4,368 karma · joined October 10, 2018
https://docs.github.com/en/github/copilot/research-recitatio...
Every literally quoted part that could infringe appears at least 10 times in the training data
In computability theory, several closely related terms are used to describe the computational power of a computational system (such as an abstract machine or programming language)
This is wrong, this is not what Turing completeness is. It applies to computational models, not hardware.
(probablilty of contracting COVID) * (probablity of heart injury in the military from COVID if someone contracts COVID)
[0]: https://ivmmeta.com/#fig_fpr
[1]: https://hcqmeta.com/#fcite_rct (Figure 7, scroll down)
Let's say that it's doing exactly what it was trained to do.
[0]: https://en.wikipedia.org/wiki/Fast_inverse_square_root#Overv...
In any case, thank you for trying to use reason in your argument, we need more around this subject.
These websites do not include protocol registration with methods, search strategies, inclusion criteria, quality assessment of the included studies nor the certainty of the evidence of the pooled estimates.
This is false. The ivmmeta.com page says:
Typical meta analyses involve subjective selection criteria, effect extraction rules, and study bias evaluation, which can be used to bias results towards a specific outcome. In order to avoid bias we include all studies and use a pre-specified method to extract results from all studies (we also present results after exclusions).
They also compare prospective and retrospective studies and see no significant difference. You could have at least scrolled through the page before blindly copy-pasting something from Google.
SETI@home achieved 50 times the computing power of the world's largest supercomputer [0], so it might actually be the only way to train the future GPT4 or GPT5.
How about 31 randomized control trials?
A nice meta analysis page:
Ivermectin is not the only substance studied: