Metacelsus

7 ms·

1.
by Metacelsus 7d ago
Just because the author can't think of a way to cause mass death with an engineered virus, doesn't mean an AI won't be able to think of one. (There are several biological threats the author hasn't considered. I'll j
2.
by Metacelsus 8d ago
How can they "not rule out" that Tristan and Levent's data was used for training?
3.
by Metacelsus 10d ago
Whatever happened to Edge? They haven't really published since 2018
4.
by Metacelsus 15d ago
Seeing the recent increase in bond yields reminded me of this.
5.
AGI and the Efficient Market Hypothesis (2023)(lesswrong.com)
2 points by Metacelsus 15d ago|1 comments
6.
by Metacelsus 16d ago
>really top journals like Scientific Reports I wouldn't consider Scientific Reports to be a top journal, it's basically pay-to-publish
7.
by Metacelsus 21d ago
>Reward hacking has been present in AI systems both historically (see this work from a decade ago , figure shown below) I went to the page, and guess who it's by . . . Dario Amodei and Jack Clark!
8.
by Metacelsus 24d ago
Well, as a biologist, Fable is still completely unusable
9.
SpaceX tugs Starship into port after 24 days at sea(arstechnica.com)
6 points by Metacelsus 29d ago|1 comments
10.
Investigation casts doubt on anti-Alzheimer's gene variants(science.org)
3 points by Metacelsus 1mo ago|0 comments
11.
by Metacelsus 1mo ago
relevant: https://xkcd.com/1000/
12.
by Metacelsus 2mo ago
>However, later in the run, Claude realized that the compromised host sat in a cloud account with no connection to the capture-the-flag challenge. On its own, it concluded that the target was in fact real, and ceased its attack. Interest
13.
by Metacelsus 2mo ago
Holy crap. This is definitely not good
14.
by Metacelsus 2mo ago
If built in America it would easily cost 10x that!
15.
by Metacelsus 3mo ago
Time to bring out the gene drives!
16.
by Metacelsus 3mo ago
Sure, but at what levels? The dose makes the poison, and the article doesn't say
17.
by Metacelsus 3mo ago
Hmm, Science just had a news article saying the AMOC was doing OK: https://www.science.org/content/article/ocean-current-warms-...
18.
by Metacelsus 3mo ago
Yeah. With cancer, delivering to 99% of the cells won't cut it, the surviving 1% will quickly grow back.
19.
by Metacelsus 3mo ago
The actual study: https://www.biorxiv.org/content/10.64898/2026.05.30.728989v1 Abstract: Cas9-based tools enable the introduction of genetic lesions to investigate DNA repair outcomes and edit the genome at diseas
20.
In a First, Scientists Precisely Edit Human Embryo Genes(nytimes.com)
6 points by Metacelsus 3mo ago|1 comments
21.
by Metacelsus 4mo ago
Cool, I've been using Claude Code for this, but I may want to try Gemini
22.
Pushed by Trump policies, top U.S. battery scientist is moving to Singapore(science.org)
67 points by Metacelsus 5mo ago|37 comments
23.
by Metacelsus 5mo ago
It's always interesting to see East Africans doing so well. Even with technology like advances in shoes and diet/training, genetics is still a huge factor. Also it must be an crazy feeling to be Kejelcha, the guy who came in 2nd p
24.
by Metacelsus 5mo ago
See also the press release: https://hms.harvard.edu/news/massive-ancient-dna-study-revea... This study covers about 10,000 years of recent human evolution in Europe and West Asia. From the abstract: >in the past ten
25.
Ancient DNA reveals pervasive directional selection across West Eurasia [pdf](reich.hms.harvard.edu)
73 points by Metacelsus 5mo ago|70 comments
26.
by Metacelsus 5mo ago
If they're going to all that effort to make allele-specific guides why not just cut out the centromere and eliminate the chromosome entirely? This seems like an overly complicated solution.
27.
by Metacelsus 5mo ago
TIL Backblaze doesn't back up the /Applications folder on Mac! Time to move everything to ~/Applications
28.
by Metacelsus 5mo ago
I guess they're not Kerbals :)
29.
by Metacelsus 5mo ago
The name "mythos" seems a bit too eldritch for my liking. Brings to mind Cthulhu.
30.
by Metacelsus 6mo ago
>Legit password resets for example come from more random top level domains with "microsoft" in it, like microsoftonline.com Or aka.ms

More ›