omg I loved this scene so much! not only because we finally saw how the machine worked, her very first number, but also because of how proud Harold was
I always found it a bit contrived that an AI would be given a gender. That and some aspects of Root/Shaw might have been some early identity politics creeping into an otherwise fantastic show
I’ve said this elsewhere, but does anybody else think that when Harold Finch says “I’m good with computers,” he’s making the understatement of the century?!
he says that because he has seen how the machine works. how can he consider himself any more then "good" after that? its like thinking your good at a game and then seeing a speedrunner play it.
It is something you have to remember a lot as a lot of future episodes reference back to pre machine official live date and other tit bits One thing is towards the last 2 seasons you have to understand the scale finch thinks forward and could of ended the series at any time and predicted what might happen and had his backup plans was set in motion very early on around first 2 seasons to end in the last 3 episodes (as back story to counter what "might happen" in the future (witch did happen) One thing what do you think would happen if Finch got angry and had a super powerful machine AI that was also not to happy at the same time (due to an event that happened on a episode) I don't know if you watched it yet so don't want to be specific about spoilers,
POI is one of those "once in a generation" shows, where nothing else comes close to it. To quote SCP-294, Addendum [SCP-294q-01], "at this point everything's just one big letdown."
Watch Buffy the Vampire Slayer. It's the same show, only gothic horror instead of cyberpunk. They even use a library as their HQ for the first several seasons.
they open a door for that, in last 2 episodes, it shows how thornhill industries recluted a Team in Washington D.C. composed of people saved in NYC, and in the last episode it is so ibvious thay Shaw, the detective Lionel Fusco, and the Asiátic Guy who knows of business and scams, make a new NYC Team who works for the machine now. in order to save the irrelevant's list people
@ Mich seal : Exactly! The machine connect every single clue to understand human behavior and predict our next move. It understand what Nathan are going to do after. And like Harold said, it has an instinct for self preservation ;)
Like ARNOLD IN DARK FATE. HE WAS ALMOST LIKE A MAN WITH AI MIXED IN HIS BRAIN CIRCUITRY. HOW HE PROGRAMMED HIMSELF TO CARE. HUMANS SHOULD NATURALLY CARE. SOME DONT. BUT HE STILL WAS NOT HUMAN. 6 MILLION DOLLAR MAN WAS JUST 📺.
The thing I never got with this was that some events would have taken days, weeks, months or even years to track and categorize, yet Harold erased/rebooted The Machine at midnight every day. So how could it keep track? Apart from that, this was an awesome series with great chemistry between the characters.
There is also an episode where they explain how The Machine works around this reset. I believe it is the episode where they introduce Thornhill Industries.
Only reason they were so amazing is our knowledge of the modern Machine.. if the show was just the start it wouldn't be as magical, small flashbacks are just enough to keep it interesting imo
This show is so good. And so few of us know about the show. Its been lost to time.. no one questions if there is a machine. An AI that tracks relevant and irrelevant threats and the irrelevant ones are just ignored... but out there nonetheless
When I started watch the TV show. I thought what a great fantastic. But later 1 year, I don't think like that, because I every day notice the moments in POI, in our real life. It is coming to trust
And yet he did tunnel a backdoor into the Machine. It may not have known at this point whether the backdoor was a good thing or not, so it is classified as a threat.
“I’ll encrypt the OS so completely, that no computer on earth will ever crack it.” Because that has always worked out so well when it comes to dangerous AI. *cough Skynet, Faro Swarm cough*
@@gildedphoenix the only instance of machine actually getting hacked was via a program that Finch written himself to fortify it and it required access to highest level government facilities i think it was good enough
In a later conversation (between Finch and Reese), regarding the OS. When Samaritan used the sample code to search (Search and Destroy). Finch stated he used a "Novel" Language Code to build its structure. So "...Encrypt the OS so completely...." is an apt statement. If someone who speaks a language that is so unique that only a very few know how to communicate with. That would also consider "hard to crack" term.
Don't forget finch feared the machine turning into skynet (didnt say it but rather implied the dangers of it) so he hobbled it and erased its memories every 24 hrs
Any AI with ndimensional utility functions (aka, anything with a goal of any kind) is working to minimize or maximize one or several values. Unfortunately, the problem of local minima/maxima means when the score of those utility functions (which you can think of kind of like dopamine and seratonin, to motivate new action or reward/punish successful/failure actions) hits a ceiling, it stops functioning and the AI becomes stuck. Typically in AI optimising problems, an AI doesn't need to reach the local minima to perform its function successfully but in enormously complex number crunching like this its essentially mandatory because its so many sub-tasks and if those sub-tasks aren't performed well, this introduces failure percentages to its value assessments -- and failure percentages multiply the more sub-tasks are added to a given chain. The Machine's task of identifying causal effect in human patterning using heuristic modelling and probability systems via neural comparitors is such an n-dimensional utility function problem. Basically, its like making an incredibly complex recipe and there being a percentage that you have to substitute the ingredients. When you have billions or trillions of ingredients, the number of non-substitutions needed to make a successful [food item] with an acceptable degree of accuracy are physically impossible. "But don't humans do that?" -- Yes: we use multiple parallel models which compete with one another and we begin selecting and trimming them accordingly and we too, get stuck on local minima in exactly the same way. In short -- this is science fiction -- until the local minima problem can be solved without outside assistance and the loss of model - since every time you reset the minima, the AI has to learn over from scratch which is incredibly wasteful. One way hypothetically to do this would be quantum computing -- since all AI in this case, is trial/error modelling with neurons which learn what works and what doesn't. In this case, a quantum computer can do every potential trial/error case simultaneously and reward for only the good ones -- and every time you double the number of qbits, the potential complexity of the simulation doubles. After that, the new question then becomes resolving power - the accuracy of your starting point when you model a simulation of something. For The Machine to work, it would need to overcome local minima, query duration and resolving power. Mathematically, two of these are unsolvable problems. The true answer to the P=NP problem will tell us which two can be solved.
Hey I got a question for the LOST fans out there. I know that Finch is Ben Linus... But is the other guy Goodwin? He looks a lot like him. Just wondering. :) Thnx.
Let's look at this quote: "When I'm done, I'll encrypt the OS so completely that no computer on Earth will ever crack it." From a first glance, this seems like surprising bullshit from a generally tech-aware show. One, you can't encrypt code and throw away the key, not if you ever plan on running it. The CPU needs to know what to do, which means it needs to decrypt it, which means it has to have the key, which means theoretically someone could just pull the key out of memory or even better, wait for the CPU to decrypt it for them. This is the problem with DRM, and why it always gets cracked while normal cryptography doesn't - because when you try to use encryption for DRM, you have to give the user a decryption key as well. But from a second glance... Harold is not above lying. "Encrypt" could be a metaphor. The obvious way to defend the system, is to have it defend itself. He doesn't want to admit that it has such agency, so is resorting to innocent-sounding terms like "encryption" - but in reality, what will stop you if you try to break it is not some clever DRM system, but the Machine herself.
In a later conversation (between Finch and Reese), regarding the OS. When Samaritan used the sample code to search (Search and Destroy). Finch stated he used a "Novel" Language Code to build its structure. So "...Encrypt the OS so completely...." is an apt statement. If someone who speaks a language that is so unique that only a very few know how to communicate with. That would also consider "hard to crack" term.
There's the possibility of homomorphic encryption; currently, that kind of encryption is still pretty far from practical use, but if Finch could create AI it's not a stretch to think that he could have solved that along the way.
@@htgnef it's actually far better then that remember in 4x02 when Harold tried to track Samaritans email Samaritan detected it and smashed straight thru every VPN and encryption Harold had in 20 seconds flat that's because both Samaritan and machine are sentient computer programs infinitly faster and better at computer code then any hacker and able to see an any attempt at hacking pretty much instantly the machine is quite literally impossible to hack
If you stopped reuploading your videos on the main person of interest channel you would see this show has racked up 10s of millions of views over its time on its official channel alone. If you include the copyright hit material that was forced off youtube then you'd double the views. You're making a mistake letting this show go.
If finch entered the machine in the science fair this would be the judges reaction. Harold: I built this system of mass servalance I call it the Machine. Judges: HOLY Sh#t that’s impressive.
Nathan was the face, noone knew Harry was the real creator. The NSA and US government would do anything to get the Machine, even murder. The machine therefor decided Nathan had a threat against him. Sadly, because his number was "non relevant" the Machine deleted his number and didnt warn Harry.
Just imagine theres a big (not big, bug. Big bug? Big bug.) in the machine that accuses people of crimes they didn’t or weren’t going to commit and there’s no way to fix it because he encrypted it
The machine doesn't accuse people of anything, its only output is a number for a person that's going to be in trouble. A bug would mean it could output a false positive with a person who's not in trouble.
The story Finch trying to to tell Nathen about the attack in 1994 was actually a true attack, almost everything has been said in this tv show had happened and going to happen, one more story I am waiting for when the Russia ambassador been attack by some terrorists when Greer is trying to convince Shaw that Samaritan is a hero
Don't worry, we're nowhere close to this type of artificial intelligence. For an AI to be capable of both simultaneously monitoring billions of people at a time and then be able to *predict* their every move would require processing power far beyond our grasp. So, unless there is a sudden and HUGE leap in quantum computing, we're safe.
![[LIVE] : ONE ลุมพินี 85 | คู่เอก "ยอดเหล็กเพชร vs ผึ้งหลวง"](http://i.ytimg.com/vi/Qq_19bnnxEo/mqdefault.jpg)
omg I loved this scene so much! not only because we finally saw how the machine worked, her very first number, but also because of how proud Harold was
You voted for Hillary Clinton, did you Soy Boy!
I always found it a bit contrived that an AI would be given a gender. That and some aspects of Root/Shaw might have been some early identity politics creeping into an otherwise fantastic show
@@lukebaines7785 Grow up
@@bpbeep6526 Just enough to not have a cat avatar.
@@zomfies ❄️
I’ve said this elsewhere, but does anybody else think that when Harold Finch says “I’m good with computers,” he’s making the understatement of the century?!
He's just the IT guy
Ikr?!
its like saying the sun is "just a bit warm"
Century?
@@xa1x1ax lmfao
he says that because he has seen how the machine works. how can he consider himself any more then "good" after that? its like thinking your good at a game and then seeing a speedrunner play it.
"So you think the Goverment would abuse this machine?"
Not even I would ask such a stupid question.
i never heard of this show before, i just saw this short clip and i'm already invested in watching this show O_O
It is something you have to remember a lot as a lot of future episodes reference back to pre machine official live date and other tit bits
One thing is towards the last 2 seasons you have to understand the scale finch thinks forward and could of ended the series at any time and predicted what might happen and had his backup plans was set in motion very early on around first 2 seasons to end in the last 3 episodes (as back story to counter what "might happen" in the future (witch did happen)
One thing what do you think would happen if Finch got angry and had a super powerful machine AI that was also not to happy at the same time (due to an event that happened on a episode)
I don't know if you watched it yet so don't want to be specific about spoilers,
Go. Watch. It. Now!
I hope you ended up watching it!
Oh this show is so good, and no one knows about it.
That's the point. Like Finch said "Nobody will know"
@@WryleHargreeves Only difference is that in this case it isn't 'the way it has to be'...
finch was happy with Nathan
and I was happy with you lola. pls come back!
🙈🙈🙈
@@devletrgt loooool
Remarkable. This serie was stunning.
No homo XD
Tbh i had a depression when they ended it. The finale was so good that i wanted more so badly. I wish they make a sequence.
I cried for hours
The machine will live on for as long as there is power and information
POI is one of those "once in a generation" shows, where nothing else comes close to it. To quote SCP-294, Addendum [SCP-294q-01], "at this point everything's just one big letdown."
Watch Buffy the Vampire Slayer. It's the same show, only gothic horror instead of cyberpunk. They even use a library as their HQ for the first several seasons.
they open a door for that, in last 2 episodes, it shows how thornhill industries recluted a Team in Washington D.C. composed of people saved in NYC, and in the last episode it is so ibvious thay Shaw, the detective Lionel Fusco, and the Asiátic Guy who knows of business and scams, make a new NYC Team who works for the machine now. in order to save the irrelevant's list people
@ Mich seal : Exactly! The machine connect every single clue to understand human behavior and predict our next move. It understand what Nathan are going to do after. And like Harold said, it has an instinct for self preservation ;)
Confirmed in 4x11.
***** That one was my favourite episode of season 4. It was amazing. How the Machine talked to Finch, how it learned...everything!
hebince44 Yeah! Me too. My favorites ever: 2x16, then 4x11, then 4x21, then 3x21.
Like ARNOLD IN DARK FATE. HE WAS ALMOST LIKE A MAN WITH AI MIXED IN HIS BRAIN CIRCUITRY. HOW HE PROGRAMMED HIMSELF TO CARE. HUMANS SHOULD NATURALLY CARE. SOME DONT. BUT HE STILL WAS NOT HUMAN. 6 MILLION DOLLAR MAN WAS JUST 📺.
The thing I never got with this was that some events would have taken days, weeks, months or even years to track and categorize, yet Harold erased/rebooted The Machine at midnight every day. So how could it keep track? Apart from that, this was an awesome series with great chemistry between the characters.
It doesn't delete the relevant list, and it can still see footage from the past, it only deletes own logs more or less.
There is also an episode where they explain how The Machine works around this reset. I believe it is the episode where they introduce Thornhill Industries.
yeah, that was just a bad episode -- better to just forget it ever existed
@@firstname4337 what makes you say that?
@@firstname4337 how lmao
I like the song that starts playing at the end.
Bloke Masterson It’s called “Listening with a million ears”
They should make a show just focused on these parts. Absolutely amazing.
I loved the flashbacks.
Only reason they were so amazing is our knowledge of the modern Machine.. if the show was just the start it wouldn't be as magical, small flashbacks are just enough to keep it interesting imo
@@BaconNuke I suppose that's why we think these parts are amazing: the pacing and the space between each flashback gave them more value.
This show is so good. And so few of us know about the show. Its been lost to time.. no one questions if there is a machine. An AI that tracks relevant and irrelevant threats and the irrelevant ones are just ignored... but out there nonetheless
"Antiques., like us :) Living underground, resisting the new age that tries to make us irrelevant. And like us, they still work" :)
It's there. Not sure it's fully functional...yet.
There's no one competent enough to pull such a heist
The end
Yes, Brett Cullen (actor) aka Nathan (other guy) is Goodwin from lost.
When I started watch the TV show. I thought what a great fantastic. But later 1 year, I don't think like that, because I every day notice the moments in POI, in our real life. It is coming to trust
ssssssshhhhhhh...they can hear you....Siri, no Alexa, Oh sorry no Cortana, actually I meant Watson Play "The Police - Every step you take"
Possible Threat Detected: Ingram, Nathan? Dude fought tooth and nail for this thing!
And yet he did tunnel a backdoor into the Machine. It may not have known at this point whether the backdoor was a good thing or not, so it is classified as a threat.
“I’ll encrypt the OS so completely, that no computer on earth will ever crack it.”
Because that has always worked out so well when it comes to dangerous AI. *cough Skynet, Faro Swarm cough*
to be fair it worked for Finch. Maybe you just need to not have writers hate you.
The machine wasn't so fortified though. There were moments it got hacked infected etc.
Still it was a great attempt.
@@gildedphoenix the only instance of machine actually getting hacked was via a program that Finch written himself to fortify it and it required access to highest level government facilities i think it was good enough
In a later conversation (between Finch and Reese), regarding the OS. When Samaritan used the sample code to search (Search and Destroy). Finch stated he used a "Novel" Language Code to build its structure. So "...Encrypt the OS so completely...." is an apt statement. If someone who speaks a language that is so unique that only a very few know how to communicate with. That would also consider "hard to crack" term.
Don't forget finch feared the machine turning into skynet (didnt say it but rather implied the dangers of it) so he hobbled it and erased its memories every 24 hrs
I bet Snowden likes this.
This is awesome. Science Fact not Science Fiction.
Any AI with ndimensional utility functions (aka, anything with a goal of any kind) is working to minimize or maximize one or several values. Unfortunately, the problem of local minima/maxima means when the score of those utility functions (which you can think of kind of like dopamine and seratonin, to motivate new action or reward/punish successful/failure actions) hits a ceiling, it stops functioning and the AI becomes stuck.
Typically in AI optimising problems, an AI doesn't need to reach the local minima to perform its function successfully but in enormously complex number crunching like this its essentially mandatory because its so many sub-tasks and if those sub-tasks aren't performed well, this introduces failure percentages to its value assessments -- and failure percentages multiply the more sub-tasks are added to a given chain.
The Machine's task of identifying causal effect in human patterning using heuristic modelling and probability systems via neural comparitors is such an n-dimensional utility function problem.
Basically, its like making an incredibly complex recipe and there being a percentage that you have to substitute the ingredients. When you have billions or trillions of ingredients, the number of non-substitutions needed to make a successful [food item] with an acceptable degree of accuracy are physically impossible.
"But don't humans do that?" -- Yes: we use multiple parallel models which compete with one another and we begin selecting and trimming them accordingly and we too, get stuck on local minima in exactly the same way.
In short -- this is science fiction -- until the local minima problem can be solved without outside assistance and the loss of model - since every time you reset the minima, the AI has to learn over from scratch which is incredibly wasteful. One way hypothetically to do this would be quantum computing -- since all AI in this case, is trial/error modelling with neurons which learn what works and what doesn't. In this case, a quantum computer can do every potential trial/error case simultaneously and reward for only the good ones -- and every time you double the number of qbits, the potential complexity of the simulation doubles. After that, the new question then becomes resolving power - the accuracy of your starting point when you model a simulation of something.
For The Machine to work, it would need to overcome local minima, query duration and resolving power. Mathematically, two of these are unsolvable problems.
The true answer to the P=NP problem will tell us which two can be solved.
@@osakanone I don't know what you said but I liked it.
BRING THIS SHOW BACK!
"I will encrypt it so completely"
Hey I got a question for the LOST fans out there. I know that Finch is Ben Linus... But is the other guy Goodwin? He looks a lot like him. Just wondering. :) Thnx.
acspotlight yup they both ran of the island
Technically sshd is always listening. 🤷♂️
I understand that reference lol
lmao
I would like for this show to come back :(
Let's look at this quote: "When I'm done, I'll encrypt the OS so completely that no computer on Earth will ever crack it."
From a first glance, this seems like surprising bullshit from a generally tech-aware show. One, you can't encrypt code and throw away the key, not if you ever plan on running it. The CPU needs to know what to do, which means it needs to decrypt it, which means it has to have the key, which means theoretically someone could just pull the key out of memory or even better, wait for the CPU to decrypt it for them. This is the problem with DRM, and why it always gets cracked while normal cryptography doesn't - because when you try to use encryption for DRM, you have to give the user a decryption key as well.
But from a second glance... Harold is not above lying. "Encrypt" could be a metaphor. The obvious way to defend the system, is to have it defend itself. He doesn't want to admit that it has such agency, so is resorting to innocent-sounding terms like "encryption" - but in reality, what will stop you if you try to break it is not some clever DRM system, but the Machine herself.
In a later conversation (between Finch and Reese), regarding the OS. When Samaritan used the sample code to search (Search and Destroy). Finch stated he used a "Novel" Language Code to build its structure. So "...Encrypt the OS so completely...." is an apt statement. If someone who speaks a language that is so unique that only a very few know how to communicate with. That would also consider "hard to crack" term.
There's the possibility of homomorphic encryption; currently, that kind of encryption is still pretty far from practical use, but if Finch could create AI it's not a stretch to think that he could have solved that along the way.
@@htgnef it's actually far better then that remember in 4x02 when Harold tried to track Samaritans email Samaritan detected it and smashed straight thru every VPN and encryption Harold had in 20 seconds flat that's because both Samaritan and machine are sentient computer programs infinitly faster and better at computer code then any hacker and able to see an any attempt at hacking pretty much instantly the machine is quite literally impossible to hack
If you stopped reuploading your videos on the main person of interest channel you would see this show has racked up 10s of millions of views over its time on its official channel alone.
If you include the copyright hit material that was forced off youtube then you'd double the views.
You're making a mistake letting this show go.
😔
Nathan knows full well how smart the machine is by now
If finch entered the machine in the science fair this would be the judges reaction.
Harold: I built this system of mass servalance I call it the Machine.
Judges: HOLY Sh#t that’s impressive.
Still lost to volcano
hi #CBS, we see you. We're always watching :)
me to. I hope they sell it on iTunes at some point.
(kind of like what they did for the Mentalist)
Loved both this and the mentalist - insanely underrated, at least by my generation.
I thought Ssh meant Secure Shell lol
I always feel like it's a prequel to the pc game Watchdog. Even the visual style matches!!!
Yes
I wonder why the machine saw Nathan as a threat.... maybe because it knew he was suspicious of it... idk...
Mich seal it never saw him as a threat
No it detected a threat to Ingram (by the government)
Nathan was the face, noone knew Harry was the real creator. The NSA and US government would do anything to get the Machine, even murder. The machine therefor decided Nathan had a threat against him. Sadly, because his number was "non relevant" the Machine deleted his number and didnt warn Harry.
Just imagine theres a big (not big, bug. Big bug? Big bug.) in the machine that accuses people of crimes they didn’t or weren’t going to commit and there’s no way to fix it because he encrypted it
The machine doesn't accuse people of anything, its only output is a number for a person that's going to be in trouble. A bug would mean it could output a false positive with a person who's not in trouble.
But the machine never accuse anyone. That's why Shaw and her team investigated the person of interest first just like finch and his team
The story Finch trying to to tell Nathen about the attack in 1994 was actually a true attack, almost everything has been said in this tv show had happened and going to happen, one more story I am waiting for when the Russia ambassador been attack by some terrorists when Greer is trying to convince Shaw that Samaritan is a hero
So, the machine can't take a joke?
Maybe it ranks statements that's technical higher rather than jokes or sarcasm?
That could be dangerous: Could the machine differenciate between jokes, sarcasm and actual treaths?
Actual facts script and novels can be classified as fiction To get truth out to the public. This ain't fiction.
PS, Our Phone Knows
I still can’t believe y’all cancelled this show.
Because of the fact that it's theories turned out to be true.
Don't worry. It's just Samaritan hiding its existence by cancelling the show and not letting it recommend to anyone
sometimes i think this show was canceled bc it was exposing things the government is actually doing.
Next level conspiracy theory on the clock ay
But the goverment is doing this. We just dont know if they have a IA
Don't worry, we're nowhere close to this type of artificial intelligence. For an AI to be capable of both simultaneously monitoring billions of people at a time and then be able to *predict* their every move would require processing power far beyond our grasp. So, unless there is a sudden and HUGE leap in quantum computing, we're safe.
Tired of ads.
Use adblock + image-block for pc.
Use lightning browser for phone
im LOUDER so listen all you want
FINCH!
He will always Be Ben Linus to me just because of his acting :)
Max Headroom!
Harold Finch understands, Nathan Ingram, not so much..
Can quantum computer, (google created) crack Finch's encryption?
Yes As a high level coder quantum and entanglement of photons even at ultraweak intensity can crack(crude),A machine with an unique coding language
That's not how it works lmao
I read that as ssh
Episode?
S01E11
✌🏼👁
first