Mariehane
Joined 9 years ago
Karma
52
Submissions
29
Member since
May 17
Recent submissions
Comment · 11 months ago
Will there be a large-dose trial? I imagine that’s going to be more difficult, as the weight-loss effect of GLP-1 means you cant include frail patients in your target group. And GLP-1 at high doses also has some unfavourable side-effects: nausea, vomiting, diarrhea, constipation.
Comment · 11 months ago
I think you naturally undergo that course when you are maintainer of a large OSS project.
Comment · 1 year ago
But then you’re creating an incentive for the AI slop to become so realistic it is indistinguishable from actual video.
Unless there some fundamental, technical way to distinguish the two, I wonder who would win?
Comment · 2 years ago
While Bitcoin is designed to reward high energy consumption the opposite is true for the fiancial system, since lower energy costs would make the different processing entities earn higher margins.
Thus, the financial system provides much more utility for the energy it consumes than bitcoin does.
Comment · 2 years ago
It seems most commenters here are trying to use the default model, which is akin to chatting with non-tuned GPT.
There is a 7b-chat model, which answers perfectly to questions like “who are you?”
Comment · 3 years ago
I read an absolutely heart-wrenching excerpt from this book (The People's Hospital) recently in the new yorker. See: https://www.newyorker.com/science/annals-of-medicine/the-ass...
One interesting point it emphasises is just how important insurance coverage is for survival. And that government insurance processes can be rather arbitrary or even injust.
I hope this book is successful and that it makes a change towards more equitable healthcare in the United States.
Comment · 3 years ago
The website has a '©Open AI™ 2014 - 2023' text in the bottom right if you go to another page. Maybe this supports the domain squatting / trademark abuse hypothesis?
Comment · 3 years ago
Perplexity creates permalinks for your queries. Here's an example: https://www.perplexity.ai/?s=u&uuid=605b2823-3067-4cd1-aacf-...
Comment · 3 years ago
I wouldn't want to host the tournament, but I'm also down for working on a web-client.
Comment · 3 years ago
ChatGPT already works suficiently well in my own language (Danish). It can probably be improved, but so far it has worked well enough for several news stations to report on ChatGPT while prompting it in Danish. And schools are also worried about its ability to write essays in Danish, among other things.
Comment · 3 years ago
Chrome is the iPhone of the browser world - there are probably things in other browsers that work better, or that you could customize to fit your needs, but instead of spending the time to optimize those you can just choose chrome and you'd be pretty satisfied overall.
Comment · 3 years ago
> But such is the world we find ourselves in that anything short of a ground-breaking discovery is treated as an unacceptable embarrassment. And thus, every marginal increment is touted as ground-breaking, and every deviation is repackaged to look like an intended result (which is again touted as ground-breaking)
This is a very nice quote.
Comment · 3 years ago
Hi! I thought this was an excellent recap of the whole climate crisis, from the initial discovery to the most recent developments. It also features some incredibly engaging writing and an interesting structure which works surprisingly well.
Hope you enjoy!
I am interested in the intersection of Machine Learning and Oncology.
Here's two projects I was involved in:
- An analysis of sepsis algorithms in hospitals: https://www.statnews.com/2022/02/28/sepsis-hospital-algorithms-data-shift/
- A model for lung cancer risk from a single CT: https://ascopubs.org/doi/full/10.1200/JCO.22.01345
If you're doing exciting, challenging work related to this space, Please reach out!