rhyn00 · HackerTrans

vor 3 Jahrendiscuss

If you took time to read it you would find that meat packing and hazardous chemicals are still illegal under the new bill.

rhyn00

vor 3 Jahrendiscuss

The bill explicitly states packing plants are not allowed.

rhyn00

vor 4 Jahrendiscuss

Is AI the problem here, or is this story a symptom of a bigger problem?

This is a somewhat long comment, but just thought I would share my thoughts.

The concern is that someone who has worked hard to create art wants use it to leverage it to obtain security in their life, and they are at risk of losing that leverage with new AI technology. Well, first of all, plenty of people work hard their whole lives for low compensation because their skills are not unique, and often don't have the time/capital/energy to build new skills even if they wanted to.

I question the way we live our life, there is a lot of suffering in the world because we are focused on taking from each other (in some cases using violence, in other cases laws). In the first world most people are expected to spend their childhood preparing to work away their adult life to earn minimum compensation enriching someone else, and in the third word most people are in extreme poverty, which the first world is happy to exploit for labor, resource extraction, brain drain, strategic military advantage, ect.

Artificial intelligence is about to make a lot of things that required difficult and unique training trivial. Making many more people redundant, and no longer unique/needed even if we do end up creating copyright laws enforcing that nobody can use someone's work to train the AI. Do we really think that we won't become redundant just because it doesn't train on our art/code? So like the artist here, we should all be wondering what is to come.

I think reason people are so worried is because we have it in our heads the way to conduct trade, and ourselves is by extracting value from one another. Instead of a culture of giving gifts without guarantee of reward, fostering relationships and caring for and understanding each other, we are trying to take from each other to ensure our needs/wants are met. The ultimate way to achieve security under this way of life is to create a dependence, become the owner of the assets/capital that can be used to take value/labor from others want/depend on what we have.

We act like we have only a few choices when it comes to how to live, religion, capitalism, communism, socialism, and so on. But we have the choice to try and understand the people we meet in life, and reach out to those who we don't know. Knowing each others needs/wants we can help each other out. We have the ability to be generous when we can, foster relationships, so that in our time of need we might be helped, by someone that cares (because we're there for them too). It's just that we listen to the advertisements telling us if we want to be happy, to buy more and more things (possibly at determent of our own health/well being), we listen to leaders that would have us go to war with each other, we listen to society telling us to focus on making more money, exchange labor for wealth, and use the wealth to obtain security/happiness.

What if instead we believed the most important thing was to live with balance with each other and with nature, and to communicate with each other to see how we can help. We do after all live in this world together, and when there is an imbalance we see the effects like poverty, stress/anxiety, addiction, theft, violence, wars, exploitation, hate, not to mention the obscene amount of time we spend working (40+hr/wk for 40+ years) to enrich someone that doesn't care about us.

Are we not living in a backwards way? Most people are dependent on a few, by culture, or by force (think of how this force is created - if you don't obey those with power/authority, someone else who does will threaten you with violence or revoke privileges). If someone breaks a law, or threatens our livelihood, we'll just take away their livelihood (or what little they had to begin with) - how well has this worked out?

With the advanced technology to communicate more easily than ever before, and provide what we need/want with much less labor/skill as before, and the wealth of knowledge available, it's time realize, that by giving to each other and helping one another, we will all be better off. Well, maybe not the extremely wealthy and powerful powerful people - but even they don't get to live in peace because someone is always trying to take their spot - think about all the wars fought because one authority is threatened by another, or all the companies trying to hold their grip in the face of more desperate innovate companies.

I'm not saying doing something different is easy, or straightforward, but it might be better. I think we don't need to ask anyone's permission to do it, we just start doing it within our own networks - building up each other, making new relationships, looking out for one another.

If you want your security to come from laws and by taking instead of giving, how long that will last, just how secure is that anyways?

rhyn00

vor 4 Jahrendiscuss

Wow, I'd never heard of StarBook, but it does indeed look like a wonderful machine.

rhyn00

vor 4 Jahrendiscuss

I would add that the linux-ready boutique vendors like system76, tuxedo computers, framework are also options if you don't want to fuss with drivers and what not, but still want to run linux. I definitely agree that linux doesn't run super smooth on all hardware, but it's not hard to find hardware these days where it does run smooth.

I came across tuxedo computers randomly one day, and gave it a shot. Very impressed, and am extremely happy with my tuxedo pulse 15 gen2 - running their supported version of Ubuntu+KDE, that just works out of the box. Only thing I can complain about is that: speakers are not great (but I use headphones 90% time anyways), and KDE doesn't support independent resolution scaling (I need 125% for laptop display but 100% for external monitor), so it's a bit hacky to get scaling the way I want. However, everything else runs perfectly and smoothly.

It's best laptop I've ever owned for linux. It is quite, portable, moderate power laptop, for fair price. I gave my wife my Macbook air M1 over this one. While the M1 CPU/GPU is a little more powerful than Ryzen 5700U (8 core), I get more ram (32gb 3200mhz), bigger and faster disk (1TB 980 pro pci 4), more battery life (18hr idle, 10+ working) for similar price. It's also repairable, w/ removable standard components (not cpu tho). Linux running SMOOTH.

Basically with these type of vendors, you don't need to struggle or sacrifice (much) to run linux anymore. Tuxedo computers [1] has many more models worth checking out, like with high end GPUs or smaller/more portable (even one that support external liquid cooling and an rtx 3080ti lol).

[1] Tuxedo Computer (notebooks) https://www.tuxedocomputers.com/en/Linux-Hardware/Linux-Note... [2] Pulse 15 gen2 : https://www.tuxedocomputers.com/en/Linux-Hardware/Notebooks/...

rhyn00

vor 4 Jahrendiscuss

Nice, that's pretty cool!

rhyn00

vor 4 Jahrendiscuss

Adding a few results:

Using OP's code for following

    php 1.8mb/sec
    python 3.8 Mb/sec
    node 1.0 Mb/sec

Java print 1.3 Mb/sec

    echo 'class Code {public static void main(String[] args) {while (true){System.out.print("1");}}}' >Code.java; javac Code.java ; java Code | pv>/dev/null

Java with buffering 57.4 Mb/sec

    echo 'import java.io.*;class Code2 {public static void main(String[] args) throws IOException {BufferedWriter log = new BufferedWriter(new OutputStreamWriter(System.out));while(true){log.write("1");}}}' > Code2.java ; javac Code2.java ; java Code2 | pv >/dev/null

rhyn00

vor 4 Jahrendiscuss

Spectral decomposition is pretty cool.

Application 1 - spectral clustering - an alternative to k-means for nonlinear clusters. Get a Distance matrix of your data, spectral decomp, run k-means on your k top eigen vectors and that's your clusters.

Application 2 - graph clustering - (run spectral clustering on adj matrix!)

There's some tricks to getting it to work in practice like normalizing but it's a simple and powerful method. Also the matrices can get big so it helps a lot to use sparse matrix libraries for the computations.

[1] https://towardsdatascience.com/spectral-clustering-aba2640c0....

[2] https://www.hindawi.com/journals/ddns/2020/4540302/

rhyn00

vor 4 Jahrendiscuss

Interesting, as they mention, it looks like the main advantage to an approach like this is that it can make use of "special events" (e.g. stop/starts, bumps, turning) which can null out the drift.

Combining this with other covariates/sensors could also provide more correction and better detection of "special events".

I wonder if something like this applied on top of traditional methods that incorporate the dynamics would make a better approach since you get the advantage of using the known dynamics + learnable "special event" corrections.

rhyn00

vor 5 Jahrendiscuss

I've been getting into his later work with Chiara Carletto: Constructor theory of information. It's an interesting way of looking at the world. Rather than using dynamical equations to describe the world, in constructor theory it is about looking at what things are possible and impossible with respect to information exchange. A constructor is something that can apply a transformation to something without itself changing (kinda like a catalyst in chemistry).

Here is the home page if you want to learn more:

[paper] https://www.constructortheory.org/wp-content/uploads/2016/03...

[homepage] https://www.constructortheory.org/

rhyn00

vor 5 Jahrendiscuss

Marc Lewis, a neuroscientist has a nice book on the topic, which to me makes complete sense (and is grounded in sound science). His view on the topic is that addiction is nothing special, just a learned behavior. Hebbian learning states that neurons that fire together wire together (corollary: neurons that don't, the connection weakens). The neurons in the limbic system control emotions, attention, desire, and fear. The prefrontal cortex (PFC) controls planning, expectation, and long term planning. We are motivated to do things that we desire, or avoid things we fear, which is controlled by the limbic system (also involving dopamine release).

Let's take drugs for an example but this can apply to other behaviors like internet usage, eating, falling in love, becoming obsessed with a hobby, or cronic anxiety toward something (avoiding fear). When we take a drug (or do xyz behavior ) it gives us a positive feeling (pleasure, or comfort in avoiding fear) and probably a favorable outcome occurred, so we are likely to engage in the behavior again. Each time we engage in the behavior the connection between the limbic system and the stimulus that triggers it, strengthen. When the trigger happens the limbic system lights up and triggers the PFC to create a plan and imagine how good it will be (then a positive feedback will occur - this is craving). Over time more and more stimuli will become triggers. Also, if you don't use it you lose it, and the neurons responsible for long term planning will weaken their connection to the limbic system (also, shown in brain scans), making it harder for you to assert self control.

However, your brain is plastic, and you can also unlearn the behavior over time (studies also support this). Meditation can help because you can stimulate these neural pathways in a controlled environment. For example you can imagine how a trigger leads to drugs, then you can imagine how drugs leads to bad things while meditating - over time you can strengthen the connection from the trigger to the emotion and imagery of bad things happening. Ofcourse meditation is not the only thing that can help, but the main idea is that you want to make your brain associate the bad behavior with bad outcome, and strengthen the association of good outcomes with fear/desire circuit.

Book: The Biology of Desire: Why Addiction Is Not a Disease - https://www.amazon.com/Biology-Desire-Why-Addiction-Disease/...

Youtube: The Neuroscience of addiction - https://www.youtube.com/watch?v=aOSD9rTVuWc

Edit: typos

rhyn00

vor 5 Jahrendiscuss

This article does a horrible job of explaining Turchin's ideas. It's not too many "brainy" people that are the problem. It's that if upward mobility becomes limited to those that have advanced degrees, and if there are an oversupply of people with advanced degrees, you will now have (at least) 3 groups of people: non-advanced-educated-but-not-good-enough, educated-but-not-good-enough, and elites.

The non-advanced-educated feel marginalized because it will be ridiculously hard for them to get the education, and as things get more expensive their livelihood will go down relative to the elites.

The advanced-educated-but-not-good-enough will feel marginalized because they spent their time and money and got nothing in return, and now have debt and degraded livelihood compared to the elites.

When enough people feel marginalized AND something happens that weakens the state's power/influence, political instability will occur.

It compounds in US/Europe with regards to education, given that these two groups are politically opposite of each other, and advanced education is becoming a barrier of entry to upward mobility.

Turchin's 2016 book Ages of discord explains these ideas in detail.

rhyn00

vor 5 Jahrendiscuss

Does anyone recommend a book about Solomonoff induction?

rhyn00

vor 5 Jahrendiscuss

This sort of reminds me of the book "Vehicles: Experiments in Synthetic Psychology" by Valentino Braitenberg. In this book the author starts a series of thought experiments by constructing small "vehicles" which drive around on a table top. The vehicles start with very simple behaviors, then he applies evolution (by vehicles falling off, or being selectively removed) while adding more complex behaviors until the vehicles eventually become intelligent.

In a way the slim mold is a analogous to one of the simple vehicles that ends up becoming more intelligent through the simple mechanisms and evolution.

Book link: https://mitpress.mit.edu/books/vehicles

rhyn00

vor 6 Jahrendiscuss

That would actually be a pretty difficult ask, you would have to cover the order of thousands of acers of land, spotty uplink service, supplying power or developing low power solution (also needs to be maintained by the farmers). Also, building a machine/implement to kill this or that plant would be pretty tough too. That's aside from the soil type and what kind of irrigation you're able to use.

It wasn't mentioned in the video, but you can get kinda close to what you're thinking today. When harvesting the crops it's possible create a map of the yield throughout the field. Then with combinations of random sampling of the soil throughout different parts of the field, one can adjust the fertilizer distribution thought the field. Finally, when you go to plant the crops next year, you can utilize this data to configure the planter to automatically space the plants a little closer or farther apart. Many farmers are already doing exactly this.

comments