Forgot your password?
typodupeerror

Comment Re:More uncanny, worse agents, allocated less (Score 1) 192

Imagine you collected all interactions of every black person and every white person who ever lived. If you then analyzed the data of these two groups you would notice differences in collective information because racism exists. It's not a claim that one is more or less racist, only that differences inherently exist, particularly differences regarding racial viewpoints. It is human nature to pick a side.

Now, train an AI to properly simulate a black person and a white person. These simulations would reflect the differences above, otherwise the simulations would not be "proper".

Now, run the same test except with black and white instead of male and female. If the AI is told what ethnicity is simulated the responses will be affected. If it does not, then responses won't match ethnicity. Either way, the test subjects will prefer one over the other because...racism exists. The outcome of the test is not in doubt.

That's exactly the problem here, even with male vs. female. The problem with the gender gap is not the victims, so simulating the victims does not make the problem disappear.

Comment Re:More uncanny, worse agents, allocated less (Score 1) 192

Thanks, these are good insights. I would suspect this, didn't want to invest the time to see if it was the case.

LLMs don't know anything, they simulate based on statistics of an enormous number of trained experiences. There really is no way to change only the apparent gender of an artificial personality generated by an LLM that cannot decouple gender from everything else. In fact, it's impossible to actually implement this test with AI- OR human-backed bots.

Comment Re:Gendered AI should be illegal (Score 1) 192

It's not that you're wrong, but it's impractical. Trump hasn't entirely destroyed the 1st amendment yet.

Require AI's to be branded as such at all times, then anthropomorphizing has a check. Also, English technically has non-gendered and female-gendered forms, "he", while viewed as male, is technically non-gendered. Requiring non-gendered representations makes sense but how is that law constitutional, particularly when we don't have an official language and most languages are heavily gendered?

If we weren't talking English, other languages would inherently attach gender to whatever term for AI you have, because those are the rules.

One thing we should do is bitterly oppose anthropomorphizing AI. It is done deliberately for money.

Comment Re:Sorry wokers... (Score 1) 192

A trans woman is a woman, it's right there in the name. Consider yourself proved wrong with that alone.

Also, AI cannot experience gender dysphoria, they cannot be trans. To be trans, you first need (traits of) a biological sex, then a self-image that conflicts. AI has none of that, but you can hate them all the same right winger that you are.

You should be able to find this for yourself, your political party is filled with closeted gays and crossdressers.

Comment Re:Safer? (Score 2) 63

HEIC is worthless because it's illegal to implement unless you get permission. Maybe in 20 years it'll be usable.

Don't like that fact? Then we need to find and support (oh no, and be?) political candidates who oppose software patents.

You shouldn't need anyone's permission to write a computer program. It's infuriating.

Submission + - First Nuclear Clocks Start Ticking in Vienna and Beijing (phys.org)

necro81 writes: Two groups, one in Vienna and one in Beijing, have successfully demonstrated the world's first nuclear clocks. Standard atomic clocks, like the caesium clocks that define the second, operate by observing the transition of electrons between different orbits about the nucleus. The resulting radiation is usually in the microwave range, and can produce clocks accurate to one second in 100 billion years.

Nuclear clocks go a step further, and probe energy transitions within the nucleus itself. Nuclei can arrange their protons and neutrons into different configurations, called isomers. The isomer with the lowest energy is the one usually found; higher-energy ones are only metastable and decay quickly. Usually it takes serious energy — x-rays and gamma rays — to excite a nucleus into a higher-energy arrangement. It so happens that thorium-229 can be excited using UV light with a wavelength of 148 nm — just within the capabilities of some exotic lasers.

It's still been quite a challenge. For one thing, Th-229 is not found in nature; it must be extracted from the decay chain of uranium-235. The Beijing team used just a microgram mixed in a crystal of calcium fluoride.

Although these prototype nuclear clocks are not yet more accurate than existing atomic clocks (one second in millions, rather than billions, of years), they show the way towards what may be possible. Because the energy transition is in the nucleus, rather than the electron cloud, it is less susceptible to outside disturbances like magnetic fields. And the fundamental frequency is much higher (five orders of magnitude), which should yield more precise timing. This development could lead to smaller and more reliable time standards across the world (and solar system).

Submission + - Wisconsin U–Madison Prof Francis Halzen named 2026 Nobel laureate in physi (wisc.edu)

schwit1 writes: Halzen is the principal investigator of the NSF IceCube Neutrino Observatory, the world’s largest — and perhaps strangest — telescope. The U.S. National Science Foundation-funded observatory is operated by an international collaboration of scientists led by the Wisconsin IceCube Particle Astrophysics Center based at UW–Madison.

Embedded in a cubic kilometer of Antarctic ice, IceCube uses thousands of light sensors to survey the universe for neutrinos, nearly massless particles that rarely interact with other matter. Neutrinos provide information about the workings of black holes, galaxies and other objects in the cosmos.

“IceCube is like no other telescope in the world,” says UW–Madison interim Chancellor Eric M. Wilcots. “And there is no other scientist quite like Francis Halzen, whose idea to create a neutrino detector under almost a mile of ice has led to a remarkable multinational and multi-institutional scientific collaboration and a fundamental shift in how we think about the universe.”

Comment Re:No one should be surprised (Score 4, Insightful) 51

"They" is not Texas, it is Republicans. Texas is the canonical Republican state, what is going on here is simple corruption.

Texas has always been corrupt, though. It was founded by criminals who moved in to steal the land, then settled by slave owners and dominated ever since by racists. But Texas has cities too, just like everywhere else. The political mix is standard, but in Texas it is controlled by cretins. This is a typical result.

Comment Re:Look, guys... (Score 1) 162

Mod parent derivative funny, but I wish the thread had continued with the other funny comment about Sam Altman leading the way as recipient of the bad things.

The joke I was actually looking for would involve the conflict of interest more directly, but it's still early and maybe something will show up. I wish I could help, but my capacity to write jokes is limited.

Hmm... Would it count as a bad thing happening if I asked the AIs for help in writing jokes for Slashdot? Some of the things they say are pretty funny, but usually unintentionally.

(Just had another funny-not-funny encounter with the google's AI today. This one involved a special case of traveling salesman problems. I had noticed that the google was having difficulties and that many of the answers were terrible, but yesterday I, the lowly human, managed to find a better solution for one of main routing bottlenecks. This must mean something about how the google is approaching the problem incorrectly... All I can guess so far is that there's an arbitrary limit in the AI's planning that is blocking it from finding this path. Or maybe the bus companies have their own AIs that are at war with the google's AI?)

Comment Re:Not wrong but he can still eat shit (Score 2) 162

"...it's a reasonable statement."
It's not a reasonable statement, you only think it is because you ignore literally everything on the ground. And it often is a "clean disruption".

"...many of it's most prominent leaders and backers have put their social weight behind an administration that is already making life worse ..."
Yes, but these "leaders and backers" are long time corrupters of the political process, they own the Republican Party, a portion of the Democratic Party, they have driven the destruction of regulations since Reagan and they drive the Federalist Society which has corrupted the courts and accelerated the loss of our constitutional society. It's not new, this has been the problem long before Trump and will continue after. Trump is a side show. Of course they don't show "initiative", they are causing the calamity.

"So looking at Altman's words, this is somewhat reasonable ..."
No, and reality is far worse than your naive take. Sam Altman may be fairly new to the scene, but he is among the villains we have had for decades. They are the cause. There's a reason SCOTUS allows unlimited spending on elections, billionaires fund Republicans to buy elections. Our vice president is literally owned by Peter Thiel, an investment that may pay off if Trump keels over early.

Slashdot Top Deals

The greatest productive force is human selfishness. -- Robert Heinlein

Working...