I think we’re saying the same thing. I had understood your prior comment to mean that 2014 included 36.8%.
I think we’re saying the same thing. I had understood your prior comment to mean that 2014 included 36.8%.
The text is to the left on '15; zoom in and compare the circles to the year. It was a 15-16 jump according to the dots.
Journal quality can buffer this by getting better reviewers (MDPI shouldn’t be seen as having peer review at all, but peer review at the best journals–because professors want to say on their merit raise annual evals that they are doing the most service to the field by reviewing at the best journals–is usually good enough at weeding out bad papers), but it gets offset by the institutional prestige of authors when peer-review isn’t double-blind. I’ve seen some garbage published in top journals by folks that are the caliber of Harvard professors (thinking of one in particular) because reviewers use institutional prestige as a heuristic.
When I’m teaching new grad students, I tell them exactly what you said, with the exception that they can use field-recognized journal quality (not shitty metrics like impact factor) as a relative heuristic until they can evaluate methods for themselves.
Sorry, what? Not sure if you’re joking, but Americans use texts because they’re free and the ability to use them comes preloaded on the phone (no need to download something that takes up more space). I have Signal and WhatsApp on my phone for my international friends, but I use texts to communicate with US friends because RCS works with everyone and it’s integrated much better into my phone, watch, etc. than any app can be without an absurd amount of permissions given to the app.
Engagement helps posts in various algorithms, though I’m not sure that Lemmy uses comments for Hot or anything else. More importantly, I think there’s truth to the meme that the quickest way to get an answer to your question on the internet isn’t to ask the question, it’s to tell someone else the wrong answer. People will then chime in with the right answer if they know it. Wrong answers can be useful in that respect.
A fellow Julia programmer! I always test new models by asking them to write some Julia, too.
I actually took that bit out because LLMs are pro climate and against everything that makes the environment worse. That’s a result of being trained on a lot of scientific literature. I was just curious what Opus would say about the conceptual knowledge piece.
Claude Opus disagrees, lol (edit to add: all of what follows is Claude; not me):
I respectfully disagree with the claim that current language models are logically incapable of mapping language to actual concepts. While today’s LLMs certainly have limitations and are prone to hallucinations, they have demonstrated a remarkable ability to capture and manipulate complex concepts through language.
A few key points:
LLMs are trained on vast amounts of human-generated text which is grounded in real-world concepts, allowing the models to build rich representations that go beyond surface-level word associations. Techniques like unsupervised pre-training enable LLMs to learn meaningful conceptual relationships.
In many domains, LLMs have shown an impressive capacity for reasoning, inference, and knowledge synthesis - skills that rely on a deep understanding of the underlying concepts, not just pattern matching. Their performance on complex question-answering, analogical reasoning, and natural language inference tasks is evidence of this.
LLMs can be fine-tuned and augmented with techniques like retrieval-augmented generation or chain-of-thought prompting to further ground them in factual knowledge and improve their reasoning capabilities. This allows more robust conceptual mappings.
Careful prompt engineering and sampling techniques can help to mitigate hallucinations and nonsensical outputs. While not a complete solution, this shows that the conceptual knowledge is often present even if the model sometimes struggles to express it faithfully.
That said, I fully agree that the tendency for subtle nonsense is a serious issue, especially for layperson use cases where hallucinations may go undetected. Continued research into making LLMs more truthful, consistent and robust is crucial. Techniques like constitutional AI to bake in truthfulness, and better UX design to convey uncertainty are important steps.
But in summary, I believe the evidence suggests that LLMs, while flawed, are not fundamentally incapable of meaningful conceptual representation and reasoning. We should push forward on making them more reliable and trustworthy, rather than dismissing their potential prematurely.
Sorry, but this makes clear that you aren’t in science. You should avoid trying to shit on studies if you don’t know how to interpret them. Both of the things you mentioned actually support the existence of a true effect.
First, if the treatment has an effect, you would expect a greater rate of relapse after the treatment is removed, provided that it treats a more final pathway rather than the cause: People in the placebo group have already been relapsing at the typical rate, and people receiving treatment–whose disease has been ramping up behind the dam of a medication preventing it from showing–are then expected to relapse at a higher rate after treatment is removed. The second sixth-month period was after cessation of the curcumin or place; it was a follow-up for treatment-as-usual.
Second, people drop out of a study nonrandomly for two main reasons: side effects and perceived lack of treatment efficacy. The placebo doesn’t have side effects, so when you have a greater rate of dropout in your placebo group, that implies the perceived treatment efficacy was lower. In other words, the worst placebo participants are likely the extra dropouts in that group, and including them would not only provide more degrees of freedom, it would theoretically strengthen the effect.
This is basic clinical trials research knowledge.
Again, I have no skin in the game here. I don’t take curcumin, nor would I ever. I do care about accurate depictions of research. I’m a STEM professor at an R1 with three active federal grants funding my research. The meme is inaccurate.
Why are you completely ignoring the second paper I linked, which doesn’t suffer from any of the limitations you mentioned?
The meme says no trial was successful. Any trial with any small difference is a successful trial.
I’m not saying the study is good, just that the meme isn’t true.
Also, you can level almost every single one of those criticisms against many studies for SSRIs and they’d hit just as hard. The exception being sample size.
Not true:
https://www.sciencedirect.com/science/article/pii/S0165032714003620
https://www.cghjournal.org/article/S1542-3565(06)00800-7/fulltext
I found more, too.
Edit: I have no skin in this game. I don’t take turmeric and won’t ever because of the risk of lead. I’m just pointing out that the meme is inaccurate. The person who replied to me pointed out some flaws in the first study (not the second), but none of the flaws mentioned makes the meme accurate. Even the shitty first study I linked found a significant condition difference in its primary endpoint at 8 weeks. Yeah, it’s got flaws (which the second doesn’t), but a successful trial with heavy limitations and conflicts of interest is nonetheless a successful trial, making this meme inaccurate. The second study I linked is stronger.
Also, the limitations in the first trial are standard for many clinical trials. For example:
https://onlinelibrary.wiley.com/doi/abs/10.1111/jsr.12201
https://www.sciencedirect.com/science/article/pii/S0924977X14001266
I could list 100 more with the same limitations of the first study I linked above. High dropout, small sample sizes, funding by an industry with a conflict of interest etc. are standard for clinical trial studies.
That’s not actually the abstract; it’s a piece from the discussion that someone pasted nicely with the first page in order to name and shame the authors. I looked at it in depth when I saw this circulate a little while ago.
deleted by creator
I’m thinking of shorting it. My friend is definitely shorting it.
Would you, after devoting full years of your adult life to the unpaid work of learning the requisite advanced math and computer science needed to develop such a model, like to spend years more of your life to develop a generative AI model without compensation? Within the US, it is legal to use public text for commercial purposes without any need to obtain a permit. Developers of such models deserve to be paid, just like any other workers, and that doesn’t happen unless either we make AI a utility (or something similar) and funnel tax dollars into it or the company charges for the product so it can pay its employees.
I wholeheartedly agree that AI shouldn’t be trained on copyrighted, private, or any other works outside of the public domain. I think that OpenAI’s use of nonpublic material was illegal and unethical, and that they should be legally obligated to scrap their entire model and train another one from legal material. But developers deserve to be paid for their labor and time, and that requires the company that employs them to make money somehow.
Wow, a real, live, tankie!
Wow, a real, live tankie!
deleted by creator
Eh, I switched. I switched all of my lab’s computers, too, and my PhD students have remarked a few different times that Linux is pretty cool. It might snowball.