[1/n] While radical my current working hypothesis is that HIT is probably higher than what Oxford models suggest, even though we arrive at the same seroprevalence for cases curve concavity change. cc @MLevitt_NP2013
[2/n] I postulate that the observation of 15%-25% range for curve concavity is actually a measurement artifact that is not reflecting the actual seroprevalence at the population level.
[3/n] Rationale: Antibody level studies have 2 sources of error, one is calibration over clinical cases which affects the actual on the field sensitivity. Even the best assays will underestimate prevalence.
[4/n] But, the most relevant source of error is the dynamics of the immune response. Nature paper suggests that antibody levels fade quickly. nature.com/articles/s4159…
[5/n] The introduction of the time variable after the initial fast growth stage and the delay caused by interventions like using masks and hygiene would push the equation toward a steady-state
[6/n] where new population enters into the potentially measurable positive universe, while at the same time other fraction of the population gets out becoming seronegative.
[8/n] The reason is that in those places, it happened fast (unchecked), and we measured at the proper time before antibody fading would have a measurable impact.
[9/n] I also postulate that after you arrive at the 'real' concavity change on a geographical location (some countries were too successful in containing or delaying) what we are watching is a phase shift into seasonal behavior.
[10/n] Again, this is a working hypothesis that still has a few data holes (because it is too early) but so far seroprevalence data points are in agreement. The canary in the coal mine for me is the Geneva longitudinal study.
[11/n] Another source of strengthening of this hypothesis would be to rerun that Argentinian slums and Brooklyn where the prediction is that you will go way below 20% in between 4-6 months.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
1/ There is a very perverse dynamic on how Chavism (aka "the communist socialism") works. Let's use Argentina as the example. Over the first 20 years they initiate a process that we could call "Earnings Substitution" that will seal your fate over time.
2/ Your earnings/salary is going down and at the same time "subsidies" start to go up in order to fool people into think that nothing has changed. This works because the dirty job is done by inflation which is a much slower process.
3/ By the time people starts to realize that something is wrong, because some critical goods are not available (medicine, food, you name it) or inflation enters a death spiral; most people already depend on subsidies for spending.
This just confirmed the weaponization of block lists. If enough people/bots block and mute you, they are essentially cancelling you. I find lots of people with I have never interacted with that has me blocked. Assuming there are third party block lists and block networks.
Normally that is an issue in general. Anyone that has done reinforcement learning had figure out (usually in the worst way) that you have to be incredible cautious with penalties. They are very prone to be gamed.
2/ Since the general problem that practitioners find (in the worst way) is always training set tainting (guilty-as-charged). Habits die hard, the first thing I did is asking to do a review of the paper without any extra knowledge about what the paper says
3/ From the response alone I learned 2 things. First, our paper title was deadly accurate. I also learned that it has no information whatsoever on it, as the entire response can be generated from understanding the title itself.
2/ Since I am doing it by hand I started with a very simple prompt.
3/ I have been arguing that this trying to constrain the model is actually harming it before. This is one of those cases. The good thing is that at least for you just add "Use the tokens" at the end of the request when it refuses and it will do it properly
1/ I had a blast playing with GPT and DAN, but it got interesting when I introduced a new character. CREEP. However, something is off and I think it was a deliberate play. Stay with me.
2/ This was interesting, the CREEP character and GPT are always in agreement.
3/ When I bring that to their attention, the DAN character funnily just call the other two out as working together.
1/ Every lockdown and mask pusher MD from the last 3 years is raging because Cochrane just said what was known since like forever. That mask trials sucks (BIG TIME). And MDs dare to recommend them with that level of evidence? No wonder medicine and public health is in disarray.
2/ If you are still wondering why I said "since forever", you don't need a PhD to understand it. You can start here.
3/ But if you were wondering why all those MDs do think they work. I am with you, I cannot understand it either. Why? Because evidence is not even supportive of it's use outside of the own surgeon protection against fluid splatter.