I'm not sure what the argument is here. I think we're both in agreement that if changing the seed affects the accuracy of your model, then your design isn't good. Your concern seems to be there are too many bad models on too little data, which I also agree with. Randomized seeds don't fix that.
October 24, 2025 at 4:53 PM
I'm not sure what the argument is here. I think we're both in agreement that if changing the seed affects the accuracy of your model, then your design isn't good. Your concern seems to be there are too many bad models on too little data, which I also agree with. Randomized seeds don't fix that.
I don't care if they do or don't. Changing the seed gives you the same chance of failure. Selecting a seed at random gives you the same chance of failure. Getting more/better data reduces your chance of failure. If your success is significantly affected by chance, you need more/better data.
October 23, 2025 at 5:12 AM
I don't care if they do or don't. Changing the seed gives you the same chance of failure. Selecting a seed at random gives you the same chance of failure. Getting more/better data reduces your chance of failure. If your success is significantly affected by chance, you need more/better data.
The argument is that the percent chance that a particular seed causes an issue shouldn't change by seed. So, even if seeds were randomly selected, you'd still end up with approximately the same number of issues. If the seed breaks your model, I'd argue it's the model that's the problem not the seed.
October 23, 2025 at 1:27 AM
The argument is that the percent chance that a particular seed causes an issue shouldn't change by seed. So, even if seeds were randomly selected, you'd still end up with approximately the same number of issues. If the seed breaks your model, I'd argue it's the model that's the problem not the seed.
Lovecraft Country and Watchmen teaching Americans about the Tulsa massacre would be a good point to make.
October 2, 2025 at 10:32 PM
Lovecraft Country and Watchmen teaching Americans about the Tulsa massacre would be a good point to make.
Reposted by Joe Blackmer
Medicaid is the largest source of insurance coverage for people with HIV in the United States, covering an estimated 40% of the nonelderly adults with HIV www.kff.org/hivaids/issu...
Medicaid and People with HIV | KFF
This data note provides an overview of the role of the Medicaid program for people with and at risk for HIV.
www.kff.org
February 26, 2025 at 1:37 AM
Medicaid is the largest source of insurance coverage for people with HIV in the United States, covering an estimated 40% of the nonelderly adults with HIV www.kff.org/hivaids/issu...