@Kulei@social.vivaldi.net

Luce @Kulei@social.vivaldi.net

· 7mo ago

Replying to

@notjustbikes@social.notjustbikes.com worth mentioning in a video?

4

1

0

Luce @Kulei@social.vivaldi.net

· 5mo ago

Recently researchers used LLMs for advice on making strategical decisions. Despite the context they were adding with each iteration machines persisted in offering the same solutions. They described the LLMs as outputting "trendslop", just a mere regurgitation of popular ideas – an another example of how LLMs regress to the mean and turn brilliance into stupidity, tainting every fresh idea with it's inherent rot.

https://hbr.org/2026/03/researchers-asked-llms-for-strategic-advice-they-got-trendslop-in-return

Tagging @pluralistic@mamot.fr for reach, assuming this wasn't discussed before

Harvard Business Review

Researchers Asked LLMs for Strategic Advice. They Got “Trendslop” in Return.

Leaders might assume that LLMs are able to offer a kind of unbiased, outside perspective. But new research found that leading LLMs have clear biases when it comes to strategy and consistently recommend strategies that align with modern managerial buzzwords and trends rather than context-specific strategic logic. This propensity for AI to opt for buzzy ideas over reasoned solutions is called “trendslop,” and leaders should beware of it warping their strategic planning. When using AI in strategic

1

0

Luce @Kulei@social.vivaldi.net

· 10mo ago

Replying to

@shadowdancer@mstdn.social @jargoggles@kolektiva.social @melanyabelta@wandering.shop A soon to be psychologist here. Yeah they are different ways of thinking that usually get used in different situations and people have their own preferences for thinking, for instance autistic people significantly prefer detail oriented thinking. Both styles of thinking have their pros and cons. Kids more often than not prefer using detail oriented thinking, because that is expected of them in school. It can also depend on the amount of competence/information one has, for instance an expert in a field may sometimes instinctually say "yeah that's not going to work" on a hunch without conciously thinking about the details. Same with people that get to interact with a lot of images and with time gain competence in recognizing machine generated input. Both styles of thinking may actually be useful in that situation

1

0

Luce @Kulei@social.vivaldi.net

· 12mo ago

Replying to

@forthy42@mastodon.net2o.de @noodlemaz@med-mastodon.com @Natasha_Jay@tech.lgbt There are other factors at play here, for instance autism is a difference in brain and nervous system structure which is mostly hereditary, whereas the diagnostic criteria in DSM-V or ICD-11 are about a '"disorder", which means that it the person needs to be disadvantaged in some way by the "symptoms". That's why in this case it's extremely vital to take into account cultural differences in social order and the philosophy about how diagnosis is conducted (and who is it for). Also the "symptoms" combined themself don't constitute the whole difference, they are only the most useful markers to use in a given context. Clinicians in other countries might also be less educated about other ways autism could be recognised, for instance even in black people and women in US it's underdiagnosed.

to all that there are some analyses that take into account not only the amount of diagnoses into account, but also count them retroactively to when the person was born, as many people are not diagnosed at 8 as already they might be masking spectacularly well, and it's not like the diagnoses are suddenly of a newer generation, but it fits the even distribution model better.

and some other things I've no space to put here

1

0