showmethescience

Strong convictions loosely held. Ex–Dean for Medical Education. Founder and

[showmethescience @friiedaaaaaa]

@RolandSeidl2 @BMeinl das ist schon für uns schlimm genug

[showmethescience @friiedaaaaaa]

@javilopen

[showmethescience @friiedaaaaaa]

@Knochenwerfer @motihari1903go Genau - drum gibt es parteinahe Versorgungsorganisationen, die mit Steuergeld erhalten werden.

[showmethescience @friiedaaaaaa]

@GreteSchwetz @CtrlAltDisagree @motihari1903go Ja. Und „nur“ haben Sie geschrieben.

[showmethescience @friiedaaaaaa]

@BiotechTV @EliLillyandCo @DaveARicks @DaveARicks Your “giant drug” cancer comparison for Kisunla caught my attention. We work in computational physiology and have a mechanistic model of AD patient states to identify additional drug targets. Could I send your team a one-page note?

[showmethescience @friiedaaaaaa]

@yevgets probably none, academia is changing.

[showmethescience @friiedaaaaaa]

@naurpillled at Cirsentra Labs we have built a physiological model that can fully explain this.

[showmethescience @friiedaaaaaa]

@piersmorgan @Cristiano Ronaldo obviously is in a manic episode and should go see a psychiatrist asap. The underlying narcissistic personality disorder is probably untreatable but he should get his bipolar disorder under control, it's embarrassing.

[showmethescience @friiedaaaaaa]

@FPLfrasier Ronaldo needs to go to a psychiatrist, that guy is clearly in a manic episode.

[showmethescience @friiedaaaaaa]

@shalseban what a wonderful idea!

[showmethescience @friiedaaaaaa]

@DrSusanNasif did you ever let a frontier model critically appraise published papers (e.g. from Cell, Nature)? most of them are full of grave errors, not replicable figures and insights (even when data is available), exaggerated claims. human error at its best.

[showmethescience @friiedaaaaaa]

@Don_Nozi @austrocom genau! weil der Weißmann die Chats nicht mehr auf seinem Handy hatte, sonst hätte er schon früher den Kontext herzeigen können.

[showmethescience @friiedaaaaaa]

@ross_prager 100% agree. it's similar to forcing people to use Otto Warburg's lab equipment from the 1920s when they have so much more efficient tools at hand today.

[showmethescience @friiedaaaaaa]

@LocasaleLab it's a much bigger flaw that the prize wasn't awarded to the field and so many important early collaborators have been left out.

[showmethescience @friiedaaaaaa]

@shay_fleishon @OpenAI interestingly I have this problem mostly in Codex but not in chat with Astra Pro. I am working on how infections impair receptor function in the human body. @jrm_black

[showmethescience @friiedaaaaaa]

@littmath the non-obviousness of the scientific argument and novelty should be the decisive criteria and it should be easy to evaluate that.

[showmethescience @friiedaaaaaa]

because the AI has been trained overwhelmingly on technical sounding text (I am sure there is a statistics out there about the composition of training data), but people still want the output to sound colloquial and not like a chemistry textbook. Also human writing style is massively heterogeneous, that's why it's so hard to replicate "human writing", unless you provide a (sufficiently long and diverse) style example for imitation.

[showmethescience @friiedaaaaaa]

@FlorianGallwitz interessant. war mir sicher, es geht über die Rechnungsadresse.