scheme271

6 ms·

61.
by scheme271 3mo ago
There might be "one true diagnosis" but there's no reason to believe that we'll have practical diagnostic tools to get it. If we need to sample the brain chemistry to diagnose a neurochemical disorder, it's probably
62.
by scheme271 3mo ago
There's quite a few diseases and conditions that don't have definitive tests. For example, alzheimer's and parkinsons are diagnosed based on medical history and symptoms. With alzheimer's an autopsy can tell for sure but
63.
by scheme271 3mo ago
The problem is that the diagnosis might not be known for a while. There's a few conditions and diseases that require an autopsy for a guaranteed diagnosis and therefore are diagnosis based on symptoms in clinical settings.
64.
by scheme271 3mo ago
I'm not sure how anyone can take that assertion in good faith. It's like saying that you want to bomb a country back to the stone age and then saying that you didn't realize that it meant attacking civilian infrastructure.
65.
by scheme271 3mo ago
The problem is how do you know whether the answer is just the most persuasive or actually the most accurate one? It's hard to figure this out without domain knowledge.
66.
by scheme271 3mo ago
I'm pretty sure that the reputational harms of using nuclear weapons on a country of 70M is going to be pretty severe. Surely you understand how starting a war with a country without sufficient cause and then using nuclear weapons on t
67.
by scheme271 3mo ago
Andrew Jackson was around 150+ years ago. Nixon was fairly competent diplomatically even if he was horrible domestically. Trump is a whole different breed that seems to switch his actions based on whim and who he spoke with last.
68.
by scheme271 3mo ago
Mochizuki and a group of mathematicians still claim that Stix and Scholze didn't actually identify a flaw and his proof was published in a journal (where Mochizuki is the chief editor and the reviewers were people from Mochizuki's
69.
by scheme271 3mo ago
Or alterntively, that it's a relatively modern addition to add to it's mystique for tourism purposes. When did SPQR get added to the manhole covers and does anything else in Rome use similar signage?
70.
by scheme271 4mo ago
Nah, I doubt they're going to rebase to openzfs. There's too much divergence and I don't see them putting the time to write something that converts their zfs format to openzfs without an extremely good reason.-
71.
by scheme271 4mo ago
I think the ZFS changes were due to needing a way to allow qnap systems to expand zfs pools. The raidz expansion features in openzfs probably took too long for qnap to wait.
72.
by scheme271 4mo ago
ECC support depends on the processors that the NAS uses. A few of their NASes allow you to use ECC memory but you'd need to swap the memory installed to ECC memory. A lot of their systems use Intel cpus that don't support ECC at a
73.
by scheme271 4mo ago
There's a HN article and discussion about Anthropic expanding to use Colossus 2. https://news.ycombinator.com/item?id=48214017 I think it's fairly clear that grok isn't using as much compute as expected.
74.
by scheme271 4mo ago
It's more than that though. Any repairs due to wear and tear or whatever, ends up being really expensive. Although you can probably reduce the costs a bit if you get the non-branded OEM part or potentially the same part from another ma
75.
by scheme271 4mo ago
I don't think they can withdraw a model easily, they also supply Pis to industrial markets and they probably have contracts that guarantee supply for a given time period.
76.
by scheme271 4mo ago
I thought anyone serious about mining switched to ASICs and custom hardware a while ago. At least for BTC and maybe eth, a custom asic setup is ~2-3x more profitable
77.
by scheme271 4mo ago
It could be that the instruments were looking at the differential pressure. E.g. the pressure in that section compared to another. A leak elsewhere would throw off the comparison.
78.
by scheme271 4mo ago
I think it's more some security company writing about a vulnerability they discovered in this module or a worm/backdoor and not the company that wrote the software. The security company gets publicity and potentially gets more biz
79.
by scheme271 4mo ago
It is and isn't. I think even if the germans had switched to making mass quantities of Sherman or T42 style tanks, it would have failed because they had shortages things like labor and fuel. The allies essentially had huge industrial b
80.
by scheme271 4mo ago
The German tanks were superior when functioning but quite a few of them were difficult to produce, maintain, and use. Do better armor and weapons really imply a superior tank if it requires significantly more maintenance and breaks down mor
81.
by scheme271 4mo ago
That's better but people typically won't appreciate a thermonuclear explosion overhead in order to take out a missile armed with a few hundred kilos of explosives. That is doubly the case if the nuke is from a neighboring country
82.
by scheme271 4mo ago
Brookhaven does classified research and access to the site is pretty restricted. The last time I visited there were guards with assault rifles at the gates. In comparison, Fermilab used to (and may still) let you walk onto the campus and wa
83.
by scheme271 4mo ago
Depends on what you're being attacked by. If it's just regular warheads, a nuclear interceptor is wildly inappropriate especially if the intercept happens over someone else's airspace.
84.
by scheme271 4mo ago
The interesting part is that he got charged with insider trading. A few attorneys on bluesky have pointed out that this is a novel use of this law since previously the trading occurred on regulated markets (e.g. SEC or CFTC regulated market
85.
by scheme271 5mo ago
LPDDR is used by the Nvidia rubin platform. I can see AMD using lpddr as well because it gives you denser memory at the same or reduced power budgets.
86.
by scheme271 5mo ago
Depending on the person, they may have spent years or possibly a decade or more in the US before trying to get an AOS. As such, the USCIS is probably a better adjudicator than a consulate.
87.
by scheme271 5mo ago
I was being pretty generous to the comment I was replying to. Needing 32+ H100s just strengthens my argument that people aren't going to run frontier models locally anytime soon.
88.
by scheme271 5mo ago
We don't know the parameters but it probably takes at least a H100 and possibly several to run a SOTA model. Given the pricing (25+k per H100 + hardware to run it) and power (700W per H100 + hardware to run it), I don't see how an
89.
by scheme271 5mo ago
ptyxis has a few features that gnome-terminal doesn't and which are really handy. Namely, being able to list containers running on the system and then being able to select one to get a terminal running inside the container. Not sure th
90.
by scheme271 5mo ago
I think the fees are on order of 1-3% depending on your risk and business type. Certainly an issue but it's mitigated a bit by the decreased costs of going cashless. I.e. cashless operations avoids theft by employees; overhead for stoc

More ›