rsfern

36 ms·

421.
by rsfern 5y ago
The main bottleneck in GP models is the inversion of an NxN covariance matrix, so training with the most straightforward algorithm has cubic complexity (and quadratic memory complexity). 10k instance is what I’ve seen as the limit of tracta
422.
by rsfern 5y ago
Primary source: https://dx.doi.org/10.1097/AOG.0000000000004695 Couple interesting points: - the effect size is less than one day - they don’t have any data on actual SARS-CoV-2 infection - time between doses seems to
423.
by rsfern 5y ago
Usually they pre-treat the highways and roads around the DMV, but it was raining before the snow started so that was not feasible. The temperature also dropped really fast
424.
by rsfern 5y ago
maybe in general, but for this case specifically there was a national weather service advisory that travel could be dangerous that night (I live in the DMV; there’s currently another advisory for tomorrow night). I’ve never tried to take an
425.
by rsfern 5y ago
This post [0] summarizes the early research on omicron hospitalization risk. The preprints linked are now like 2 weeks old, so probably there is better data available, but I haven’t come across any analysis yet. I’ve been reading that ICU u
426.
by rsfern 5y ago
> Covid has been decreasing in severity with each new major variety. That’s not accurate, delta was a marked increase in severity. Omicron is reduced in severity relative to delta, but it’s probably on par with the original strain once
427.
by rsfern 5y ago
Thanks for the link, it’s interesting that they have similar findings to other studies based on passive reporting systems like VAERS in the US. I don’t think the forever-recurring boosters are going to be a thing, it seems like cellular imm
428.
by rsfern 5y ago
Solid data is still hard to come by, but preliminary analyses from the week of Dec 20 have omicron as moderately less virulent than Delta, and on par (+/- 10-20%) with the original strain. Good discussion of various pieces of the pictu
429.
by rsfern 5y ago
Similar reports of what? I’m not saying that I don’t think myocarditis is a risk factor for kids and teens, just that this paper isn’t giving a quantitative assessment of its relative risk vs Covid. If you have seen research that does prov
430.
by rsfern 5y ago
The preprint discussed in that post specifically calls out that there weren’t enough myocarditis events in the 13-17 age group to reliably estimate the relative risk, so I don’t see where that conclusion is coming from Also, note the highe
431.
by rsfern 5y ago
I agree with you that many people are just echoing whatever they see on social media, but it’s easier than ever to read the primary literature for yourself. Synthesizing what you read in in introduction sections can be a good way to get a s
432.
by rsfern 5y ago
The recent flurry of papers on omicron trying to decouple the effect of its intrinsic virulence from prior immunity have been finding that prior infection (with delta) and unboosted vaccination have comparable risk reduction file hospitali
433.
by rsfern 5y ago
Surfing a bit, I found some slightly more detailed captions for this year at least. Maybe there are similar pages for previous years https://www.zyvexlabs.com/contests/2021-2/ Edit: there are links for previous ye
434.
by rsfern 5y ago
I agree with your point about epistemological hubris, but this isn’t about mandates, it’s about clear-eyed risk assessment. Without a specific hypothesis, it’s all a big game of “what if”. The scenarios you mention aren’t physically impossi
435.
by rsfern 5y ago
Obviously we don’t know everything, and we may never know with 100% certainty, but I disagree that we have “no idea”. We have prior knowledge from established vaccines about when to expect side effects, and what kinds are likely. Most known
436.
by rsfern 5y ago
Yeah, that is interesting. It does look like a large fraction of the population is going to get omicron regardless of vaccination. I’m not sure what the implications of a breakthrough infection are for this scenario, but I would imagine it
437.
by rsfern 5y ago
This is based on VAERS and also v-safe data, and the CDC has done some review. They don’t seem to be trying to estimate rates from this data, which is one of the more common errors made when looking at VAERS data Also, it’s a bit more than
438.
by rsfern 5y ago
Is there even any proposed mechanism that would give rise to long term effects (that show up months or years later) from this vaccine? I’ve never encountered any credible hypotheses, whereas we have plenty of preliminary evidence and propo
439.
Poor science reporting and premature enthusiasm(sciencebasedmedicine.org)
3 points by rsfern 5y ago|0 comments
440.
Internet Literacy Atrophy(lesswrong.com)
3 points by rsfern 5y ago|0 comments
441.
by rsfern 5y ago
I take your point, but by correct I meant “consistent with the established definition,” which was the original context we started in Like I said, maybe it’s not the optimal term to use in public health communications, and that could be beca
442.
by rsfern 5y ago
Ok, sorry if I misinterpreted your punctuation, the last instance is what read like scare quotes to me; maybe `backtick` is a better delimiter for this case > Most people think that `vaccine` has a specific meaning. It is dishonest to us
443.
by rsfern 5y ago
Thanks for the thoughtful reply. I do agree that the public discourse has not been very productive, and the news coverage has been overconfident and one-sided in one way or another, especially around omicron lately.
444.
by rsfern 5y ago
Where am I pushing mandates? I think there are valid policy positions on both sides of the issue, and that you’re misrepresenting the logic of those who still are in favor of vaccine mandates. You also haven’t offered any substantive counte
445.
by rsfern 5y ago
I agree that definitions matter, and it’s important to have a clear discourse on policy implications of public health research. But my point is that it’s incorrect to call the current Covid vaccines “not vaccines” just because they don’t 10
446.
by rsfern 5y ago
Cancer immunotherapy candidates have been referred to as cancer vaccines for a long time without controversy, and they are generally not intended to prevent cancers. They are medicines that work via an immune response. Sure, public health c
447.
by rsfern 5y ago
I think it’s generally true. Check out the discussion sections of the primary sources, they are usually pretty forthcoming about the limitations of the data and methodology. This in general has been what has kept my uncertainty factor high
448.
by rsfern 5y ago
There is barely any solid data on hospitalization yet; the first batch of preprints with enough signal to start assessing intrinsic virulence vs prior immunity and other covariates were all posted like less than a week ago, out of South Afr
449.
by rsfern 5y ago
This is the most detailed summary I could find [0] Basically there’s a private company, dsruptive subdermals, that has made NFC implants for a while. You can put whatever on there, including I guess your vaccine records, if you want. I don’
450.
by rsfern 5y ago
I will concede a misunderstanding perhaps of the income distribution in CA. But the figures you quoted are the marginal rates, and at the federal level that top bracket kicks in at over 500k in the limit of very high income, your effective

More ›