entee

7 ms·

1.
by entee 1y ago
A lot of this post relies on the recent open ai result they call GDPval (link below). They note some limitations (lack of iteration in the tasks and others) which are key complaints and possibly fundamental limitations of current models. Bu
2.
by entee 3y ago
I think what’s missing is what the software allows. It could be BMW/Merc etc are way more conservative on what the allow the system to do and when they force the driver to take over. In certain contexts Merc is actually willing to asse
3.
by entee 4y ago
Transcription factors often are partially disordered, just to name one. A bunch of others here: https://www.nature.com/articles/nrm3920
4.
by entee 4y ago
> It’s only marginally less useful to actual biology than full on X-ray structures anyway. I'm not sure what you're implying here. Are you saying both types of structures are useful, but not as useful as the hype suggests, or t
5.
by entee 4y ago
Such a database would be hugely helpful across chemistry. Right now it’s extremely expensive to access databases like Reaxys or Scifinder, and they’re not usually programmatically searchable at scale. Some databases do exist based on the pa
6.
by entee 5y ago
As a fellow perfectionist who has started a company, one thing that has helped me is realizing that most decisions are a lot more reversible than they appear. Even in the legal and financial domain, most things that you might obsess over ar
7.
by entee 5y ago
There's a lot that can be learned with building-block based experiments. If you do a building block based experiment then train a model, then predict new compounds, the models do generalize meaningfully outside the original set of buil
8.
by entee 5y ago
This is true. Getting datasets with the necessary quality and scale for molecular ML is hard and uncommon. Experimental design is also a huge value add, especially given the enormous search space (estimates suggest there are more possible d
9.
by entee 5y ago
Not a chiphead, but saw this in the article that might be a reason ARM is better for this kind of thing: "The theory goes that arm64’s fixed instruction length and relatively simple instructions make implementing extremely wide decodin
10.
by entee 5y ago
Most of the methane comes from cattle belching, so it's basically impossible to harvest at scale. The lagoons of manure also do, and that can be harvested though it's hard. https://climate.nasa.gov/faq/33&#x2F
11.
by entee 5y ago
The catalyst is the part I'm most curious about. Carbon capture is a hard problem at scale. Once you have methanol, I'm not terribly surprised that you can convert it to other things enzymatically. Hard to evaluate without full te
12.
by entee 5y ago
I'd agree if it works as advertised. Level 5 or very close to it is the key. No system has shown that, much less Tesla's. In the meantime, doing a live experiment with 3,000+ pound machines moving at 30+ mph seems like a bad idea.
13.
by entee 5y ago
I drive a lot, thanks. If you can prove me level-5 or even very good level 4 autonomous driving, and that a computational driver makes radically fewer fatal mistakes than a human, then I'm with you. In other words, if you can satisfy a
14.
by entee 5y ago
I'm not sure what you're arguing in terms of acceptable risk. Biotech is incredibly regulated, specifically because the risks are so high, effectively there is very little acceptable risk. In biotech, a patient dying due to your d
15.
by entee 5y ago
It’s pretty clearly the current best way to do deep learning on molecules in chemistry. Transformers do really well in predicting reaction outcomes but GCNNs are still better when it comes to “is this a good molecule according to this label
16.
by entee 5y ago
We already do all of this in drug discovery research before putting things into animals. We try computer modeling (very inaccurate but a good pre filter) then cell culture, and only after that do we use animals. Not only is it ethically the
17.
by entee 5y ago
Even if the amyloid hypothesis is true (and most neuroscientists I know think it's not), this is a terrible decision. 1.) We don't know when to give the drug. Maybe giving it even earlier would help, but these trials don't te
18.
by entee 5y ago
That linked post is fundamentally wrong because it rests on an assumption that has essentially zero human evidence: amyloid causes Alzheimer’s. There have been several drugs that very efficiently reduce amyloid, strictly zero (including thi
19.
by entee 5y ago
Just got a postcard from my optometrist the other day saying it’s time for my annual prescription check. My eyes are no different, but my glasses are scratched, so I need to spend an hour or two to get this fixed instead of just reordering
20.
by entee 5y ago
My biggest worry with this would be the low level of off target edits and the number of recombination events that yielded an unwelcome product. Looks like those were very low, but with an N of 4, hard to know long term. The reason being tha
21.
by entee 5y ago
The first iPhone famously only had 2G cellular internet. At the time 3G phones were not ubiquitous but were readily available (had one as a mid priced Nokia flip phone), so I’ve always thought this was a fascinating aspect of the iPhone sto
22.
by entee 5y ago
A couple other perhaps more useful quotes from the same article: “Despite Trump’s assertions, few close observers of Obama’s and Biden’s response to H1N1 consider it a “full scale disaster.” And Biden, despite his early messaging problems,
23.
by entee 5y ago
That’s a pretty disingenuous link. The Bush administration did have a plan for Flu and was thinking about the issue. In 2015 (your article recounts stories of H1N1 in 2009) the Obama administration established NSC level officials responsibl
24.
by entee 5y ago
I can't upvote this enough. We HAD a plan, both Bush and Obama put together pandemic response plans. We had serious issues in logistics and deployment for sure. We had some screw-ups in early testing for sure. Maybe we should grant som
25.
by entee 5y ago
Those who think the CDC has done way too much because the virus was overly-exaggerated are simply wrong. I have my quibbles with the CDC, especially early reactions and subsequent messaging. I don't think people who say it exaggerated
26.
by entee 5y ago
I don't think that's quite right. Even if you have a flu strain with a particular name (H1N1) there are a number of variants within the H1 and/or N1 proteins that the vaccine will be targeted to as well, so the number is quit
27.
by entee 5y ago
It has very short half life, there’s not a ton of it and it’s likely to be wrapped around a nucleosome. Uniformly, sequences in blood are quite short (<200bp) which is far too small to code for anything meaningful (for reference COVID ge
28.
by entee 5y ago
One thing to note is that mRNA therapeutics are a really tough area in part because: 1.) RNA and the lipids used to get it inside a cell are inherently pretty immunogenic (a huge plus for a vaccine). If you think about it, you basically nev
29.
by entee 5y ago
Agreed. Boosters will probably be required yearly or so, and those will then be the “second shot” for lots of people. The only caveat is that partial immunization can create the conditions for viral escape, but hopefully we’ll have boosters
30.
by entee 5y ago
Burying the lede: “ While millions of people have missed their second shots, the overall rates of follow-through, with some 92 percent getting fully vaccinated, are strong by historical standards. Roughly three-quarters of adults come back

More ›