nullc

28 ms·

661.
by nullc 1y ago
If you like crazy chemistry stories, be sure to read "Excuse Me Sir, Would You Like to Buy a Kilo of Isopropyl Bromide?". One of the first stories in it involve the high school chemistry club breaking into the local college to get
662.
by nullc 1y ago
I can't use mouser or digikey anymore thanks to anti-crawler crusading.
663.
by nullc 1y ago
Lots of bad and dumb things are done by young adults, it's not the states place to micromanage the personal decisions of adults.
664.
by nullc 1y ago
Epstein was just a con man, he wove a fog of mystery around himself as part of the con and people are still being duped by it to this day.
665.
by nullc 1y ago
The amounts involved are so small that I'm a little doubtful of this cashflow squeeze pattern actually working, but it doesn't have to work to drive markets. Fun example is the old fax pump and dumps. You'd get some 'ma
666.
by nullc 1y ago
Time for bloodletting.
667.
by nullc 1y ago
> No program can tell whether any program will halt or not. Might try for another wording which isn't easily misunderstood. I was going to suggest every instead of any but that supports a different misunderstanding.
668.
by nullc 1y ago
typical room temperatures is doing a lot of work there.
669.
by nullc 1y ago
> This is also a reason why much of the traditional "optical computer" research was fundamentally flawed presumably also because photons at wavelengths we can work with are BIG
670.
by nullc 1y ago
You could coat them with an ultraemissive material and point them at the sky, if its not cloudy it will get significantly below the ambient air temp. But the paper suggest that it will condense at ambient anyways, because it gets warmer so
671.
by nullc 1y ago
Kinda backwards there, _billions_ of dollars a year are stolen from US residents that have the money in banks, the FTC says 12.5 billion last year. It's so easy that you don't even have to apply a wrench.
672.
by nullc 1y ago
The real danger is generating hysteria to justify a regulatory crackdown that restricts the public's access to AI and, conveniently, limits their competition.
673.
by nullc 1y ago
As the 90s commercials said: "I learned it by watch YOU" Pull the thriller and evil ai crap out of the training material and most of this behavior will go away. There is no particular reason for some coding assistant AI to even kn
674.
by nullc 1y ago
They did that initially, then the FDA started threatening companies that did that, so they responded by actually adding the sesame.
675.
by nullc 1y ago
Galvanizing still works if there are scratches in the finish because there is a radius of protection from the galvanization.
676.
by nullc 1y ago
My mistake, in the future I'll refrain from using Toki pona for making a rhetorical point. :)
677.
by nullc 1y ago
The AI doom hysteria is a big enabler for this kind of control, imagine if google admitted that a major goal of google search was to influence people's thinking according to their objectives? And on top of that was lobbying to make i
678.
by nullc 1y ago
It can happen more or less no matter what language the model uses, so long as its reinforcement trained. It's just in English we have an illusion of thinking we understand the meaning. An example of this is toki pona, a minimalist con
679.
by nullc 1y ago
In some sense it's more human than a model trained with no RL and which has absolutely no exposure to its own output. We have our own personal 'culture' too-- it's just less obvious because its tied up with our own hidde
680.
by nullc 1y ago
Does it require that all exit conditions of a loop are taken/not-taken? I've found lots of bugs around if(foo)break; logic.
681.
by nullc 1y ago
Even when you train AI on human language, the tokens can have "subtext" that is only legible to the AI. And, unfortunately, it's not even legible to the AI in ways that it could ever explain it to us. It's no different
682.
by nullc 1y ago
> whims of the backwards ultra hard right conservative leadership at the credit card companies the credit card companies executives like money, this isn't originating with them. It originated in the US federal government. (e.g. htt
683.
by nullc 1y ago
Would be a nice fine tuning target: training to close that loop.
684.
by nullc 1y ago
"Safety Levels" will continue to increase until state power is deployed to suppress competition.
685.
by nullc 1y ago
> It's not clear why openssl never adopted the patch OpenSSL already had an option to safely disable the bad behavior, -DPURIFY.
686.
by nullc 1y ago
The crazy thing is that after this incident they restored the uninitialized usage and retained it there for the next half decade. It wasn't as mild as being a risk of future compilers destroying the universe: it made valgrind much les
687.
by nullc 1y ago
> When asked to explain rationales, these LLMs are observed to lie frequently. It's not that they "lie" they can't know. LLM lives in the movie Dark City, some frozen mind formed from other peoples (written) memories
688.
by nullc 1y ago
> The issue is likely that left-wing people are attracted to professions whose primary output is words rather than things. Actors, journalists and academics are all quite left-biased professions whose output consists entirely of words, a
689.
by nullc 1y ago
I am not surprised to see grok failing on this. Practically everything gets trained on extracts from other LLMs, I assume this is true for Grok too. The issue is that even if you manually cull 'biased' (for whatever definition you
690.
by nullc 1y ago
I find ChatGPT is a fairly mediocre writer: It uses a lot of overwrought and exaggerated language that becomes fatiguing rather quickly. It's also frequently used by fraudsters and in spam, so you run a real risk of 'your' w

More ›