It is stated that they use the same annotators that trained/filtered chatGPT’s output. I would assume its a rather large group (my company has 10 auditors in Nicaragua). The label biases are mostly stemming from that group and - as suggested - could be removed by using experts in each field to annotate the labels. But given some responses here by experts, I am sure those expert labels would have their very own biases :p
The paper is not of highest quality indicated by typos and mislabels but the analysis is likely as good as it can get for the given methodology. Dismissing any signal is just pure hubris.
Am I the only one thinking this no-checkout comes at a huge cost of extreme surveillance? This is creepy as fuck. I would never want it in a store, no matter how much time I save. The self checkout machines at CVS make checkout already x-times faster. This would only be a small gain a price I am not willing to pay. Also there is some satisfaction in paying for things/checking out.
first AMD with Ryzen and now FF. Working with nightly in the past few days and don't miss a thing. Of course not all plugins are ported yet, but ublock works already. keep up the phenomal work.
Totally agree. Just look at the mess with mobile OS. It is nearly impossible to develop a native app yourself without excluding half of the world population. In my opinion, native mobile apps have been always an intermediary and I wouldn't mind seeing all of them going away if some better web app alternative comes along.
As founder of openrev.org, a platform to share and discuss research papers openly (or privately), I am very pleased with this announcement. It is more than just baby steps towards an open, scientific communication system.
In physics we do already have that with the arxiv.org (though there are no public discussions) but it works darn well in my sub-field and others. We cannot wait for most other disciplines to have an epiphany. So grant agencies forcing publishers AND authors to open access their material will bring us a huge step forward towards solving the most challenging problems of our time.
I am the co-founder of Open Rev. (www.openrev.org), a platform where researchers can collaboratively annotate research publications and scientific books. I am a physicist and we designed our platform such that scientists can explain complex scientific contexts to each other (latex, markdown, image support).
I presume that there will be little “where is this reference from” but more “what does this word mean” (for example look at all these acronyms in biology and chemistry papers).
Third kind of annotations is simply posting links to good references. From tests in Physics classes it turns out this is what students did and liked a lot. Sometimes its just damn hard to find a good explanation or image of a physical concept.
That brings me to my last group of annotations: Explanations. That is the rephrasing of the scientific language that is sometimes written truly encrypted. We also anticipate people to improve the scientific content by suggestions new solutions, highlight flaws, post missing derivations, etc. This is likely the hardest annotation form, since it requires "academic rigour" as some of you have noted. And this is what we are shooting for.