be a little young to remember any of them.
In terms of lies here’s one that was a central tenant of the part of the party line for decades. Social security isn’t an income tax because we have a tax called the income tax.
>but you may be a little young to remember any of them.
I'm 45
> Social security isn’t an income tax because we have a tax called the income tax.
I took too long to be able to edit my other comment but I should have asked who said this? Because "social security" isn't a tax at all. In my other comment I assumed you meant tax we pay to fund SS but this still leaves me confused, can you provide me with a quote that shows the lie?
> Homophobia isn't lying, it's an opinion/ judgement
The lies about gay people are actual lies, even if they come from homophobia.
Luring people into gayness, crap about destroying the institution of marriage etc etc.
> SS tax is a tax on your income, "Income Tax" is a type of tax. Both are true
When taking about “tax burden,” there’s no excuse around the names. I’ve got little interest in digging up 20 year old clips, but you’re 45 you should remember that phrase.
your trust scale needs more dynamic range- the sackler fiasco genuinely should have bumped everyone’s trust in doctors a lot, but probably should not have bumped them below supplement peddling minecraft youtubers.
Not a fan of Google, but if you use Gemini through AI studio with a custom prompt and filters disabled it's by far the least censored commercial model in my experience.
Words mean what they're defined to mean. Talking about "general intelligence" without a clear definition is just woo, muddy thinking that achieves nothing. A fundamental tenet of the scientific method is that only testable claims are meaningful claims.
reply