3 ms·

> The evidence and defeatablity that is built into STEM I don't know if evidence and defeatability is really built into STEM. I'm an academic in computer scien

> The evidence and defeatablity that is built into STEM I don't know if evidence and defeatability is really built into STEM. I'm an academic in computer science, and in the field of computer science we recently had a PhD candidate commit suicide in a major academic institution due to corrupt approval rings in major conferences. Similarly, several fields of STEM have reproducibility crises, including my own field, as well as politically motivated science as is being explored in this article. In theory, STEM has evidence and has defeatability through the scientific method. In reality, the people working in the field of knowledge are still spending most of their time wading through bullshit and only taking the opinions of their friends to be true.

anoxor 6y ago
I totally agree. I have a chem PdD and have seen the imperfections of the system. I would also suggest that corrupt approval rings being a problem is evidence of defeatablity - although a very messy one. You lose your job if shown to be corrupt. In chem as well, something like 30 % of published work is unreproducible, but there is at least a mechanism where that can be fixed. The example I like is white fragility. Agree with it - affirm it's truth. Disagree with it - affirm it's truth. Such a simple circular logic that an undergrad shouldn't get a passing grade on the work is now near secular gospel.
- majormajor 6y ago
  If you have a Chem PhD where did you gain the expertise on the humanities to be able to summarize and dismiss it so quickly? I see similar generalizations frequently, but rarely links to primary sources.
  - anoxor 6y ago
    Because I was told by my dean of STEM that "it's not advantageous for this position that you're a white scientist" followed by "You came from a well-educated family right?". My lower class farmer dad that that was funny. I was forced to get up to speed pretty quickly and strongly motivated to do so. people don't understand how far this is gone in the college system right now, and how that is now being incorporated into every curriculum and taught to every student as what seems to me to be a secular morality that you would be an immoral person to not agree with. Check out table a and B from the following Berkeley initiative. there's a clear biased towards diverse candidates and diversity is simply defined as non-white straight or male. https://ofew.berkeley.edu/sites/default/files/life_sciences_inititatve.year_end_report_summary.pdf https://ofew.berkeley.edu/sites/default/files/life_sciences_...
- TeaDrunk 6y ago
  > You lose your job if shown to be corrupt This is not true in computer science. There have been no major oustings following the literal suicide of a PhD candidate who had directly witnessed the corruption for themselves.
Nasrudith 6y ago
That fittingly enough sounds like a description of a NP hard problem in nature and theoretical vs practical verifiability. Individually every instance may receive validation relative to the local. In practice doing it and cross comparing across all is impossible due to scale - and that is just for simple propositions. Propositions themselves may be too massive to hard prove. I suppose that in itself is a small but significant difference about more mathematically practicable fields. If someone comes out with a P=NP proposes algorithm we can test it for various conditions and say "No that is bullshit, just obfuscating it as - setting the state up requires (N!)." If it really isn't. The smaller units of operability can be verified and there are no "anomalous unreproducible artifacts" to contend with.

anoxor 6y ago

I totally agree. I have a chem PdD and have seen the imperfections of the system. I would also suggest that corrupt approval rings being a problem is evidence of defeatablity - although a very messy one. You lose your job if shown to be corrupt. In chem as well, something like 30 % of published work is unreproducible, but there is at least a mechanism where that can be fixed. The example I like is white fragility. Agree with it - affirm it's truth. Disagree with it - affirm it's truth. Such a simple circular logic that an undergrad shouldn't get a passing grade on the work is now near secular gospel.

majormajor 6y ago
If you have a Chem PhD where did you gain the expertise on the humanities to be able to summarize and dismiss it so quickly? I see similar generalizations frequently, but rarely links to primary sources.
- anoxor 6y ago
  Because I was told by my dean of STEM that "it's not advantageous for this position that you're a white scientist" followed by "You came from a well-educated family right?". My lower class farmer dad that that was funny. I was forced to get up to speed pretty quickly and strongly motivated to do so. people don't understand how far this is gone in the college system right now, and how that is now being incorporated into every curriculum and taught to every student as what seems to me to be a secular morality that you would be an immoral person to not agree with. Check out table a and B from the following Berkeley initiative. there's a clear biased towards diverse candidates and diversity is simply defined as non-white straight or male. https://ofew.berkeley.edu/sites/default/files/life_sciences_inititatve.year_end_report_summary.pdf https://ofew.berkeley.edu/sites/default/files/life_sciences_...
TeaDrunk 6y ago
> You lose your job if shown to be corrupt This is not true in computer science. There have been no major oustings following the literal suicide of a PhD candidate who had directly witnessed the corruption for themselves.

Nasrudith 6y ago

That fittingly enough sounds like a description of a NP hard problem in nature and theoretical vs practical verifiability. Individually every instance may receive validation relative to the local. In practice doing it and cross comparing across all is impossible due to scale - and that is just for simple propositions. Propositions themselves may be too massive to hard prove. I suppose that in itself is a small but significant difference about more mathematically practicable fields. If someone comes out with a P=NP proposes algorithm we can test it for various conditions and say "No that is bullshit, just obfuscating it as - setting the state up requires (N!)." If it really isn't. The smaller units of operability can be verified and there are no "anomalous unreproducible artifacts" to contend with.