←back to thread

Grok 3: Another win for the bitter lesson

(www.thealgorithmicbridge.com)
129 points kiyanwang | 1 comments | | HN request time: 0s | source
Show context
bambax ◴[] No.43112611[source]
This article is weak and just general speculation.

Many people doubt the actual performance of Grok 3 and suspect it has been specifically trained on the benchmarks. And Sabine Hossenfelder says this:

> Asked Grok 3 to explain Bell's theorem. It gets it wrong just like all other LLMs I have asked because it just repeats confused stuff that has been written elsewhere rather than looking at the actual theorem.

https://x.com/skdh/status/1892432032644354192

Which shows that "massive scaling", even enormous, gigantic scaling, doesn't improve intelligence one bit; it improves scope, maybe, or flexibility, or coverage, or something, but not "intelligence".

replies(7): >>43112886 #>>43112908 #>>43113270 #>>43113312 #>>43113843 #>>43114290 #>>43115189 #
cardanome ◴[] No.43113843[source]
> Sabine Hossenfelder

She really needs to stop commenting on topics outside of theoretical physics.

Even in physics she does not represent the scientific consensus but has some very questionable fringe beliefs like labeling whole sub-fields as "scams to get funding".

She regularly speaks with "scientific authority" about topics she barely knows anything about.

Her video on autism is considered super harmful and misleading by actual autistic people. She also thinks she is an expert on trans-issues and climate change. And I doubt she know enough about artificial intelligence and computer science to comment on LLMs.

replies(8): >>43113905 #>>43114246 #>>43114537 #>>43115760 #>>43116614 #>>43116644 #>>43117944 #>>43121775 #
Mekoloto ◴[] No.43113905[source]
Your statement is missleading.

She doesn't say she is an expert on trans-issues at all! She analyzed the studies and looked at data and stated that there is no real transpendemic but highlighed an statistical increased numbers in young woman without stating a clear opinion on this finding.

The climate change videos do the same thing. She evaluates these studies discusses them to clarify that for her, certain numbers are unspecific and she also is not coming to a clear conclusion in sense of climate change yes, no, bad, good.

She is for sure not an expert in all fields, but her way of discussing these topics are based on studies, numbers and is a good viewpoint.

The funding scam you mention is a reference of "these people get billions for particle research but the outcome for us as society is way to small"

replies(3): >>43114307 #>>43115852 #>>43120600 #
bccdee ◴[] No.43115852[source]
> for her, certain numbers are unspecific and she also is not coming to a clear conclusion in sense of climate change yes, no, bad, good.

Climate chance is settled science. To claim that "certain numbers are unspecific, so I can't say whether climate change is real or not, or whether it's good or bad" (which, based on your paraphrasing, is what it sounds like she said) is an unacceptable position. It's muddying the waters.

I'm not going to go watch her content about trans people, but it sounds like the same thing: Muddying the waters by Just Asking Questions about anti-trans "social contagion" talking points.

---

EDIT: Okay I went back and watched some clips of her anti-trans video. She takes a pseudoscientific theory based on an opinion poll of parents active on an anti-trans web forum and suggests we take it seriously because "there is no conclusive evidence for or against it," as if the burden of proof weren't on the party making the positive claim, and as if the preponderance of evidence and academic consensus didn't overwhelmingly weigh against it. It's textbook laundering of pseudoscience. You've significantly misrepresented her position.

replies(3): >>43116149 #>>43116242 #>>43120698 #
toolz ◴[] No.43116242[source]
There's no such thing as "settled science". You can not prove that any scientific consensus has no flaws in the same way you can't prove the absence of bugs in any software. It's unproductive to treat science as anything more than an ongoing, constantly improving process.
replies(1): >>43116595 #
bccdee ◴[] No.43116595[source]
Yes there is. Germ theory is settled science. Is it theoretically possible that we'll overturn it? Sure. Is it likely? No. In the absence of any groundbreaking experimental results, it worth wasting time entertaining germ theory skepticism? Also no.

> It's unproductive to treat science as anything more than an ongoing, constantly improving process.

It's unproductive to constantly re-litigate questions like "is germ theory true" or "is global warming real" in the absence of any experimental results that seriously challenge those theories. Instead, we should put our effort into advancing medicine and fixing climate change, predicated on the settled science which makes both those fields possible.

replies(3): >>43116826 #>>43116949 #>>43118721 #
Matthyze ◴[] No.43116826[source]
Spot on. Reminds me of that old approach by evangelicals to frame scientific consensus as 'just a theory.'
replies(2): >>43116866 #>>43117289 #
1. toolz ◴[] No.43116866[source]
Ironically, it's often attributed to religion that they claim settled truths that can't be proven.