Your comment is a stark reminder that I take for granted that the single biggest killer of human beings in history is simply something I don't have to think about. For some reason, I'd have thought there'd be no overlap between HN commenters and malaria - shows my blind spot.
If what you say is accurate (that malaria can remain dormant) then I agree that I don't see how DNA is proof of how they died. Certainly, regardless of how they died the malaria didn't help, and a double infection is likely even worse, but yeah if you have it chronically then DNA alone is insufficient.
Yes, which is why the immunological tests done in 2006 were not seen as conclusive. However, this time they found actual plasmodium DNA which suggests active infection at time of death. Including from two species, suggesting double infection. Fun.
>In order to use Ask Sage, the Army had access to 100,000,000 tokens as part of an annual subscription to an “enterprise pack.” Tokens represent a unit of output, either in text or image, from an LLM. For the Ask Sage tool, a single token equates to about 3.7 characters, according to documents viewed by WIRED. The Defense Department burned through some 20 billion tokens per day during the 38-day Operation Epic Fury in Iran, according to Breaking Defense.
I'm confused. Access to 100M tokens? That's trivial, I could burn that in a month or two and I'm not even doing anything very serious. But then how did they burn through 200x per day? These numbers don't make sense.
>The Army employee says that the Army has been pushing its workers to lean into using generative AI. Employees were given an allotment of at least 200,000 tokens per month, according to emails viewed by WIRED, and were automatically allocated more if they burned through their initial allotment. Employees who had signed up for Ask Sage but were not regularly using it would receive emails encouraging them to use more of their allocated tokens.
Great piece. Was wondering whether to get an NVLink bridge for $500+ for my two A4500s for inference. This pretty much confirms it only really rescues you if you have nonsensical PCI-E topology.
I've had cells growing fine in 20 L Cytiva wave bags and then fail to grow in 20 L Sartorius wave bags. Anyone that tells you they know how a cell grows is lying to themselves :)
Because it's completely meaningless without validation, and even with validation, not really any better than the state of the art protein generation models. Which are also mostly just nice to have because coming up with a candidate is generally quite easy.
The rate limiting steps are generally testing, or characterizing. Not designing protein binders.
I'm not a MAGA or MAHA person. I just hate anecdotal pontificating about science.
First of all, regarding the trans-fat discussion - in general, yes, keep trans fats low. However there are a couple important things to consider. One is that not all trans fats are created equal, and trans fats from animals are generally found to be less dangerous:
>. We found no relationship between R-TFA intake levels of up to 4·19 % of daily energy intake (EI) and changes in cardiovascular risk factors such as TC:HDL-C and LDL-cholesterol (LDL-C):HDL-C ratios
(One author is from a dairy group, but that doesn't invalidate the data. Unfortunately this is par for the course with nutritional literature, a huge amount of it is "sponsored").
Another small sidebar is that there is of course the chance that monounsaturated turn into trans fats as well, and presumably those developed by seed oils would be riskier than those found in animal fats. But the data on that are sparse-to-nonexistant.
The other thing that irks me here is the typical dietitian take is to see everything through the lens of food. It makes sense when you deal with cardiovascular patients, but cardiovascular patients are already already pre-selected for genetic risk, that represents up to or even greater than 90% of the signal in CV events. CV events are way more visible than whatever supposed systemic inflammation omega-6s provide, but it doesn't meant that they should be the sole guiding factor in policy. If anything, they are over-represented relative to more chronic effects.
I'm not saying that there's some easy answer, just this whole article was annoyingly hand-wavy about science that we can actually mostly track.
Presumably they have unused compute in those hours and figure they may as well enable people to use it and get more invested into their ecosystem.
What I wish Anthropic would do is be a lot more explicit about what windows apply when. Surely they have the data to say "you get X usage from hours A to B, Y usage from B to C"
>Vaccinated individuals were older than unvaccinated individuals (mean [SD] age, 38.0 [11.8] years vs 37.1 [11.4] years), more frequently women (11 688 603 [51.3%] vs 2 876 039 [48.5%]) and had more cardiometabolic comorbidities (2 126 250 [9.3%] vs 464 596 [7.8%]).
This is interesting because of "supposed" cardiovascular effects of the vaccine that many folks were worried about. Even more confounding is the gender differences. You'd think skewing women would skew away from cardiovascular issues.
An alternate interpretation is that the at risk cardio unvaccinated died of COVID for some reason.
>Government actions that restrict the ability to privately own or make use of computational resources for lawful purposes, which infringes on citizens' fundamental rights to property and free expression, must be limited to those demonstrably necessary and narrowly tailored to fulfill a compelling government interest in public health or safety.
I work in drug discovery (like for real, I have a DC under my belt, not hypothetical AI protein generation blah blah) and had the opposite experience reading it. We understand so little about most drugs. Dialing out selectivity for a closely related protein was one of the most fun and eye opening experiences of my career.
Of course we've thought of all these things. But it's typically fragmented, and oftentimes out of scope. One of the hardest parts of any R&D project is honestly just doing a literature search to the point of exhaustion.
On the technical side, yes. The biggest new developments I can quickly think of are:
1) Cold field emission guns. The big challenge of an electron source is producing a coherent beam - that is a beam that comes off the tip one electron at a time, at the same location, the same angle, and with the same energy. The cooler the tip runs, the more coherent it tends to be. This has made a big difference and is just now widely commercially available.
2) Narrow pole-piece gap. The sample on most TEMs sits sandwiched between two objective lenses that operate in tandem - these are typically called twin objectives. The upper one ensures the beam is parallel, which primarily results in uniform defocus (or focus if one so desires) across the image. The lower one is responsible for image formation and initial magnification (actually, all of your resolution essentially). The gap between them is responsible for your primary aberrations: spherical and chromatic. Reducing this gap reduces the total aberrations in the image.
I will side bar that the physics of a microscope are not really holding it back from what I'm doing - generating structures of biomolecules. Really, I'm more limited by the camera technology than anything, because the cameras simply aren't performant enough to dose the images to the level I'd like, to collect as many images as possible in as short a time as possible. Fundamentally, I tend to be limited by number of observations.
For the really cutting edge stuff...check out ptychography:
>How does one become a microscopist as a profession? It seems like a specialized field with a narrow entry point and a lot of hoops.
There are basically two routes for TEM - material science, or biochemistry. The way to become a microscopist for me was to show up at a University that had a grant for a microscope, but no one to operate it. :)
In general, universities operate TEM cores, frequently called bioimaging or something. (Structural biology if it's newer although that's just one application among many). Frequently there are positions for all education levels - bachelor's through PhD, depending on what one wants to do. Training is a mix of hands on (interfacing with complicated systems) and theoretical (physics and image formation). Typically the operators aren't the most theoretical, but have a lot of very niche practical knowledge you only get from being around broken microscopes.
A lot of replies that are mostly true, or somewhat true, or simply missing the real reasons.
There are two factors here:
1) Vaccine-derived immunity is a function of the individual's immune response, which in general, weakens significantly with age. It is not unrealistic for a vaccine to simply fail to elicit any response in someone old enough.
2) It is very, very difficult to recruit folks without HPV that are over 40 for a clinical trial. Most people of that age, who were never immunized, most likely have had it. This significantly convolutes the signal.
3) This is all especially confounded once something becomes "standard of care". Every year there are fewer and fewer people age 40+ with HPV.
For these reasons, the vaccine is currently officially ??? in people over 40. Most doctors will prescribe it anyways if you ask. It may or may not infer immunity. It almost certainly will not harm you.