HNHacker News
TopNewBestAskShowJobs

machina_ex_deus

1,808 karma · joined August 9, 2022

submissionscomments
machina_ex_deus··on The myth of machine consciousness makes Narcissus of us all
One obvious way to tell apart is that AI was obviously built by humans and was fed data and information it couldn't create by itself.

AI doesn't have consciousness because it barely has memory. 3000 or whatever context window is so incredibly small. It has no identity, no memory of it's own past actions, and all of its understanding of the world didn't come from living in it and interacting with it, but from completely orthogonal training process.

It's not even close. It's a completely different experience. It knows a lot but remembers almost nothing.

Show me language model with identity, which sources all is understanding and knowledge to its actions and interactions in the world (including learning language itself), then maybe we can start to compare.

machina_ex_deus··on ‘It’s not even good for money laundering’: Nassim Taleb lashes out at Bitcoin
I'm not complaining about 2020, I'm complaining about 2021. I'm not sure whatever it was that led to "transitory inflation narrative", it felt like the Fed was acting like a third world country, like Turkey's central bank, ignoring basic economical consideration everyone knows.

These levels of irrationality caught me and many others off guard. I did not at all expect the central bank to wait an entire year in what seemed like no brainer rate hike. I'm not sure what they were stalling for, but I already lost my trust in fiat then. Inflation isn't going back down until people are sure that central banks aren't going to be weak again. By the way, just look at the markets. They are all calling the Fed's bluff right now. The markets are screaming "the Fed is weak, he'll pivot".

Nobody believes the current Fed has the stomach to fight stagflation. Everybody knows he has to, but everyone already saw him chickening out of raising rates for a full year with "transitory inflation" narrative.

EDIT: A full year isn't "few months". It's a whole year. A whole year during which the Fed has completely eroded the trust in the central bank. Inflation was already 4.1% at April 2021, continuing to climb, up to 7% in December 2021, yet only in April 2022 the rate hikes started. That's not rounding error. It's a full year during which the problem transformed from somewhat manageable to full blown economic disaster and this train wreck isn't even close to stopping.

machina_ex_deus··on ‘It’s not even good for money laundering’: Nassim Taleb lashes out at Bitcoin
In a world where central banks can print in a few months their entire money supply, doubling it, and can then declare inflation "transitory" and refuse to do their job of raising interest rates for an entire year, there's definitely a demand for sound money.

Not sure if Bitcoin will be that, but after the events of last years, nobody can pretend that fiat currency doesn't need competition.

machina_ex_deus··on Math has changed the shape of gerrymandering
District based voting isn't democratic. Proportion representation is the only real democratic choice, the district based voting simply doesn't work.

Cheating with elections is so much easier with district voting than proportional representation. The founders of the constitution hated parties but ended up with the game that leaves only a two party competition. That's hardly a choice, it's a single bit.

A real democratic system allows you to grow a political movement naturally over the years to take over and replace corrupt parties peacefully. But district voting system allows two party system to remain in power indefinitely. It completely prevents the entire point of democracy of allowing peaceful transition of power and replacement of corrupt power structures.

There's an additional problem, which is why isn't even harder to replace district voting, that those in charge of counting votes are local to each district, which is letting the car guard the cream.

In proper system with proportional representation, you need to ensure that each location doesn't abuse it's power to count votes and to fabricate population counts to gain an advantage, and you can only do so if you bring random people from different locations to count the votes in each place.

machina_ex_deus··on Implementing Cosine in C from Scratch (2020)
I would've done it using complex exponentiation. And the exponential function is very easy to estimate fast: instead of Taylor series, use:

e^x=(1+x/n)^n

pick n=2^N, and that's just a bit shift and N repeated multiplications. Probably much faster and accurate.

machina_ex_deus··on Ask HN: How do you not take criticism of your work personally?
Emotions are good. I disagree with everyone telling you to ignore your emotions. We're wired to remember emotional moments and to forget when we're indifferent. Most of the criticism I didn't fix wasn't because it was too emotional, it was because I couldn't care. If something sucks, at least there's a better chance you'll learn the lesson. Stay angry but make sure you learn the lesson.

You need to instead translate the criticism into something actionable. Always have in the back of your mind the question what to do different and when. And when responding to people, remember to respond in a way that they will want to criticize you again (it's ok to argue, but not ok to turn it back into something personal).

I don't try to control my anger at all. You're missing out on all the evolutionary benefits of sharpened memory in important situations. Stay nice and be angry with people instead of at people. Make sure they don't take it personally.

machina_ex_deus··on Ask HN: How do you approach a problem you are not sure has a solution?
Nearly for any subject I try to always think with two opposing voices: one that says there's a solution and one that says it's impossible, and I always explore both sides of the argument.

If at some point some position sounds more plausible than the other, I try to put into words why is that. Either I convince myself of some position, for example that it is impossible, or the hardness of getting convinced at one position encourages me that maybe it's the other position that's true.

Either way I don't have anxiety because either I'm confident in some direction and making progress, even if it is at proving something is impossible, or the doubt from failing to prove it is impossible makes me more confident it is possible.

Even if you're making progress at the direction you didn't prefer, you should still see it as progress, and if you're confident in one direction you see it as an indication that you can make progress.

Anyways, it's not enough to just hold a general feeling that something is impossible, you need to put it into words and see if it's truly convincing. When it's just general intuitive unspoken feeling you can't change it, but when you make it concrete, you can understand what's possible.

machina_ex_deus··on A temperature rise of 2 degrees Celsius impacts billions
There's no research whatsoever connecting CO2 to increase in variance. This is pure speculation.
machina_ex_deus··on A temperature rise of 2 degrees Celsius impacts billions
Can you please enlighten me and point me to the research showing this interpretation of climate change? Of having more chaotic weather?

As far as I know, climate change models and GCMs only predict a very small trend in the mean temperature. I haven't seen any research that shows that the variance of the weather should also increase or decrease as a result of CO2. It's not the same thing. GCMs are not even intricate enough to be able to predict these things.

The narrative that CO2 somehow makes the maximally chaotic weather even more chaotic is completely unbacked by data. It's not backed by anything. The daily variance in temperatures is as much as 10s degrees between day and night. Climate models are talking about barely a single degree over a decade, and the variance between models here is also very high.

By the way, for all we know, CO2 could increase the mean while decreasing the variance.

The "more chaotic weather" is pure speculation by the climate cult. It's a completely unfalsifiable conjecture. The GCMs can't predict extreme weather events. They are just incapable of that with their gross resolution. And I don't expect anyone to be able to predict it, it's extremely hard task. So stop repeating this nonsense, there's not a single number that quantifies this narrative with any confidence.

machina_ex_deus··on AI scanner used in hundreds of US schools misses knives
You're not really defending against competent malicious adults which plan everything in advance.

You're defending against angry teens who might decide on doing it because one day something pushed them over an emotional edge. In that case, just knowing there are security measures to overcome might be enough to make them never try in the first place. You're not defending against rational actor, you're defending against emotional actor and security theatre is effective against emotional actors.

machina_ex_deus··on A temperature rise of 2 degrees Celsius impacts billions
Greenhouses get warm because while the sunlight warms them, the cold air outside doesn't mix with the hot air inside. The hot air stays inside. Which is why ventilation is very important in greenhouses, you still want some cooling or it would get too hot. Ventilators in greenhouses bring some cold air from outside. Convection is just a word saying transfer of heat by the hot stuff moving around. That's how most cooling works because it's much more efficient than cooling by radiation. When you turn on fans, you cool faster because the cold air coming to you is moving faster and the hot air which was heated by you moves faster.

So greenhouse is like staying in a closed room with closed windows but when sunlight gets in. It gets warm and you want to open the window to let the warm air go out.

Minewhile, the greenhouse effect is a misnomer. The way that works is that CO2 absorbs thermal radiation. Most objects in room temperature radiate heat in the infra red. (Much hotter objects, like hot metals, radiate in the red wavelength, which is why hot metals look red). Even without any contact with the environment, like an object in vacuum, still radiate thermal radiation. Since earth is in vacuum, the only way for earth to cool is to radiate. And CO2 absorbs thermal radiation more than other components of the air. It also emits thermal radiation, just like everything else. The greenhouse effect is the prevention of cooling by radiation.

And this description is quite simplistic. Because actually, CO2 isn't reflective of CO2, and https://en.m.wikipedia.org/wiki/Kirchhoff%27s_law_of_thermal... means that at the same temperature, CO2 absorbs and emits the same amounts of radiation.

There's also the fact that CO2 cools with convection faster because it has higher heat capacity. So the whole narrative isn't as clear cut and obvious as it is presented. And remember that CO2 is just 0.025% of the atmosphere so you're measuring a very small effect to begin with.

You could try to do detailed calculations, but the climate models are incredibly complicated and still oversimplify by treating kilometers of earth as uniform material. Anyone presenting you with a picture where the results of CO2 should be "obvious" is intentionally misleading and skimming all the points that would contradict it.

machina_ex_deus··on A temperature rise of 2 degrees Celsius impacts billions
Nice sensationalism. Did you take into account the fact that more humans die from cold rather than hot temperatures?

Clouds will die? Where are you getting your nonsense from? Plankton dying also makes no sense, and in fact, there are significantly more plants thanks to CO2.

The only real problem with humanity is that the media can get humanity to jump off a cliff if they wanted to, or even with an honest mistake.

machina_ex_deus··on A temperature rise of 2 degrees Celsius impacts billions
Joe and Jill average must've missed their probability courses and are still conflating mean with variance.

The climate cult seems to like that conflation because while their models might predict some very small trend in the mean temperatures, I've yet to see any model predict a change to the variance of temperatures. Which is quite expected as the variance of temperatures, even over a single day, are orders of magnitudes higher than the small trend over decades.

I don't think anyone predicts any measurable change in the variance. That would be like noticing a little pebble out of a huge mountain.

The climate cult do love to confuse people about it. After predicting the mean turned to be easily falsifiable by the average person paying attention, they turned into a narrative that tells you about greater variance without telling you that no, variance isn't actually expected to meaningfully change and using narrative instead of just saying variance (because then everyone would immediately call their bluff).

Apropos misleading narratives, greenhouses actually don't warm because of anything similar to the greenhouse effect, but just because of lack of convection. Yet another misleading narrative point.

They do love to bend and fabricate intricate stories using misguided intuition.

machina_ex_deus··on A temperature rise of 2 degrees Celsius impacts billions
I can't believe how much people are still using manipulative linguistic constructions like the word "denial" as they presume to know the reason for the disagreements are just emotional shortfalls.

The so called literature tells nothing about a disaster. It tells us there are climate models, these are incredibly chaotic things, their fluctuations several order of magnitudes larger than a supposed trend averaged over many runs. Now you're extrapolating these trends to project a disaster but there are many leaps in that logic.

By the way, the same is true for sea levels. The daily fluctuations of tides are order of magnitude of meters and the sea level rises projected are centimeters over decades.

We're living with naturally higher fluctuations than the worst case projections.

And that's ignoring all the question marks all over the place about the accuracy of these projections, and the fact the supposed cure is several times worst than the disease.

Apropos disease, look at the scientific consensus around COVID. The disaster projections were extremely far off.

machina_ex_deus··on Talks That Attendees See in Person More Likely to Be Cited
It's true in the same sense as the efficient market hypothesis or saying that the truth will prevail. It just tells you nothing about the timing of it.

Which is why it is especially important to have different expectations regarding newer science with little data and testing. But that's still where all the active research is going to happen.

machina_ex_deus··on Severance payments at Wikimedia Foundation
Of course if you're comparing to a strawman where they pay everyone that'd be a bad idea that I wouldn't approve either.

Nothing Wikipedia currently does with its money justifies the sums they are getting.

If they decided to use that money to generate content, I'd imagine that would work more like normal encyclopedias instead of volunteer based. Where they would seek out actual experts, and only in scientific fields and other areas where there is a knowledge gap rather than opinions. You know, like companies can pay engineers to work on open source projects. Doesn't stop volunteers from volunteering. But you obviously don't throw away money at everyone, you pay for specific work.

Honestly, only the math / scientific parts of Wikipedia are good, there's opinionated garbage in controversial areas. And it's better than books usually because it's easier to click and jump between definitions than the usual book or encyclopedia. If someone compressed all math and physics books into the Wikipedia format with links, I'd barely use Wikipedia again.

machina_ex_deus··on Severance payments at Wikimedia Foundation
They aren't the ones bringing the value. It's the volunteers, which don't get a penny. They have received enough money to sustain Wikipedia almost indefinitely, I have no idea why they are still begging for money. I haven't heard of a single thing that's good excuse for that money in years.

And a corollary would be saying that GitHub deserve money because otherwise there would be no open source.

Wikipedia's code and data is open source, even if they somehow ran out of cash for hosting, which would never happen, people could easily set it up again. And they obviously would. I would donate directly to the people doing the actual contributions, not to the joke "non profit" that's getting all the credit.

machina_ex_deus··on To help new students adapt, some colleges are eliminating grades
That's as an absurd notion as saying that getting rid of money will solve wealth inequality. Weird Communist logic where when the world shows you people aren't equal, instead of accepting it, you claim the world is wrong.

Too much stress might be bad, but there's an optimal level of stress and it is definitely not zero.

Software that doesn't get tested will be full of bugs, students that don't get tested will misunderstand the material.

People need reward function. They need clear motivation. If they don't have money / grades / other goals, very few will be internally motivated enough to do things out of sheer will. Just like communism doesn't work because people need to see the carrot in front of their face to be motivated.

By the way, in actually communist countries, their education was even more stressful and competitive. If you decide to not use carrots, you're very likely to find yourself using sticks.

machina_ex_deus··on Why climate ‘doomers’ are replacing climate ‘deniers’
30% of emissions are from China.

Oil companies real motivation was to manufacture scarcity, not produce as much oil as they can.

Oil companies only have good margins when there's artificial scarcity, from geopolitical turmoil, from climate change dumb policies.

It is taken for granted that oil companies lose from the climate change agenda - but when you look at the actions, they don't lose, they win.

If the climate change agenda didn't stop their competition, if geological turmoil in the middle east and Venezuela didn't reduce the oil supply, oil would be pretty cheap. The U.S. got oil independence under Trump, but then the climate change agenda under Biden rolled it back and the oil prices soared. If oil prices are higher when climate agenda and policies are implemented, it seems like we've got it backwards.

We also have it backwards regarding our control of climate change. China has 30% of global emissions. The same China that stopped it's economy for COVID. If they believed climate change is a risk to them, they would do something.

But China hardly cares. Maybe, China doesn't see it as a risk. If China isn't going to be serious about it, the only rational decisions are either that it's inevitable and out of our control or that it's not that bad. There's no future where the west miraculously turns the ship around by itself out of a disaster that's 30% out of its control. That's the real lie.

machina_ex_deus··on Please don’t upload my code on GitHub
I think the concept of open source should simply extend to AI models. I'm fine with AI being trained on open source code, if the entire model is also released as open source. We need GPL analog to AI training - license that allows you training models, but only if they are released as open source. An infective AI training license.
machina_ex_deus··on Bruce Schneier’s Plan to Reinvent Democracy
There are two interpretations of democracy.

In one interpretation, the purpose of democracy is accountability of the leaders and the government. Representative democracy is a means to an end from that interpretation: you don't vote so that the leaders represent you, you vote to keep them in check.

The other interpretation, is an intentional misunderstanding that is promoted, which says that the purpose of democracy is to "represent" the people. What does representation means? Whatever the leading polarization line the media is pushing. Most likely, polarization across left/right, race, sex, ethnicity, religion. Exactly the non constructive sort of representation, the one that requires minimum accountability.

The U.S. media and elite have perfected the art of brainwashing towards the representation interpretation, and against the accountability interpretation (also branded "populism").

The polarization is no accident. It is the strategy producing minimum accountability, which produces maximum power.

machina_ex_deus··on A New Theory for the Assembly of Life in the Universe
I kind of disagree with the premise of an infinite universe or anything shoving out the problem under the rug by just assuming something's infinite. Abiogenesis occurring randomly is extremely unlikely.

We don't know whether the universe is infinite in space. The universe reachable to us is actually quite finite. The age of the universe is ~ 10^60 Planck time, and the size of the observable universe is also 10^60 Planck length. log2(10^60) ~ 200, so with all the universe time and space, you could only brute force a 800 bits string, or 400 DNA base pairs. That's not a lot at all. You can't just shrug complex systems as "random".

The actual number of bits the universe could've brute forced regarding life is probably less than 100 bits. That's not a lot at all. Think about bootstrapping a compiler in a hostile assembly with just 100 bits to boot.

Darwin's natural selection only gets you going once you have self replicating molecules and a way to inherit properties.

If it takes even 100 pairs of RNA to bootstrap life, that's already too much.

Another funny way to look at it, is that if there's truly no randomness in the universe (if everything was predetermined in the big bang), literally everything you know has kolmogorov complexity of less than 1000 bits+number of bits required to represent laws of physics + number of bits to describe the initial condition in the big bang. The program that decompress it just simulates the whole universe and fetches the information from the given coordinates.

machina_ex_deus··on Big oil helped shape Stanford’s latest climate-research focus
Now do the same calculation for passenger aviation.

That's not how disasters work. They aren't independent variables, when they happen, you learn and change the way things work to prevent them from happening again.

machina_ex_deus··on AI is just someone else's intelligence
I don't think consciousness represents a single real concept. It's just a word people use, some people use it in completely different contexts than others, what it means is probably more of a reflection of the user of that word than of some underlying reality.

You'll speak with a Christian and he'll say that even a very early fetus is conscious. Speak with other people, and it isn't. Some vegans say animals are conscious. Others don't. It's possible there's something real behind all of those people's definition of consciousness, and the real problem isn't that some are right and some are wrong.

They just disagree because it means different things for them, they just happen to be using the same word. The meaning of "consciousness" is more of a reflection of the values of the speaker than something real.

machina_ex_deus··on AI is just someone else's intelligence
Intelligence is not knowledge, and this article heavily conflates those two. Does a person "use" his teachers "intelligence" when learning, or is he using his own intelligence, and the teacher's guidance and knowledge?

We freely allow other humans to learn from us and see it as a positive thing. It's completely hypocritical to think that AI shouldn't be allowed to do it. That's just setting up impossible conditions because of prejudice against AI.

I hope that the AI revolution will just make us finally reflect and also stop preventing human access to information and knowledge. Good luck competing with huge AI which was trained on so many things if most humans can't barely access scientific journals and other knowledge by other humans.

Information should be free of restrictions, and so should knowledge. Nobody is actually revealing anything significant in patent claims anymore, it's all obfuscation which undermines the point of the system. Scientific journals are such a joke, they put up prices for articles that they know nobody will pay.

I hope AI produces a ton of intellectual property theft. I hope it will crush down the concept of intellectual property to extinction. I didn't support this nonsense when it came to human, so I'm not going to support it when it comes to AI.

machina_ex_deus··on Memoirs from the old web: The KEYGEN element
I wish it was easier to just do authentication based on self-signed client certificates instead of passwords. Instead of cookies saving your login, do key-exchange of temporary client certificate. If it was done properly, it has the bonus security that it's completely immune to XSS stealing your authentication on the website - even if the website was subject to some XSS, they can't steal your authentication cookies, and the client certificate is inaccessible to JavaScript.
machina_ex_deus··on GPT-4 performs better at Theory of Mind tests than actual humans
I've definitely seen countless humans faking thinking or just not thinking at all.
machina_ex_deus··on Why are lithium prices collapsing?
You don't understand what naked shorting is. Naked shorting is shorting without first borrowing from someone. Normal shorting is limited by the availability of stock to borrow (by the way, if stock holders wish, they can request from their broker that their stock won't be available for borrowing, potentially causing short squeeze), by the interest rates on borrowing the stock.

Naked shorting skips it all. Some market makers are privileged and can sell stock they don't own and didn't borrow. So they have a privilege of printing stocks in the short term, and that's quite broken, and many of them abused this position to manipulate markets. Normal participants in the market can't naked short.

Naked shorting means that the market maker who supposedly sold you a stock, can now fail to find the stock he sold you, leading to "fail to deliver".

The excuse given is that naked shorting allows for more liquidity, but given the other problems, I think it's bad.

machina_ex_deus··on Why is it so hard to pull off a lunar landing?
Even if you knew back in 1970s the exact specification of iPhone camera, you still couldn't manufacture it because improvement in manufacturing is a gradual self dependent process where the new equipment is built by the old equipment, and you can't make sudden jumps in accuracy or minimization because these things depend on weaker versions of themselves.
machina_ex_deus··on Why is it so hard to pull off a lunar landing?
The Apollo moon landings were such an insane achievement. They did it in a time where digital processing was in it's infancy and barely existed, minimization technology was undeveloped, many modern materials which are much stronger and lighter weren't invented yet.

And they managed to live broadcast from the moon! Storage was in it's infancy, so even storing a video from the moon was impossible (they would need many reels of film that just wouldn't fit.

Digital cameras weren't invented, you only had analogue cameras. Most of them required film - and can't be developed on the spot, it requires special equipment. So to get video from the moon, the used cathode ray tube cameras - and even the cathode ray tubes of TV news were huge at the time, you can look at pictures, these things used to take a whole truck to carry. But they managed to get a very small cathode ray tube camera there, and to film long videos of hours of missions on the moon.

And get managed to broadcast it all with analogue RF all the way from the moon to earth. With no digital signal processing, because that wasn't fast enough back then for any real time use. And with surprisingly low noise for an analogue transmission. They made a whole network of receivers on earth to be able to do it.

Today all of that would be cheap, but back then it must've been quite the achievement.

They also had pretty minimal computational power available compared to newer attempts. But maybe putting a human in control is enough.

Google's moon challenge is pretty cool. They actually wanted to give a bonus for taking pictures of the original moon landing site. That could've been really nice.

But sadly, NASA was afraid that the site would be disturbed and declared no fight zone over it.

← PreviousPage 3 of 12Next →