One of the most egregious cases of this, in my opinion, is only publishing metrics that cover part of the confusion matrix. “The false-negative rate? That could not possibly matter for a variant effect prediction model; why would we include that in the paper?” Example: AlphaMissense.
I really want to like Julia. It has a nice type system, a good ecosystem, reasonable syntax, and it’s far faster than Python. But there are several issues with its DX that prevent me from using it: the most severe of which being the complete lack of a cache for the JIT (or JIT like system), inducing multi second compile times for scripts that run in <100ms. There are some external packages that try to solve this issue, but they are far from first-party quality, and, in my experience, are quite buggy.
(I last tried Julia a few years ago; perhaps this has been improved since?)
pkgimages are keyed on package (and dependency) version so if you frequently add or update packages they don't help as much as they should. this new cache is at the function granularity and caches the llvm optimization and native codegen steps.
I imagine this will only become more important as we have more agents working on codebases in parallel; so a clean solution would be extremely valuable.
On the flip side, the rise of agentic coding means that the library/batteries mismatch against python will largely stop mattering. It should be very easy to have an agent implement large libraries -- especially if they can be just a translation from one language to another (esp. one with with better primitives!)
The benefit is probably from the removal of fat, not a direct antibacterial/antiviral effect. Fat plays a complex immunoregulatory role in human physiology: it down-regulates some pathways, while up-regulating others (notoriously, the production of IL6 is carried out, in part, by adipocytes). The overall effect of fat on the immune system, however, is negative: it tends to increase the chances of rheumatological disorders, cancers, and many other diseases. Alternatively, the effect may be due to some sociological factor that their analysis failed to account for.
Nobody's denying this, but the interesting story about GLP-1s is that after you control for fat loss (for instance, by taking cohorts of patients that aren't losing weight) you still get evidence of these off-target effects.
The official ARC-AGI 3 score—-without OpenAI’s custom harness—-can be found here: https://arcprize.org/leaderboard. Astra scores 62.7% at max reasoning for the low-low price of 26,000 dollars.
I’m no expert, but I don’t believe that “high-assay low-enriched uranium” is the sort you can use in nuclear weapons, per it being “low-enriched”. According to the DOE, HALEU only goes up to 20% enrichment [0]. Nuclear weapons, it seems, need it to be enriched beyond 20% [1]
Uranium-235 isn't useful for dirty bombs: it's too long-lived and not radioactive enough. Making dangerous levels of radiation requires short-lived isotopes. Some of these are the products of uranium fission, which is why spent uranium fuel is a radiation hazard. But isolating U-235 just to fission it to get radioactive isotopes like cesium-137 would not be the easiest way to get them.
It is when your goal is to throw a ton of alpha emitters into the environment for the population to ingest. Do you remember your nuclear cookie thought experiment?
It emits so slowly that the main hazard is heavy metal poisoning, which you'd also get from the depleted uranium that's a waste product from enrichment.
From the numbers I see on https://www.iaea.org/topics/spent-fuel-management/depleted-u..., if a dirty bomb with 1kg of 100% pure Uranium-235 exploded and evenly deposited a fine dust over a 170 by 170 meter square of typical soil, the total Uranium-235 (and specifically that isotope) of the top meter would only double; the total mass of uranium in that volume would only go up by 0.7%.
The average concentration of natural uranium in soil is about 2 parts per million, which is equivalent to 2 grams of uranium in 1000 kg of soil. This means that the top metre of soil in a typical 10 m ´ 40 m garden contains about 2 kg of uranium (corresponding to about 50,000,000 Bq of activity just from the decay of the uranium isotopes and ignoring the considerable activity associated with the decay of the progeny. Concentrations of uranium in granite range from 2 parts per million to 20 parts per million. Uranium in higher concentrations (50 - 1000 mg per kg of soil) can be found in soil associated with phosphate deposits. In air, uranium exists as dust. Very small, dust-like particles of uranium in the air are deposited onto surface water, plant surfaces, and soil. These particles of uranium eventually end up back in the soil or in the bottom of lakes, rivers, and ponds, where they mix with the natural uranium that is already there. Typical activity concentrations of uranium in air are around 2 µBq per cubic metre. (UNSCEAR 2000).
I agree that purchases on e-commerce sites are virtually never this fast; but I have used some extremely fast physical payment terminals (e.g., Square): sometimes it seems like the transaction is sub 300ms, which I find very impressive given the complexity of payment processing.
I'm honestly quite shocked that the physicians/scientists involved would choose to use an AAV for a brain-targeted gene therapy. There is just so much data demonstrating that these vectors are quite immunoreactive: most of the approved gene therapies based on AAVs carry black box labels for liver failure caused by an immune reaction to the viral capsid. Admittedly, AAVs are the most derisked vector for gene therapies, but infusing them directly into someone's brain and expecting nothing bad to happen is, in my view, crazy.
[Note that I work in this field and have co-founded a CNS AAV company]
This isn't the correct takeaway. AAV are one of the most complex drug modalities and carry considerable risk when used incorrectly. This story is tragic and violates pretty much every ethical consideration for a clinician researcher. Especially ones that are treating children of desperate parents.
That said, AAV are one of the most powerful delivery mechanism we have to deliver gene therapies to the brain. Uniqure has shown the first efficiacious treatment of Huntington's disease with intraparenchymal delivery of AAV5, Zolgensma is a brain targeted AAV9 to treat SMA, Kebilidi is an intraparenchymal AAV2 that treats AADC deficiency.
The general approach should be to keep dose as low as possible and minimally expose the periphery. AAV9 at large doses delivered intrathecally without standardized immunosuppression is simply insane.
I agree with you that AAV is clearly a useful tool but at times the low levels of self discipline and scientific rigor that the field has applied to dosing is disappointing at best and scary at times. This includes use of other vector types such as by bluebirdbio and even the Jesse Gelsinger tragedy. “If a little is good then more is better” is a crazy and lazy way to apply and optimize these technologies.
Another compounding issue is that they had to package the vector into two parts, which then have to both infect the same cell to get any effect. Which means you have to at least double the dose to get similar coverage compared to a single AAV vector (and actually more than double). Seems like a easy recipe for liver toxicity. Which is why most companies doing AAV therapy either target the liver or the eye (where AAV doesn't escape to the liver).
I feel like the parents were not well enough informed of the risks, and the PI rushed the therapy to be famous. Not the first time this has happened, and not the last, sadly.
These are all fair complaints, but DB-OTO works the same way for the otoferlin-related hearing loss: it's packaged on two viral AAV vectors. It's given to much smaller babies, though there are some Chinese reports of a similar gene therapy in teens. So it's clearly not just the dual-vector but also the target etc.
IANAMD. I remember something about that virus in the brain are super bad, and there is an additional protection to avoid virus and random substances entering the brain. They injected the virus in the medula, that is inside the protection membrane. In an ideal case, is it a good idea?
reply