yellowcake0

"We’re nowhere near the end of the shortage,” said Brad Gastwirth, Circular’s head of global research. “I think it lasts through the end of 2026 and at least the first half of 2027.”

You heard it here first folks, close your memory positions Q2 2027 at the latest to avoid getting caught flat-footed when prices drop.

Enterprise is what determines commoditization as that's where the lion's share of the revenue comes from. Maybe it will eventually happen to MS Office, maybe it won't, but until it happens it hasn't happened. Access to the full MS ecosystem, technical support and seamless integration, these things matter a lot to businesses, and I'm not even saying you're wrong, but I'm not convinced yet that something similar won't play out with coding agents.

LibreOffice didn't replace MS Office, and Octave didn't replace Matlab. It seems to me that there is even less of a moat with these products than there is with Claude Code, yet neither was commoditized.

The economics of bitcoin mining dictate that the work must have no other utility. If you increase its profitability by using it for auxiliary winter heating, then more people will mine bitcoin until there is an oversupply of heating and we return to the current equilibrium amount of "wasted" heat.

I'm not sure OpenAI can realistically afford not to IPO given its spending commitments.

I'm just trying to get an idea of what the operation is such that man - man + man = woman, but it's like pulling teeth.

I think you misinterpreted the tone of my original comment as some sort of gotcha. Presumably you're overloading the addition symbol with some other operational meaning in the context of vector embeddings. I'm just calling it addition because you're using a plus sign and I don't know what else to call it, I wasn't referring to addition as it's commonly understood which is clearly associative.

It's a pretty exotic type of addition that would lead to the second set of examples, just trying to get an idea of its nature.

so addition is not associative?

Isn't returning to an era of chip architecture experimentation exactly what would be required to explore new and better alternatives?

Are you not based in the US or unable to relocate? I understand the tech job market is at a low point right now but there are still enough programming jobs across the entire continental US that you should be able to fill 40 hours a week applying to them + auxiliary tasks such as coordinating interviews, interview prep, working on your resume, etc.

Like others have said you need to switch to employment mode, and finding a job is your full-time job

AI is just a tool to help you do your job, like an IDE, or stackoverflow, why do you have so much antipathy towards it?

apparently Seattleites have more sense

are these Russian bots in the room with us right now

the killing will stop, and Russia gets to keep the eastern oblasts.

can't say it was doing the planet much good in any case

no thanks

Information entropy is literally the strict lower bound on how efficiently information can be communicated (expected number of transmitted bits) if the probability distribution which generates this information is known, that's it. Even in contexts such as calculating the information entropy of a bit string, or the English language, you're just taking this data and constructing some empirical probability distribution from it using the relative frequencies of zeros and ones or letters or n-grams or whatever, and then calculating the entropy of that distribution.

I can't say I'm overly fond of Baez's definition, but far be it from me to question someone of his stature.

The disinformation landscape of this conflict is insane. This article paints a decidedly optimistic picture, however I fear the reality is far more bleak.

Unreliable casualty numbers aside, Russia simply has way more men, and can eventually attrite Ukraine into a demographic death spiral. Indeed, with the exodus of ~20% of its population since the start of the war, this demographic trajectory may now be inexorable. Eastern Ukraine, where much of the most fertile land is, and where most of the fighting has occurred, is now likely littered with bomb parts and unexploded ordinances, and the prospect of the jackals of the American arms industry turning it into a live weapons laboratory is particularly grim.

but it will give you some sleepless nights!

The variants in ClinVar will have genomic coordinates associated with them, each given with respect to a particular reference genome, e.g.

GRCh38 - chr2:41785474-41785474

or something like that.

You will need to align your FASTQ to one of these reference genomes using a short-read aligner such as BWA-MEM (assuming your genome was sequenced using something like Illumina NGS) to produce a BAM (binary alignment map) file, and then from there use a variant calling tool such as GATK's HaplotypeCaller or Google's DeepVariant to produce a VCF (variant call format) file, which will contain the mutations in your genome. Then you can see if any of the mutations found in your genome match those in ClinVar.

As humans are diploid, you will also need to be cognizant of the zygosity of each mutation, i.e., whether you have a particular mutation on just one of your two chromosomal homologs (heterozygote) or both (homozygous mutant), as this may affect whether or not the mutation is of concern, the variant caller will attempt to determine this for you and report it in the VCF file.

Unfortunately, the preponderance of very low value research in the literature puts a significant burden on the scientists who have to sift through a lot of garbage to find what they're looking for. Even if the work is ostensibly correct (much of it is not), it really doesn't do anyone much good, except for the authors of course. But now every undergraduate and every parent's vanity project at Andover wants a first author contrib., so here we are.

It looks like from the preprint that they sequenced the Y chromosome of HG002, which was one of the original 1000 genomes samples from way back in the day, still held in deep freeze at a number of biobanks.

Short-read sequencing data is a notoriously bad datatype for reconstructing the low-complexity / repetitive regions of genomes, so up until recently the most commonly used reference genomes have left many of these regions "dark". According to the preprint, the Y chromosome has the highest density of these low-complexity regions. It's also something of a bioinformatic nuisance when constructing a generic human reference genome, as it's only present in 50% of the population.

from months to a few weeks

So a 4x speedup? A completely irrelevant performance improvement for a "quantum computer".

maybe the Treasury bond market remained unperturbed but the SP500 dropped 7%, I would not call that investor calm.

I don't think he claims it's levitating does he? It's just standing up on its end I think.

There's a lot of optimism in this thread, but does DFT (or any theoretical model really) actually have much predictive value in quantum chemistry? I've always gotten the impression that in this field the proof is in the pudding.

That's only for prehistory.

In any case it's a pretty silly contrivance, how about the oil age? or the silicon age? Both of those are likely to remain more impactful than the LK-99 age, in the event that it does turn out to be an ambient temp/pressure superconductor.

it's not confusing, it's just silly, the Iron Age ended over 2000 years ago, so there were just no ages between the birth of Christ and LK-99?

this doesn't even make sense