I really don't think we should rely on AI for personal decisions. It's better to discuss and work on personal relationships and decisions rather than let statistics make our decisions.

@catdogmousecheese

24 күн бұрын

Yeah, tell that to people who believe in psychics and pay fortune tellers to see their future.

@ThrillSeeker3524

24 күн бұрын

@catdogmousecheese To be fair, that isn't pattern analysis

@spacepopeXIV

24 күн бұрын

I don't even think we should use AI to rely on public decisions that can and will affect countless people and with the ways things are looking now, especially minorities. AI has a tendency to forget about or directly target minorities and who knows if we'll ever fix that issue. I think it's best to just use AI for research and support in other fields of study that don't affect the social fabric, if that's even possible.

@CrashSable

24 күн бұрын

There's a difference between relying on an outside source to make decisions for you and using that outside source as a tool to inform your decisions

@maxave7448

24 күн бұрын

Statistics are also problematic because they can easily be manipulated and misunderstood. I remember a quote from some comedian that went like „About 1 in 1000000 people get mauled by a bear. So that means there could be a bear right next to me and its probably not gonna maul me“.

I wouldn’t mind a machine that would tell me the winning numbers of the lottery or which horse to bet on.

@freddyjosereginomontalvo4667

24 күн бұрын

Thats a good one. Also, I would like a machine that can predict what kind of illness I will probably suffer in the future so I can be prepared. Not only with me but with my family and pets too.

@ghostderazgriz

24 күн бұрын

Correction: you'd like a machine only you have access too to tell you the winning numbers.

@notfunny3397

24 күн бұрын

Just bet on the "women" in women's sports

@lfcbpro

24 күн бұрын

@@freddyjosereginomontalvo4667 ah, but it may be what you do to be prepared that causes you to get ill.......

@tycurt5016

23 күн бұрын

Ai's prediction is based on past situation,if everyone use it to predict, the accuracy will decline

“Life is not a problem to be solved, but a reality to be experienced.” Søren Kierkegaard (Danish theologian)

@evildragon1774

24 күн бұрын

Wait, i thought thats a quote from the first film of dune. The part where paul hallucinating about jamis

@punelopepunstop5515

24 күн бұрын

@@evildragon1774 I can’t answer that because I’ve never seen it…

@me0101001000

24 күн бұрын

@@evildragon1774 it is. But you should know that Frank Herbert was a voracious reader of nihilist and nihilist-adjacent work, such as Nietzsche and Kierkegaard. Considering the plot of Dune, I'm sure that doesn't surprise you.

@topherthe11th23

24 күн бұрын

@punelopepunstop5515 - What does that even mean? (If anything.) What is the difference (described in material terms) between a life that is a problem to be solved, and a life that is a reality to be experienced? Please describe a hypothetical situation in which a person, at the beginning, doesn't know whether life is (or is not) a problem to be solved, and also doesn't know whether life is (or is not) a reality to be experienced, and also doesn't know whether those two conjectures are mutually exclusive (i.e. the person doesn't know if life can be BOTH a problem to be solved AND a reality to be experienced, or can be neither). What, in this hypothetical, is the experiment that a person can conduct that proves that the conjecture "Life is a problem to be solved" is false, and that the conjecture "life is a reality to be experienced" is true, and that the truth-values of these conjectures must be opposite. In the absence of even a HYPOTHETICAL story about how a HYPOTHETICAL person who starts out without knowledge goes about acquiring knowledge and thus feels that Kierkegaard is right, I'm sure you've already figured out that I will think Kierkegaard's assertion is not true, and that it isn't false, but, rather, that it's a nonsensical assemblage of words that doesn't express anything concretely enough for it to be either true or false. It lacks even the HYPOTHETICAL ability to be confirmed or refuted by making observations. But if you have that story about how a scientist might go about testing Kierkegaard's assertion, please post it and I will read it.

@alberteinstein8862

24 күн бұрын

​@@topherthe11th23come back from your real account, Nietzsche

If the AI says we fail, we stay together out of sheer spite.

@jakekn7304

24 күн бұрын

Not sure of the ai cares either way

@voltmatrix1250

24 күн бұрын

@@jakekn7304 I’m gonna pretend it does for the sake of my hypothetical marriage

@M1LKMAN

17 күн бұрын

🤖: 🤬

@logixindie

6 күн бұрын

Yes it is almost a good idea to tell people they gonna fail so they put effort to defy the prediction.

You're changing the outcome by observing it, and planting doubt

@alberteinstein8862

24 күн бұрын

The outcome was never changed, doubting was akways a part of it

@KevinHorecka

22 күн бұрын

Doubt + perceived evidence in agreement or disagreement with the doubt is different than the base doubt. If the people involved truly have no belief in the evidence, then they may be the same, but then why did they observe it and why is it being considered as part of the decision?

If someone breaks up with you over an AI prediction, congratulations. You just dodged a bullet.

@gelosyguat1211

21 күн бұрын

Exactly😅

@tyronebonquaviusdaquan3562

10 күн бұрын

what if your relationship made because of AI?

Interpretability research is important. Its one thing to predict what will happen given initial conditions, its another to say what should be changed about the initial conditions to ensure a given outcome...

@TheAncientColossus

24 күн бұрын

It's a boundary value problem in differential equations. There is no need for all those words. It's that simple. 🤡🤮

@zyansheep

24 күн бұрын

@@TheAncientColossus not sure what you mean by "boundary value problem in differential equations" but given my lack of understanding, I theorize that other people may also not understand this reformulation of what I said, and thus I suspect that lengthy descriptions of complex topics in simple language may be helpful for some readers.

@TheAncientColossus

24 күн бұрын

@zyansheep Complexity is the bane of modernity. Luckily, everyone has access to Google. There is no need to overcomplicate everything in life. Life is simple. Mental illness is rampant.

@sarveshpadav2881

18 күн бұрын

Your comment reminded me of bayes theorem.

@TheAncientColossus

18 күн бұрын

@@sarveshpadav2881 Thank you.

Holy Matrimonial Minority Report, Batman!

@jain_swati

24 күн бұрын

😂😂

@221b_Bakerstreet

Күн бұрын

Absolutely😂

AI could easily without training and proper context, confuse correlation for causation. Just because a data point is common between events, doesn’t mean it necessarily causes it.

@ghostderazgriz

24 күн бұрын

That's not confusion. That's just math. Correlation is the only factor that is important for predicting events. Determining the cause is a far more involved process that requires investigation from many different angles and many tests to uncover the truth. Even in this video, they account for this by saying "The AI doesn't tell you why, only that you will"

@EldestZelot

24 күн бұрын

​@@ghostderazgrizCorrelations aren't the only thing. Relationships, effect size and power are all just as important in understanding the relationship between two dimensions.

@pageturner2958

24 күн бұрын

That is actually a bit of a problem with AIs that can happen. Algorithmic Bias and all. Lets say a company is making an AI that hires people based on their resume. If the company historically hired men more often than women, the AI might take that correlation and decide men are more hireable than women. Correlation (men are hired more than women) that the AI is assuming means causation (men are more hireable than women).

Marriage rates are at an all time low, but I think people still make such important decisions without thinking them through enough. Be it marriage, engagement, children, buying a house with your partner. Think about how much time most people spend researching what car to buy and collection data, comparing, but the decisions above are made on gut feeling alone. Crazy in my opinion.

The specific problem you propose is interesting in that the AI divorce prediction tool is trained on data sets that pre-date the AI divorce prediction tool so, the prediction is probably only valid if the couple marries and does not use the prediction to influence their marriage. One big factor that affects the likelihood of a divorce is communications (or lack of it). If the tool predicted a divorce and that spurred the couple to communicate more about what they wanted in the future, that could protect their marriage. One the other hand if the tool predicted they would not divorce, maybe the couple would not talk about their future since it seemed set already and that lack of communication could cause a divorce. When using a tool, it is important to understand what the assumptions and parameters it is built on.

@eruben2

3 күн бұрын

Great points

Remember that when you hear “AI isn’t transparent” it doesn’t mean that AIs try to hide information from you; it means that even if we read the AI’s code, we can’t always make out what certain parts of the code do. In not being able to figure those parts out we say that said parts “aren’t transparent” to us.

@adisage

22 күн бұрын

That is indeed an important distinction, thank you for pointing that out! But can we not build a way for the AI to explain its decision making? Like the newer ChatGPT models can explain the solution of a math problem by breaking it down ?

@wigosas

21 күн бұрын

​@@adisage It depends on the AI technique, neural networks have this problem but other methods can showing the reasoning

@adisage

21 күн бұрын

@@wigosas I'd call it a feature rather than a problem 😅 but I was just wondering if there could be a workaround of some kind

@wigosas

21 күн бұрын

@@adisage with neural networks there isn't, we would need to understand them on a deeper level

Let’s get metaphysical. A couple looks takes the test and it comes out yes. The once healthy couple is afraid the other will trigger a divorce and they grow unhappy and distant. That causes them to get divorced. If they hadn’t looked at the results, they would continue in their marital bliss and remain married. Are the results in a Schrodinger’s cat quantum state until they check the results?

@CDexie

24 күн бұрын

The video addresses this, as a self fulfilling prophecy. It could artificially maintain the statistics that drove it to that conclusion in the first place

@veritasdeutsch6608

24 күн бұрын

But they could have divorced for unrelated reasons anyway

@naimasophie

24 күн бұрын

I feel like healthy couple wouldn't be afraid in the first place.... it's usually when we have doubts that something can influence us

@bert-qu3iq

24 күн бұрын

It's somewhat like insisting on a prenuptial agreement before marriage. If you're 100% positive of the strength of your relationship then there's no need for a prenup yet, in order to be safe, by insisting on one, you're acknowledging that there could exist potential cracks that would leave you vulnerable in the future. In essence you are foreshadowing your relationship's demise by giving credence to your less than 100% sense of security. Not exactly reassuring to a potential spouse.

We appreciate your insights. Keep working hard.

@afdalowhida899

24 күн бұрын

مينست١ص٦٧بعثثضج١ؤ. تذتة. ؤق ؤ ضى😅عضعذ LOQ

This problem encompasses plenty if not all fields of study, showing just how important each one is and is going to be as we march further into the future. It encompasses computer science, AI science, physics, engineering, ethics, philosophy, sociology, psychology, and statistics among other fields of math.

Tbf if you distill it, AI is statistics in action. You're not asking a thing that thinks through reason, so why would you expect a reason? It's not much different than looking at statistical data to find correlations The AI isn't judging you. It's just showing patterns of correlation in statistical data

Forty two years ago our friends thought we’d get divorced in five years. They were wrong. We’re still happily together. 😊

We often do "irrational" things such as deliberately go against predictions just to prove them wrong. Humans are not lifeless, thoughtless natural processes, the very act of telling us that something is impossible is enough motivation for us to do it anyways. Our daring and willingness to follow our dreams and go against the grain is what makes humans unique from any AI.

It sounds like a useful tool if it comes with reasons. If it gave reasons then you can look into it. If it is a raw number nothing else, then it not providing anything useful.

In the last month, I’ve seen AI mistake the Suez Canal for the Panama Canal, mistake the diameter of Olympius mons by a factor of 6, forget that Francis Drake was in the military, and it gave me a completely different amswer on which snake had the most dangerous venom depending on what i asked it previously (identical phrasing). AI is cool, it can find you information quickly, but anything it says needs to be checked first, and we can’t do any checking on predictions if we don‘t know what is behind a given prediction.

@CDexie

24 күн бұрын

Keep in mind that what we colloquially call AI isn't just chatbots and ChatGPT. Large Language Models are still new, and it's truly wondrous they work as well as they do. That doesn't mean that AI can or will be able to make your decisions for you, but in areas like medicine, are you willing to sacrifice explainability for accuracy? Who cares why, if it's going to save the patient? But in judicial matters, the why is very, VERY important. That's the issue the video tackles

@wildersilly653

24 күн бұрын

@@CDexie If AIs make mistakes, it's not going to possible to fully understand what went wrong. It's fine to use medical AIs to get an initial diagnosis, but a real, trained doctor is going to need to look that over before action is taken.

I would only use it to predict things I would do anyways, regardless of the result. I've taught myself to be comfortable with taking risks, no matter how unfavourable they may be, as long as I believe it's the right thing to do. I would find using the machine to know exactly how much of a risk I'm taking to be empowering, without letting it dictate my choices

@naimasophie

24 күн бұрын

We would all take risks if we believe it's the right thing to do. Truly taking risks means doing something while you have absolutely no idea if it's right or not

@char1194

24 күн бұрын

@@naimasophie _believe,_ not _know._ I _believe_ something is right but I don't _know_ for sure. Your comment doesnt make sense because its possible to believe something is the right thing to do without having any idea if what you believe is actually right or not, so in those cases, by your definition, you both are and arent taking risks? And in any case I think both your statements are false even if they didnt contradict

The type of person who would trust an AI's prediction of a relationship is exactly the type of person who is most likely to fail at a relationship.

There is a fourth option, using that information to commit on understanding strengths and weakness in the relationship to make it stronger, go to a couple's therapy and say "hey, we're great right now, and we want to keep that way, how can we do that?" Prevention is the best in most cases, getting married or not is not the issue, it's the actions after that that can change the outcome.

Thank you for sharing your expertise and creativity through this outstanding video.

excited to discover new learning opportunities

My answer is no, but the potential of AI is crazy and should not just be ignored and used to measly detect marriage success, it should be used for weather forecast, earthquake, volcano, storm detection, etc with its accuracy it might even just save lives instead of making people heartbroken type shi.

You can't predict the future because of quantum mechanics; A good example is of nuclei of a radio active particle which has a 99.999% chances of breaking but still it could break the next instant or after a millennia, we could never can correctly guess the outcome.

@MsKoffeinjunky

24 күн бұрын

And if you average it out on a large number of particles you prediction gets really really good. You don't understand statistics and quantum mechanics.

@The_Observer_god

24 күн бұрын

@@MsKoffeinjunky Yes however if you are counting on a large scale you could can almost predict what would happen, but individually it's harder than you think

Has anyone remember the original name is How a Machine tells a future?

Can you please let us know what tool or platform you use for this animated video? Because your animations are very cool and i want to learn this type of animations

Ted -ed sugestion to next history video: Los Angeles ritos of 1992

I’m sure someone mentioned this already, but wouldn’t using the machine and learning the outcome, change said outcome? Ie. If someone saw a 5% success rate but loved their partner enough to marry them regardless, but with the knowledge of the 5% - depending on their disposition, that knowledge could ensure they treasure every moment of the relationship, because they know the very high probability of it ending - but this act of cherishing keeps them together, thus the success rate, in actuality , increased to something like 90%, simply from the initial knowledge of the 5%. The 5% prediction could only be true if the Ai doesn’t give them the knowledge - which defeats the purpose.

There is a saying in Hindi that goes something like, there is no cure to suspicion. Once planted in your mind, it manifests in the ugliest of ways. It doesn’t matter whether the AI is right or wrong, if someone could even remotely suggest the possibility of something happening, it is human nature to ruminate over it. The principle of manifestation is very much real - no matter good or bad, the more you think about something, the more you are inclined to take decisions that would lead to that very outcome. We create our own problems and cry over them. Even if not for AI, marriages these days are so fragile that break over the stupidest of things (not talking about abusive marriages here). AI will perhaps inflate the problem even more. Because who is to say that a generation that relies on AI to do even the most menial tasks will not become so mentally handicapped that it would also rely on AI to make such big decisions?

@namikazeswapnil671

24 күн бұрын

Omkara

A machine which can predict my future gives me the power to change it well, a nice bet

@dylanvellut

24 күн бұрын

I'd advise you to watch Jojo bizarre adevnture part 5

@Timbik123

24 күн бұрын

@@dylanvellut am accepting your recommendation

your videos help me a lots in my journey learning english thank you very much

In order to test the accuracy of the prediction, you need a placebo. For instance, a random prediction. That is, without using data and just randomly guessing divorce or not, this result is assigned. Then, the difference between the predicted results and the random results would be the strength of using prediction with data. Without the placebo, it’s impossible to know how knowing the result of your marriage changes the prediction.

I don't think it is out sourcing your decision and deferring your agency to AI if you choose to follow the prediction. It is simply getting more information about external factors that will affect you. Learning there is a 95% chance of rain when deciding whether to go out is not outsourcing your decision and deferring your agency if you choose to stay inside. In this scenario about marriage, I would tell my partner about the AI's prediction and still stay together if they are willing because in my mind if we break up know then it is pretty much the same as divorcing except with the divorce we would have had those years of happiness. But, if my partner decided to end it, I would also understand.

To be honest, since most people are damaged and faulty by default, always making bad decisions about what to do with themselves and others around them, I think that just the simple act of checking or feeling the need to check it out already demonstrates that the couple isn't meant to be together, along with other unrelated subjects where the actor might have doubts or is unable to understand his role in society That's why there are so many divorces already, and that's why nothing's ever going to change

A divorse prediction machine would fall into the classic greek tragedy plot, it happens because you try to make it not happen and wouldnt if you didnt

I could. Depends on what I want it to predict, though.

AI still can't produce for me a decent classic country song in arabic language. Why should we trust predictions without understanding the reason?

I believe this where we should set the boundaries for AI

Yes so I would know if I should end it from now or not.

I think the simplest solution for this particular ai is rather than focusing on displaying the results of their divorce, it needs to focus on how each person need to develop or change certain behaviors that could lead to divorce. So the machine could predict how and where they need to improve. And leave it to them…

You know it really doesn't matter because under multiverse theory if you did see your future you could still change it by going on a different time line. Unless, of course, the future you saw was fixed point in time in which case it'd be impossible for you to change it because the universe itself will actively stop you.

00:30 one of my friends got married for only 4 months till they got a divorce , they spent more time together during the divorce process than they did married

that is for early machine learning/AI systems, like years ago. these days, AI products can be developed that can also explain its predictions/analysis.

If it's that accurate then wouldn't it make sense to check with the AI about all potential partners and limit your dating pool ahead of time rather than waiting until you're already considering marriage and realizing you had a high statistical probability of never reaching that point in your relationship in the first place? I get that how long you've been together is an indicator for how long the relationship is likely to continue but if the model is predicting you're going to break up in spite of that then there must be a lot of factors that can cause problems for you you've already been ignoring. Why not limit your options to the ones who will give you a stronger statistical foundation to build the relationship on? It's really not as if the AI will tell you "you only have two people on Earth that you have a chance with and they're both married already"

Would I use a machine that could predict the future? I dunno, machine, you tell me

Wouldn't it be better to live in smaller communities and have an entire small town who knows you both extremely well help you make a decision like this?

I don't even need AI to predict the fact that I only have a few minutes left until I get to the end of this delicious slice of pie I'm about to eat. But there's no question about it, I'm definitely going to eat it anyway, and with no regrets. Those 5 minutes are worth living the experience, even if it has an end.

I say it’s the same reason we don’t have machines judge whether someone committed a crime or not: there’d be no one to blame/punish. If an AI has a false prediction or is just wrong, you can’t really do anything about that

Interesting direction of the subject matter you choose to take this conversation in . I thought you were going to talk about the origin paradox and the accuracy paradox. ♑✍️🇧🇻🇦🇺

But… self fulfilling prophecy

I feel the real idea here is to try and I peel back what the AI is actually tracking through the study of its predictions. That in of itself will lead to Intresting data

Look at it this way. There's a 5% chance for the two of you to achieve happiness. That's better odds than most of us will ever get.

People change, and AI uses current info. So it may be wrong.

Would not knowing any divorce statistics affect the outcome where ignorance is bliss? I feel like knowing the high divorce rate in itself would affect the marriage.

Where did you beginning quote go? 😢

would the AI determine the impact of the information it provides on the end result? would it withhold information? does it have a preference over the result? what is the preference based off of? would it benefit from return customers? would that influence the output?

It is better to love and fail than to have not loved at all.

I think it would be a useful tool. No one else would really tell you not to marry someone. Think of it as advice, just really accurate advice. If it told you that you're not a good match, it's up to the couple to figure why not and are you going to try and fix it.

I remember there's a film similar about this, Don't ever let the AI do everything

I won’t decide important decisions by AI. Our future is ours, so we should not rely on other ideas.

I was just discussing this issue with chatGPT last night when I asked it how it arrived at a solution to a problem. The algorithm itself has some sort of reasoning process but since the AI isn't self aware of its own algorithm, it can't explain how it arrives at solutions. There's probably a way to get an AI to be self aware & capable of explaining its own process. This would probably give it the appearance of consciousness as well.

It remind me of the scene in the movie Gattaca when Uma Thurman's character brings a hair sample to a clinic to have her potential partner tested for compatibility. The AI could be programmed to tell you how and why you will break up but it will be long and boring as it's just a statistical and pattern analysis.

Do you think an ai could theoretically create a predictive model that could take a portion of reality and extrapolate the reaction of every molecule that came into contact to create them into that state all the way back to the big bang could we create a way to somewhat "see" a model of the past. Like seeing into the future into the past.

statisically knowing something in the future will affect the future

The last phrase is the key!

Statistical Models are an important tool which can be used to make clear decisions. Of course it should not be the only tool used. Statistical models are not made up numbers but numbers based on actual outcomes from past behaviors.

Wasn't there a movie based on this same theme??? I just remember the couple decided to give their relationship a chance in the end. They were together till the end of their times. Anyway, AI doing this kinda predictions is just like those streetlight self-proclaimed diviners who divine your future out of the air.

You can't predict the future because of quantum mechanics; A good example is of nuclei of a radio active particle which has a 99.999% chances of breaking but still it could break the next instant or after a millennia, we could never can correctly guess the outcome

@Cristi4n_Ariel

24 күн бұрын

This is a bot that copied someone's else comment. Sad that KZread hasn't been able to ge rid of them.

@a-walpatches6460

23 күн бұрын

@@tafdiz Of course it does! How could a detailed view of the universe have no effect on a more general view The macroscopic world is 100% governed by quantum mechanics, it's what quantum mechanics looks like when you zoom out.

@shanggosteen9804

10 күн бұрын

Such is the rules of entropy and quantum mechanics

Could the knowledge of the prediction change the outcome? Do you know what I mean?

I think the AI is like the Oracle of Delphi in ancient Greece, always providing prophecy to those who seek answers and usually the prophecy was fulfilled because of paranoia and doubt.

No. No I wouldn't use the machine and if my partner is so wrapped up in the statistics then they're not living in reality. In the present moment. And honestly, it doesn't seem like they're invested in the relationship much at all if they're worried about the *statistics* and not talking to me in the moment.

I have done some analysis with shap values and help a lot with interpretability. So there are some research done in that direction. But even if that it is the case the ethical dilemma of taking important decision based on an AI that doesnt feel what you do, it is indeed problematic, in a sense what we as a human should do? Where are we entitle to do as we pleased even if it is not optimal? And other point is, even if you do break up, was it worth it? That question is I think more important and nobody will know it. Not even an AI. And lastly, nobody fit an statistical metric perfectly, we are near that measure but not exactly. So the ai will not know the history and learning lesson you truly have in your marrige. (Sorry for my english is not my first language)

There's a well known concept of "Self fulfilling prophecy" in sociology. There is a well known experiment where a teacher was told a class of students were determined to be gifted, and so the teacher gave extra attention to them and encouraged them to work hard and so this class got much higher scores than other classes. But in reality they weren't determined to be more gifted, it was an ordinary class like any other in the school. If we are told that something is true about a person or a group's behaviour then that alone can modify how we approach them so that the original claim turns out to be true. So if you sow a seed of doubt that a couple will get divorced, whether through AI or some fortune telling, then they will start distrusting each other and likely destroy their own relationship and get divorced. Such predictions about how people will behave is fraught with danger as humans aren't really predictable and are also unfortunately very impressionable by such bold statements and predictions.

@eruben2

3 күн бұрын

Boy, this couple better not talk to their friends or family then. One little indication - “hey, are you sure Alex is right for you?” - will then cause doubt, possibly leading to destruction.

@litetaker

3 күн бұрын

@@eruben2 you may be saying it sarcastically, but it's not impossible. Of course one single statement may not lead to it, but repeatedly saying it can. And it depends on how it's stated. If it's an opinion and you don't trust that person to give good opinions, then you may not care. If it is coming from someone you truly trust and it's stated very confidently, authoritatively, then it will lead to doubts in your mind as well. In any case, these may not lead to the self-fulfilling prophecy. But people who believe astrology to their very core put so much belief into it that it can very much lead to a self-fulfilling prophecy. This is why astrology plays a big role in marriage in many cultures.

@eruben2

3 күн бұрын

@@litetaker I do believe the concept of a self-fulfilling prophecy is possible; certainly, someone who believes they can do something will try harder and be more likely to do such a thing. I just don't see this as a big reason not to have AI make predictions like this discussion is covering. My point is that being against AI making a prediction is like being against anyone talking to their friends about whether a partner is good for them. Sure, it's possible that either one becomes a self-fulfilling prophecy. But realistically you always have to think for yourself if you want a good outcome. Astrology can be dangerous, just because I'm a Scorpio then I should be cautious marrying another Scorpio? Utterly ridiculous.

@litetaker

3 күн бұрын

@@eruben2 but the example given in this is not simply an AI giving advice, but the AI will apparently do some magic modelling to give you a definitive answer of whether someone is suitable or not. That to me sounds like a perfect scenario for self-fulfilling prophecies to occur. Because, if you defy the AI and still go ahead, you may be constantly wondering if the relationship won't work out. And that lack of trust in the relationship could potentially cause it to break down. And if you believe the AI, then you already fulfilled it's claim. It can lead to situations where you no longer give the relationship a fair shot. Of course this same issue will occur if you believe in astrology as if it's 100% true. In many traditional marriages, they will just not go ahead with the marriage due to it, and if they do go ahead anyways, any issues that occur in the marriage will likely be attributed to the poor astrological compatibility. I am saying that an AI system will likely cause the same exact issues because how can one accurately predict what will happen with two human beings? I feel like that is a fundamentally flawed point of view and humans are simply not predictable like that and claiming to build such a system, whether astrology or AI is essentially a recipe for disaster.

@litetaker

3 күн бұрын

So I think what I'm saying is I don't believe AI or anything or anyone can predict the future and if some silly tech guy from silicon valley comes and says they built an AI that can do it, even with more advanced technologies than today's, it's best to run away from such things 😂

I honestly like surprises

Loving ship of diamond! 😊

I love the fact they didn't use AI to draw images ❤

Most definitely and I haven’t even watched the video yet.

As someone who is currently engaged to someone named Alex, this was a weird TedEd to watch 😂

Maybe that’s a wrong question to ask. Even if the AI prediction is correct, one still worth to get married. Nothing last forever, maybe that 5 years will give us something and shaped us to a better person, and most importantly we chose it at our own will, our life is not governed by a machine. We are all mortal, doesn’t mean that if we can’t own things forever then there is no more reason to pursue it.

Isn't the ia own accuracy also a self fulfilling prophecy? Because, who are the most likely to ask an ia for relationship advice? People who are already insecure in their relationship. So if the ia especially give negative answers, then the likelihood of the couple breaking up and the IA being proven right increase, because the IA answer was just an excuse the participant wanted to hear to convince them that there is indeed something wrong with their relationship, even if the IA will never explain what exactly.

So, the question arising is, what happens when one person is predicted to divorce and the other is predicted to stay together?

95% chance is high? The general public should be mandated to play XCOM, lol

I wonder why you've haven't done a video on cancel culture. The topic is EXTREMELY relevant today, and it seems like it can get wildly out of hand.

Current AIs don't reason therefore they don't predict either. Perhaps the video should have been clearer about that.

No, I would rather rely on planetary positions.

If it was truly 100% accurate, and I do end up getting a horrible prediction, then I guess that's it really, can't do anything about it, doesn't mean I'll stop caring about my life. We all gotta go someday, be it tomorrow, next month, next year, next decade, next century or next millenia, who knows somebody might even make it to the end of it all, but we will all die someday and you just have to get comfortable with that. What matters is how you lived your life, if you think you did good, even if in the end you die in a horrible way then, you still did good in life.

@dylanvellut

24 күн бұрын

Have you seen the movie "Big Fish" yet ?

@dud3655

24 күн бұрын

@@dylanvellut Oh I haven't seen it yet

Why does this paradox have to include that the machine won't explain its reasoning behind its decision? It's a weird limit to apply when AI is able to explain its decisions. If an AI were advanced enough to go through all that data and parse through it and make connections it would be able to note those connections made that led it to a failed/successful marriage prediction. And that bit of information changes this whole thought experiment and makes AI sound less dubious. Knowing the AIs reasoning you can make a better educated decision on how much you want to trust it's choice.

95% accuracy doesnt mean that it will happen in itself...

there is some research that Quantum computer could predict the future above 99 percent, and one left could be computer error but some doubts about its capacities. Therefore, it has currently still in testing and experiment.

Cold calculations isn't going to account for emotions

If it was me, i would not get married but would stay in the relationship.

For as long as mankind has existed, there have been attempts to foretell the future. From the Oracle of Delphi to horoscopes to crystal balls to tarot cards to weather reports to almanacs to advanced statistics and probability to name just a few. It’s one of the many ways we deal with our anxiety over uncertainty. None have worked. None have been accurate. It’s at best a calculated guess. But a guess is all it is and all it’s ever been.

Statistics do not inform the experience of the individual.

this is more of a didn't even ask topic tbh!

Watch Dogs has already predicted Bellwether. Doesn't seem all that too far-fetched to me.

How am I now worried about a fictional relationship from Ted-Ed?

There's a biological machine in my head that constantly predicts the future and is almost never right.