@llallawg@Sauers_ How can you administer something that is not even manufactured? Manufacturers will wait for legal approval and they will not care if a specific stage 3 cancer wants something
@AndrewCurran_ I agree that it's poorly communicated. Do we want to avoid AI expoiting the trial process? Yes. But the answer should be making the trial more bulletproof, not the mechanism understood first. Biochemistry is complex and patients don't have time to wait.
@Harrlarn@Pherekydez@ChrisHayduk Not really. Phase 3 could fail for several reasons:
- Unknown/Ignored adversarial effect
- Skewed statistics due to selected population e.g. chosen population may have hidden external factor
- Compromised random selection process e.g. hacking the patient database
@bjarkeroune@LimYenKheng He spent his entire life verifying his works, not just writing a bunch of unfinished nonreadable stuff with barely any review. Btw OpenAI made a big announcement on social media, they didn't simply put it on Github.
@etteuree Wel Akhtually, you have to either measure or assume that because it is a variable that depends on the location and the moment at which the experiment is taken due to effect from the Earth's rotation and the Sun and Moon's gravitational field
I was shocked that AI broke the nlog(n) multiplication barrier until I learned that human-made algorithm that actually achieved nlog(n) takes 40 pages to explain and somehow involves a 1729-dimension Fourier transform
https://t.co/muZcwozpPu