Wrote a small script to test how LLMs handle growing context sizes. It sends larger prompts step by step and measures slowdown or failure. Pretty rough but useful so far. Curious if others have ideas for improving it.
no, that makes no sense. How does having identical genetics negate problem of different lifestyles?
Even the authors agree with awestley:
> Additionally, having a tattoo, especially among adolescents, has been suggested as an indicator of risky lifestyle highly associated with e.g. smoking [4] and alcohol consumption [46] – both risk factors of certain cancer types. Hence, evidence of an association between tattoo ink exposure and occurrence of cancer may be confounded by other health-related lifestyle factors. We intend to exploit the remainder of the information gathered in the survey in the future.
Indeed, it does seem like a big problem that should really be accounted in the initial analysis.
Twin studies can help control for things like genetics and environmental factors in childhood (exposures, socio-economic factors, etc). They can't control for lifestyle choices made by adults. So, if a person has tattoos, are they more likely to smoke? To have sun exposure? To drink heavily? All of these factors would need to be addressed to see how confounding they are to having a tattoo.
You could also look at tattoo coverage, as in how much of the body is covered in ink? Would a small tattoo on the shoulder have as much risk as a full back tattoo? There are a lot of extra confounders here that could be better explored, but it gets difficult to get a full dataset. However, given their survey data, they should have more analysis options with more time.
The lack of confounder analysis is a bit surprising, but perhaps the paper was long enough already.
Are people who have tattoos more likely to want to have their skin exposed to the sun (to show off the tattoos)? That seems like a reasonable association, but it's probably dependent upon tattoo location.
It's therapy for a lot of people. The act of organizing helps them deal with information. That act of organizing and storing can help with recall for me but not as much as writing something down on paper does. The number of times people pull info out of these systems is questionable.
An argument people use is that these systems help you later in life. I find these systems really hard to adopt and also find it difficult to work with people who expose these systems outwardly.
I think it is for power users who document a lot, just think of it as a better way to organise your documentation. I think the secret behind good documentation is the effort which will go daily to maintain it vs following a specific system.
If you are putting in daily effort you will automatically find a system which suits your needs.
Oh no no no. Companies are supposed to be perfect and clairvoyant and never make a mistake. So it’s entirely reasonable to expect them to have no need for learning or fixing or reversing a decision.
The Wired article explores the promising practice of carbon farming and its potential to revolutionize agriculture. Carbon farming involves adopting regenerative techniques that not only enhance soil health and biodiversity but also mitigate climate change by sequestering carbon in the soil. By implementing methods such as rotational grazing, cover cropping, and agroforestry, farmers can enhance soil carbon levels while improving crop yields and resilience to extreme weather events. The piece highlights successful carbon farming initiatives around the world and underscores the urgent need for widespread adoption to combat climate change and ensure food security in the face of environmental challenges.
