[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 363 KB, 2857x1734, 171202595179.jpg [View same] [iqdb] [saucenao] [google]
16116849 No.16116849 [Reply] [Original]

Academics are literal hivemind charlatan scammers. This has been known by many already, but now there's hard evidence to prove it. So-called "experts" are cheaters and liars just like the average student who cheats on his or her exams. And the evidence focuses on STEM papers specifically, including medicine, and not what are considered bullshit majors where of course everyone is expected to be a bullshitter and plagiarizer.

>Are medical studies being written with ChatGPT?
>Well, we all know ChatGPT overuses the word "delve".
>Look below at how often the word 'delve' is used in papers on PubMed (2023 was the first full year of ChatGPT).
https://twitter.com/JeremyNguyenPhD/status/1774021645709295840
>Here is the one for Math.
https://twitter.com/outscape/status/1774568789239095550
>Same trend in economics! Used ChatGPT (GPT-4) to write a script to scrape this data and make this plot
https://twitter.com/williamhickman_/status/1774097806224920852

>A normalized chart, showing % of papers with "delve" in title or abstract). Source: OpenAlex, articles (includes papers & preprints). Use of "delve" was gradually increasing through to 2022 but then jumped in '23! Some 46% of all 90-24 papers using delve came out in 23 & 24!
https://twitter.com/philipshapira/status/1774497589980742071

>I asked Gemini why this happens: “Sounding Authoritative: Both "delve" and "notable" have a slightly formal, academic tone. ChatGPT might use them in an attempt to sound more intelligent or credible, even if simpler synonyms would be more natural.”
https://twitter.com/GaryMWatson/status/1774090530227446050

>It's hilarious. Same for "multifaceted"
https://twitter.com/padolsey/status/1774046174695108943
https://blog.j11y.io/2023-11-22_multifaceted/

>> No.16116850
File: 31 KB, 1520x770, 1712025951.png [View same] [iqdb] [saucenao] [google]
16116850

heres the multifaceted chart

>> No.16116853

We have to find unique challenge

>> No.16116860

>>16116849
useless graph, add a second line demonstrating percentage of total

>> No.16116863
File: 40 KB, 680x470, delve.jpg [View same] [iqdb] [saucenao] [google]
16116863

>>16116860
check your eyes. Paper output is exponential, this isnt.

>> No.16116866

A very good Idea for a proxy, but what does this change? Everyone knows papers are just word slop these days. ChatGPT helps to churn them out faster

>> No.16116900

You forgot one tiny detail. Do you know the provenance of most of those papers? I tell you, they're chinese. They constantly churn out shoddy articles at an obscene rate in order to get as much coverage as possible, nevermind the quality.The fact that you find patterns in the grammar and syntax of certain abstracts might come from zhang simply using google translate.

>> No.16116910

>>16116849
You sure do seem to care a lot about shit which has no actual impact on your life

>> No.16116969
File: 22 KB, 518x565, cob'd.jpg [View same] [iqdb] [saucenao] [google]
16116969

>>16116910

>> No.16116972

>>16116910
You must not write papers..

I avoid all academics, they tend to be highly ignorant about their fields and I would rather deal with it myself.

>> No.16117569 [DELETED] 
File: 45 KB, 320x320, twilight zone.jpg [View same] [iqdb] [saucenao] [google]
16117569

>>16116860
how come you don't want to delve into this multifaceted issue?

>> No.16117570

>>16116849
they are corrupted by people like you, or pushed in those positions by people like you.

>> No.16117609

I'm surprised more people aren't aware of the replication crisis happening right now in psychology and medicine.

>> No.16117611

>>16116849
What is the advantage of using chatGPT in this case? You still have to read over it to verify that it is corroborating the outline. It can't interpret data. It can't be trusted for citations. It doesn't have formatting functions. You have to coach it, line by line, which means you have already written the shit. Ah, it would be good for translating.

>> No.16117614

lowering the bar so more people get into academia will only make things even worse.

>> No.16118079 [DELETED] 

>>16117614
It will also accelerate it's demise

>> No.16118189

I point this out as much as possible, truly intelligent people do not spend time in academia nor writing retarded circle jerk papers.

>> No.16118205

Delving into the use of ChatGPT for academic research papers reveals multifaceted dangers. While the tool offers notable assistance in generating content, its reliance solely on pre-existing data poses risks. The lack of real-time verification and potential for biased or inaccurate information compromise the credibility and integrity of scholarly work. Thus, prudent caution is advisable when incorporating ChatGPT outputs into academic discourse.

>> No.16119730
File: 361 KB, 448x704, 1709913853784.png [View same] [iqdb] [saucenao] [google]
16119730

>>16118205

>> No.16119747

>>16116849
Why isn't it the other way round? AI has been fed a bunch of papers and it's been known since before you were born that scientists like any other groupl copy eachothers' language. ''More research needed'' and such buzz phrases.

>> No.16120154 [DELETED] 

>>16119747
Scientists are a bunch of dumb cattle who can't think independently LOL

>> No.16120572 [DELETED] 

>>16117609
its not limited to those two fields, nothing published is truthful anymore

>> No.16120580

>language drift is a conspiracy
Delve into trash compactor OP

>> No.16120864

>>16118205
This post lives rent free in my head; the more I think about it the funnier it gets.
I am at the point where I genuinely think that this was written by a human, because there is no way that ChatGPT can sound this ChatGPT.

>> No.16121879 [DELETED] 
File: 47 KB, 527x365, oyyyyyyy.jpg [View same] [iqdb] [saucenao] [google]
16121879

that efficacious intricate multifaceted delve when no LMM symbiosis

>> No.16121890

>>16117611
It's like pair programming, or brainstorming with an intern. You need to iterate upon the results, tell it to shut the fuck up at times, or be concise, point out errors in reasoning, change course. I've been using this from the beginning and have derived a tremendous amount of productivity, particularly from GPT4.

>> No.16123104

>>16120864
>the more I think about it the funnier it gets.
delve delirium

>> No.16123175

>>16116969
Nta but lmao it looks to me you are the crybaby one

>> No.16124682 [DELETED] 
File: 45 KB, 721x789, 1702185189130.jpg [View same] [iqdb] [saucenao] [google]
16124682

>>16123175
>Nta but lmao it looks to me you are the crybaby one

>> No.16125267 [DELETED] 

>>16118205
177 IQ post according to https://www.writingtoiq.com

>> No.16126339 [DELETED] 

>>16117614
academia always has been and always will be for do nothing losers and frauds. capable people are able to move on from school once they become adults, academics are perpetually immature, which is why they stay in school like children for their entire lives

>> No.16126929 [DELETED] 

>>16118205
Give this guy a PhD

>> No.16126933

>>16116849

how to explain use of "delve" increasing even before chatgpt?

>> No.16128018 [DELETED] 

>>16126933
>how to explain use of "delve" increasing even before chatgpt went public and easy to access for gayphone retards

>> No.16129364

>>16126933
plagiarism

>> No.16130002

>>16129364
Probably, the period before the "AI" software came out was the period during which minorities were told that they would never be held responsible for any type of academic fraud they committed

>> No.16130434
File: 77 KB, 1010x615, 001.jpg [View same] [iqdb] [saucenao] [google]
16130434

>>16116910
>sorry chud, you don't belong to the right social caste so you're not allowed to talk about that

>> No.16130437

>>16116849
Not my problem. This is the natural result of dysgenics and leftist polices.

>> No.16130966

>>16116850
weird how intricate, symbiotic and multifaceted all dip in 2021 before undergoing explosive growth

>> No.16130970

>>16116969
>>16130434
faggot

>> No.16132271

>>16130970
At least we can all agree that being a homosexual is reprehensible and that anyone who is accused of being a homosexual should be deeply offended

>> No.16132935

>>16117611
It's shit for translating. Used it to translate a bunch of legal papers, it went off the rails at some point and wrote a bunch of stuff about weed.

>> No.16132939

>>16132271
don't forget chemically castrated, ostracized, and their family shamed my nigger.

>> No.16133063

>>16116910
>no actual impact
I really want to crack open your skull with a pickax, you stupid moron.

>> No.16133206

>>16120864
>>16118205


How come robots trained by human written text still sound so robotic

>> No.16134143

>>16133206
the longer "AI" software exists, the more the training sets becomes polluted with "AI" software output

>> No.16135103

It's important to consider that naive Bayes on 1-grams won't capture all of the LLM's phony catchphrases.

>> No.16135522

>>16134143
This loop will eventually lead to hilarious consequences

>> No.16135671
File: 3.33 MB, 4364x3198, Théodore Maurisset - La Daguerreotypomanie (1839) - People lined up for photos while painters rent gallows.jpg [View same] [iqdb] [saucenao] [google]
16135671

>>16116849
Are the papers incorrect? If not keep seething.

>> No.16136665

>>16120864
the deeper you delve into the humor the more multifaceted it becomes

>> No.16137627

This thread was moved to >>>/pol/465628070