It's clear from this comment that you did not read the full article. If you did then you'd have seen that the author addresses this criticism you're making here.
I did read it. She doesn’t mention mathematics or RLVR training once, so I assume you’re referring to my point about empirical testability. Well, I think her statement that the claim “LLMs are stochastic parrots” is not an empirical claim is false, and she’s being disingenuous there with a classic motte-and-bailey fallacy. She quotes her own original paper thus:
> Text generated by an LM is not grounded in communicative intent, any model of the world, or any model of the reader’s state of mind. It can’t have been, because the training data never included sharing thoughts with a listener, nor does the machine have the ability to do that. This can seem counter-intuitive given the increasingly fluent qualities of automatically generated text, but we have to account for the fact that our perception of natural language text, regardless of how it was generated, is mediated by our own linguistic competence and our predisposition to interpret communicative acts as conveying coherent meaning and intent, whether or not they do [89, 140]. The problem is, if one side of the communication does not have meaning, then the comprehension of the implicit meaning is an illusion arising from our singular human understanding of language (independent of the model). Contrary to how it may seem when we observe its output, an LM is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning: a stochastic parrot.
Do you really think that claiming the output of an LLM “has no reference to meaning” is not an empirical claim? That it doesn’t attempt to place any bounds whatsoever on what LLMs can and cannot do? LLMs can solve some very difficult mathematical problems quite well now: see the article from Gowers that was on here recently. Do you think that the output in a situation like that “has no reference to meaning?” If so, you’ll have to explain why, because I don’t understand at all.
> still i suspect the largest barrier is merely that all the popular social media sites are actively captured by ad-driven development / leaders. That cant last forever, people are sick of it.
This is why it's a good idea to make the switch to federated alternatives like Lemmy/Piefed. The more people who do this the more people will see it as a viable alternative, making it easier to get away from the ad-driven model of social media.
It is modular and built to be adapted to the different needs of online communities. They also have a more fine-grained boundaries system that allows to post only to pre-defined groups (or Circles in Diaspora / Google+ lingo) etc. I don't think they have an Events module yet, but technically speaking, this could be implemented.
It's not necessarily the case that Loops is just as addictive as TikTok. Because TikTok is more than just short term videos. It's also a recommendation algorithm that slurps up as much information as it can about you to predict what you'll want to watch next. This recommendation algorithm plays a big role in making TikTok addictive. And as far as I'm aware, Loops does not have this functionality. It will just show you videos based on a much simpler algorithm that takes into account how recent a video is and how many likes/comments it has, or something like that, which will make it less addictive.
> The self-assuaging fantasy that g itself doesn't exist is a classic example of a psychological defensive mechanism of rejection, one rooted in a need to defend a worldview that holds all people as inherently equal, when we're measurably, biologically not.
I think this statement conflates two different senses of the word “equality”. Equality of abilities is different from moral equality. It is perfectly coherent to accept that people aren’t equal in terms of their abilities but are still morally equal. For example, just because Person A is smarter than Person B it does not follow that the interests of Person A matter more than those of Person B, or that the suffering of Person A matters more than the like suffering of Person B, etc. So the view that g is real and people have different IQ scores is consistent with the idea that all people are inherently equal. Because in most contexts the concept of inherent equality is not a biological or psychological concept but a moral concept.
I don't disagree with moral equality one bit - the golden rule should absolutely apply to everyone and we should all strive to look at each other with compassion, tolerance, empathy, understanding, grace, humility, and goodwill - but there is a loud, vocal subset of people that truly believe in absolute biological equality - not just between people with different IQ's, but even between different biological sexes and all other categories of humans; "tabula rasa" proponents who argue that ALL differences in outcomes along the lines of categorical differences (e.g. sex, race, gender, ethnicity, nationality, religion, culture) are exclusively and irrefutably explained strictly by discrimination and discrimination alone, which is a patently absurd assertion that should be refuted. Discrimination is real and should be confronted vocally, but the idea that it's the only factor explaining differences in outcomes between groups is a harmful myth.
The fact of the matter, relating back to the original discussion, is that sex/race/gender/ethnicity/nationality/religion/culture-blind IQ testing is not only a strong predictor of job performance, it is perhaps one of the best tools we have for eliminating discrimination based on sex/race/gender/ethnicity/nationality/religion/culture in hiring, as it explicitly controls for differences along these lines by exclusively targeting an assessment of g in abstract ways that are explicitly stripped of cultural, religious, racial, and gendered biases.
Pseudonymized hiring that relied exclusively on IQ tests, with zero indications of race/sex/gender (e.g. legal name), stripped of proxies for SES and/or parental SES (e.g. which university was attended, if any) would be significantly less biased than current hiring practices. Throw in job-specific pseudonymized skill evaluations (so, no voice calls, no video calls, just direct assessments to candidates) and you've got a system to dramatically reduce hiring discrimination along protected classes.
... or you can just discard all the hokum and only use job-specific pseudonymized skill evaluations --- a concept more widely known as "work samples" --- and be using the gold-standard tool for candidate selection, supported by research going all the way back to Deming.
reply