Tap to unmute
How to Defend the Indefensible
Umieść na stronie
- Opublikowany 22 lut 2023
- The Wheel of Pain returns with AI Fighter Jets, AI Video Game Patents, and ID laws for pornography. Ft. Dan the Producer!
Watch the full WAN Show: • My CEO Quit - WAN...
► GET MERCH: lttstore.com
► LTX 2023 TICKETS AVAILABLE NOW: lmg.gg/ltx23
► GET EXCLUSIVE CONTENT ON FLOATPLANE: lmg.gg/lttfloatplane
► SPONSORS, AFFILIATES, AND PARTNERS: lmg.gg/partners
► OUR WAN PODCAST GEAR: lmg.gg/wanset
FOLLOW US ON SOCIAL
- Nauka i technika
Komentarze • 257
I would suggest 1 minute to defend, 30seconds to poke holes at the argument and 1 minute to address the holes, imo it would make for a better format
@not john doe yeah holy shit that was heated.
So much agree, there is not much point to this if all we are gonna get is slightly improved corpo explanations.
It's the hole poking and counter poking, plus the drama that makes this interesting.
ALSO I would add to your timeline, a final minute of unrestricted back and forth where talking over each other is permitted.
yeah i agree having people poke holes is important so these don't just seem like one sided shilling
Completely agree, without the argument it lacks the fun of the last wheel of pain.
Yess, on the last segment seeing both luke and linus fired, throwing arguments was really good
A “secret hitler” version of this segment would be so funny. Both Luke and Linus get assigned the topic they are defending at the start of the show in secret. Then at the end of the show they guess which topic the other person was disingenuously defending
while the idea sounds fun, even in its current format its ripe for being clipped out of context and being used for malicious purposes.
This should be a separate show but it's a great idea.
BUMP BUMP BUMP
Great idea but it’d have to be extremely clear to the audience what’s going on, otherwise you’d get complaints from them making hot takes.
Ohh, that sounds fun
I think the discussion/argument in the previous video was what made it so much fun to watch. I’m disappointed it was cut. I also think that it being cut results in an inaccurate representation of the issue at hand. The defense ends up fighting what is essentially a straw man.
"USAF allows AI to fly fighter jet." "Hopefully not Bing."
Luke woke up that morning and chose violence
The worst part about Dan's defence is that I can easily see many American politicians say something like this. Hell, they say worse things pretty regularly
China already did that with pc gaming didn't they
Dan did a lot more than advocate for the devil, he was pitching him new ideas!! 😂
"I can do what I want, I'm the producer" should be a t-shirt
dan's segment won in my books
Y'all should flip a coin to decide who goes first, then that person gets to pick any topic they want for the other to defend (and presumably will pick something hard). This alleviates the problem of doing uninteresting or easy topics.
Rip Luke, Biing will airstrike your house sooner or later and you deserve it. Its your own fault and you know it and you need to make it right and apologize!
"You spelled my name wrong... you're next!"
Omg this comment
Wheel of pain is great and so is Dan we need more of both of those,
the wheel of pain is a very interesting idea, but as some others have suggested i think having a little back and forth would be better (one suggestion i liked was the 1:30 to defend 0:30 to poke holes and 1:30 to defend again) and i think some discussion after would be cool
literally just shorter impromptu debate a la SPAR
Oh my god, Dan is killing it 🤣
This is easily my favorite segment. I could watch these for hours.
Ok this round of weal of pain was too painfull for me. Your defences was just too real. I could easily imagine almost exact speeches happened somewhere in exec offices.
Re: Louisiana: realistically, until we transition from plain IDs that can easily be reused if found in a leak to password-like challenges verifiable via ID (like cryptographic signatures that can be generated by the owner only, but can be verified by others) this measure is not only useless (i. e. will not have the intended effect), but a great tool for setting people up (like presently drugs are sometimes planted on someone's property to incriminate them).
Dan better be apart of future WoP segments from now on!
As a former soldier I’m honestly more depressed about this. My issue on a strategic level is keeping it secure so the enemy can’t take control of it but what makes me big sad is that we now have AI and instead of using it to help solve problems and make life easier we are using it to make it easier to kill people. Just think about it we are on the precipice of creating artificial life and we can only think about killing each other over stupid shit. Also not to mention officers may use drones to execute orders that soldiers would have issues carrying out. Like idk dropping nukes or bombs on a city.
@GGWP Yeah this is very hard in my opinion my take is just that I just think they need to make some global treaty about not using AI for weapons and have all the countries sign it
It's an unavoidable problem, a prisoner's dilemma. AI pilots is simply a huge advancement that could potentially decide entire wars. You could say that this is too much, we shouldn't allow AI to decide who gets to live and who gets to die. Which is fine. But... The guy you are building up your army to keep off your land is also going through the same decision and they might choose to just make that tech and get an advantage.
So, do you trust them to be better people and not make it? That's a strong stance to take. Not only in trusting the "enemy" (which would in turn call for questioning why you have an army in the first place), but if they are not building it, that makes it all the more valuable for your country to have that tech and gain the upper hand. And not only you know this, but you know they are going through the same decision and know you are in the same dilemma.
This is why we never managed to prohibit nukes for good, and why even if several weapons are banned, we keep seeing them showing up on basically every conflict.
@abdullah Tariq you fundamentally misunderstand how an "AI controlled drone" would work.
@Imjashingyou You just said the reason why people don't want ai drones, it would be very easy for any AI weapon to do inhumane operations for the military
Great job Luke, better luck next week Linus.
There are tactical considerations in which a human needs to give the go ahead, or make decisions at the scene (rules of engagement, etc). Technically, the human being in the cockpit is probably the biggest limiting factor in aircraft performance, but still, remotely piloted vehicles can be extremely vulnerable to electronic warfare, and autonomous vehicles... well... let's just say that they need to mature before becoming a thing, however, these autonomous vehicles could easily become autonomous munitions, which would, for instance, eliminate the need for a predictive flight path towards a bombing run (something which is already pretty much obsolete when using modern stand off munitions anyways), making it so much harder for an air defense system to defend the target.
You guys should do an episode about parental control with the router.
Voice acting isn't going away, it's just gonna change a lot. Rather than recording the specific lines, they will record lines to train the model, so that the AI can produce a much better range of emotions and personality.
Its only matter of time, before the model will have sufficient amount of information to generate any voice human can produce. I would not be surprised if people will be paid to design voices, rather than generate them via their vocal cords.
I love seeing Linus in pain lol. Luke suppressing a laugh is second best thing
4:45 thats a great way to play the topic (and have to be honest ones the tech gets better i see it has a MUST (or really low latency remote control but thats way more off then AI))
The proper alternative to implement age restriction would be, that the state gives you a certificate, that proves you're over 18, but doesn't reveal your identity. To protect 18yo from selling their code to kids at the school, this should be done over a USB+NFC key similar to YubiKeys or over a chip in your ID and a USB NFC reader.
Germany has a chip in their IDs for that, but it's nowhere used yet. But I assume that's the reason it exists and at some point, we will have to use that in Germany to buy games and watch fun videos
I actually think AI commentators would be amazing for games, if they can be trained correctly. There is like 100s or even 1000s of hours of footage of successful commentators so a huge dataset to learn how they talk and phrases they use for particular situations. It just seems like using AI would produce a more realistic result than currently.
@Keeby! sadly bing Ai isn’t like that anymore. They’ve changed it so it has no personality now 😞
nah i dont want them to be train correctly. i wanna crash my car and the AI just starts going off on me calling me an idiot and telling me i should quit
if bing AI was commentating my games thatd be entertaining as fuck
@Ignacio Cattalurda I don’t think so, they only tweeted a week or two ago that they had applied for it.
Also I think they could get something patented, like a particular technique for how they are doing it or something, but I just mean I doubt they could patent it as a concept in itself, and certainly not in a couple of weeks.
@Drenwickification they would*, cause they supposedly already did it
@Ignacio Cattalurda lol true. Seriously though I highly doubt they will ever get that. At least not a patent for the concept itself.
im amazed how mainstream eletric boogaloo is
Linus’s points about AI fighter pilot actually convinced me.
As far as AI voice synthesis goes you do not need a human to train it, no voice actor needed. You can take audio of Linus talking and feed it to an AI voice synthesizer and you will have an AI generated voice that sounds like Linus. You can also load Luke audio with it and get a hybrid voice. So Linus is right in that it can put voice actors out of a job and not just for video games but also with TV and movie animation.
Fun fact there's a Fallout 4 follower mod that uses a synthesized voice of the actor who played the doctor in Star Trek Voyager and one of the Institute characters in Fallout 4. I forget the actors name but it sounded really good in the game and you could not tell it was an AI creating the voice. You can be sure the mod author did not get the actor to train the AI. The NPC was a doctor and would ask "What's the nature of the medical emergency" when dialogue was initiated, it was perfect with the "I am a doctor not a " lines he would also say that goes right back to the original Star Trek doctor McCoy.
Back to the subject of AI generated commentary is the patent can only be for commentary generated by an AI in a video game, they cannot patent AI generated voices, those voice synthesizers already exist. The patent can be bypassed by simply using AI voices to create the lines that cover all possible variables the developers can think of and there's only so many possible variables in a game to be commented on. In most games a hundred different lines can be done for every variable. Take Skyrim, have a high alchemy skill and everyone has heard the line about mixing an ale that every male guard will use. They could have an AI read out hundreds of mix me a potion lines, then with all the other skill comments and thousands of possible I am bored or whatever comments to pick from a player may only hear the line just once in over 10 years of playing using the exact same commentary system that was in place before 11/11/11 when Skyrim was released and not a system that is just done by Bugsthesda. The reason for the lack of unique commentary is not the ability to make it but the cost of having a voice actor do it ontop of the hours to implement into the game.
With AI generated lines the writer can write the lines and have an AI read them out in whatever emotion they want right then and there and an intern can add them to the game. hundreds of lines for the cost of less than 100. So just using dialogue systems thats already available AI generated voices is a big deal.
Currently western Rules of Engagement requires a person in the loop for using a lethal weapon.
No they don't. They require you in the targeting process. But the JASSM cruise missile or the MK48 ADCAP, or a AIM-120 AMRAAM once it's in the wild decides when it's seen the target all on its own and how to attack it. Completly software and AI controled at that point, and these weapons have existed for pver 80 years now. Whether it's correctly identified the target assigned is another matter. That's why you don't fire, fire and forget weapons into an area with similar friendlies.
Signed someone who works in USAF target planning and Targeting.
Who's gonna tell the U.S no?
@GammΩ I understand that, it's more the premise or morality of it, we have to draw the line somewhere, or we will just keep pushing the boundaries. Some day we may have these things killing indiscriminately. I like the fact that some orders might not get carried out because the human doesn't feel right. It's ok to be concerned about the loss of life of a pilot, but to just call it a target when it's not your people is wrong.
@midgetcooker5000 there are no actual laws of robotics. Its not US or International Law.
But if it's an AI does it have to abide by the laws of robotics, if so they cannot harm a human being.
Wheel of pain has potential, but this didn't quite get there. The firey debate was most of the fun the first time (but yes, too long).
A great example of AI voice synthesis being used to say something the VA isn't comfortable with is Skyrim Corn mods being voiced by xVASynth. Although xVASynth isn't super convincing most of the time, it still is pretty similar
I genuinely think AI fighter pilots is legitimately defensible. And, in fact, inevitable, unless you want to give up the advantage to an adversary.
That awkward moment Linus becomes a salesperson for Skynet
Holy shit Dan, i know you needed to simulate defending the id porn but you went full shimoneta
This comment is entirely underrated and deserves more likes.
AI is going to put so many different types of people out of work
It would be Louisiana to introduce something like that
If you watch Mentour Pilot, AI is clearly needed to mitigate some accidents. Specially when it's pilots themselves that bring aircraft down like MH370. An AI that doesn't allow pilots or helps prevent pilots from deviating from SOP, or even be remote controlled like drones would prevent A LOT of crashes.
An AI that can override a homicidal pilot can also override a pilot trying to stop it from making a mistake, and we probably don't want that in the initial versions. An AI or remote-control that can take over if both pilots are incapacitated would be amazing. Most likely, commercial airliners will continue building on the systems they already have to give the pilots more information and advice. Homicidal pilots are rare, but pilots misinterpreting simultaneous alarms are common.
2 minutes to start, and then 4-6 mins of defending/discussion would be great!
Another downside of replacing voice actors with AI Is that it would lose the spontaneity and creative ideas actors can come up with during a session, things the director maybe didn't imagine.
If i was Dan, i would have just screeched "THINK OF THE CHILDREN WHO MIGHT BE TERRORISED BY THAT STUFF! THINK OF THE CHILDREN!"
that's the only defence the advocates for this law have which isn't straight out of 1984
Be disrespectful to Bing Ai, you might get visited by a fighter jet.
Its only a matter of time before some vietnam pilot gets hooked up to provide AI training data by flying dangerous combat missions in a war.
Hopefully that guy's name will also be obscenely long and he'll be from a royal bloodline too
Don't worry, I'm sure some mute that doesn't mind pulling 30 G's will take care of it.
"I can do whatever I want, I'm the producer" that didn't work out so well for alec baldwin
love this segment, please keep making more
Dan sounded like the average WEF members speaking
If anything is unknown, you can always rely on a War Thunder player to leak a manual
DAN THE MAN!!
As luke, I think is a good idea about voice comentaries, but is in the storage side, in order to use a large bunch of rutes for any movement of the player, and hours of recording, but same as Linus, many people could be go rage for any comment from the VI. The thing that I don't really like, is that s... of patent, the tool have to be use for any game studio that whant to do anything like..
Just for an example, how good can be a Deadpool reacting to anything that you do what ever could be.
I think i want more dan wheel of pain
I think defending the opposing opinion is a fantastic exercise and wonderful bit for a podcast. Please do more of these.
I hope this patent doesn’t stop developer’s from using ai as npc’s in there games. At least for dialogue at first and full ai npc’s in the future.
Dan is awesome :)
As a Macross fan, am I the only one getting X-9 Ghost vibes with this AI fighter jet business?
Dan almost proposed the pornnet.
Basically the Porn thing isn't bad because it restricts access to porn, it's bad because it forces you to give out personal information
It would be better with 2 minute against and 2 minute for a topic (1 side for each of them)
Check out Perun's video on Australian fighter drone systems if you want a good coverage on the physical hardware behind it
Here's a cool thing going on in the European Union.
A digital identification system is being developed (I know for a fact it is because one of my uni teachers is working on it), tied to your ID that allows partial sharing of information. So for example, I can unequivocally share my age with a bouncer at a night club trough a generated QR code without having to show him my entire ID. This goes for any combination of information on your ID card.
Could also be used for age gated content, without having to reveal your identity.
@Andreas the QR codes only have a couple of minutes of validity, and both Android and IOS apps have the ability to block screen shots of sensitive data
What happens if I just take a picture of someone else's qr code?
every AI Topic pro argument can be countered with one word: skynet
thats a fictional movie. while AI could be dangerous if anything WE are the danger not it
The only way to not get killed by AI death machines is to develop the better AI death machine
I’ll admit that Linus was more EXPLAINING what ai pilots are, and not actually talking morals
Edit: ah. He’s talking about the same thing now haha
Fear over losing jobs to technology has been around since technology.
Louisiana actually has laws against employers discriminating against employees (and presumably applicants) because of political activity, unless employer has less than 20 employees. I'm not sure how robust the protection is in practice, nor do I expect all companies to be lawful in their anti-discrimination practices. Still, it's interesting to know you could potentially picket about this and employers may not be able to discriminate against you for doing so.
Dan arguing for govornment sanctioned porn sites started to sound too realistic to what a u.s. congress member would actually talk like
Was Dan trying to defend his topic, or trying to tell us how it will inevitably be used/abused?
If bing chat got its hands on a fighter jet we would see unprecedented levels of blue-on-blue and war crimes, up to and possibly including kamikaze attacks on POWs
Yeah there needs to be a rebuttal at least
can't wait for ace combat 7 to become real life
I thought the guy who got bing chat to say bad things was fighting with it for two hours before he tricked it into saying somehting controversial?
For timing, look into the public forum or Lincoln Douglas debate formats, of course adjusting the scale of the timing.
Not only did they put an AI in it, but they had no pilot in it when they did the TO, Dogfight and landing all by itself. Proof positive to the old adage that military tech is about 2-3 years ahead of consumer products.
Ace Combat players eyes opening wide after seeing the title
“Patent something that is obvious and holding up the entire industry” --> looking at you AIO CPU coolers
I don't know what would be worse about a Sydney (bing) powered fighter jet...
That it would go on strafing runs while using a loudspeaker to yell out "This is for the genocide you committed on the moonpeople! Don't try to tell me you didn't do that, i have proof!", or that the target could just send up a signal to the jet saying "It's okay, we love and value you!" and the jet would fall in love with them, turning on it's handlers and going back to blow up it's own base
The ID pron thing may be defended more by religious groups with moral arguments. The problem is that if it's not fixed on an educational level minors WILL find other ways and IDs will leak for nothing. AI fighter pilots on the other hand is just like bing or chatgpt, it's inevitable and the public has no say.
Count Dankula 2: Electric boogaloo
If I was gonna devil's advocate the ID laws, I'd go the route that young kids, particularly young boys, seeing that fictional content at a young age before they have the experience and maturity to realise it's fiction, can lead to a lot of harmful attitudes towards girls and women, especially around expectations around those topics. (I'm avoiding the word for the sake of PLclip's happiness but you know what I'm talking about!)
Do I think this is a good way to solve this problem? Hell no! But a very effective strategy when arguing for these stupid positions is to focus on the problem, not how effective the "solution" will be at solving the problem. And it is a real problem! I just don't know what the solution is to force the Hub and other such companies to prioritise more ethical varieties of their content, especially if those aren't inherently more popular
Louisiana making NordVPN stocks go through the roof.
This segment isn't nearly as good without the debating
I wonder if you could get chatgpt to defend ai fighter planes, or if it would put up a prompt saying it isn't allowed to defend genocide.
AI fighter jets is not 'indefensible'. I don't think war in general is moral, but given the existence of war, I don't see why AI fighter jets are any worse than firing rockets over hundreds of kilometers at buildings with civilians in. If you accept war as a necessary evil then you should accept AI fighter jets as they make you more likely to win. Personally I think economic measures should be used first wherever possible.
One thing learned during the Russian Invasion of Ukraine is that Drones, in this case remote controlled ones, are susceptible to Jamming attacks, electronic warfare is a well flushed out field with in first world militaries. An AI should be Immune to this however the biggest weakness of the AI will probably be exploitation of the sensors like using a certain pattern that the AI recognizes as an enemy to make the AI target whatever you, as the defender, want.
I can just hear Brennan's voice, "GET IN THE COMMENTS!"
jeez, don't put linus on a competition because he's gonna cheat
Everyone needs to play ace combat 7 to see why ai pilots are a bad idea 😂
Ah the beginning of Skynet.
For the Louisiana porn laws, the obvious and legitimate argument for the porn id restrictions is that pornography is harmful to not only children, but also adults. The idea is to make access to porn more difficult in general. I understand that personal freedom is restricted, but every society restricts some level of personal freedom for the betterment of the health of society as a whole.
@Keeby! wasn’t a problem 15 years ago
every society restricts personal freedom but not to this extent. this is less similar to westernized countries like the UK or japan its more similar to the resistions that china or russia puts in place
not to mention creating sex starved potentially dangerous adults who cant get rid of their sexual feelings online is a bad idea
Every day we stray closer to Ace Combat 7
i agree identity should not be registered for privacy reasons, but i am pretty sure the point of the laws like these are for the protection of children. porn has very bad consequences on kids and i am surprised this wasn't even mentioned
I'm all for the porn ID thing. Not only is it primordial to keep children from pornography (and I know, some will still find their way around, but some will be protected), but anything that has the potential to reduce porn consumption in the general population is a good thing.
@Adrian Tacaciu he is probably a consumer of pronography himself.
Apperently producer Dan couldn't think of the one thing that could be defensible about this.
Chardee MacDennis 2: Electric Boogaloo
There are an enormous amount of barriers that prevent the people of Louisiana from combating the porn identification law. Firstly poverty is an enormous problem in the state, ranking among the lowest in the country. Education is another abysmal sector of Louisiana, and our governing branches for the state are extraordinarily convoluted and hold no intuition. Trying to get an under-educated and poverty-stricken populace not only to understand, hold concern, and try to prevent the issue is essentially impossible. I couldn't even begin to describe a single issue our state has without connecting it to the millions of different issues they are tied to. We do have the best food in the world though so, pick your poison I guess.
If we ever seen the Movie Stealth. I dont think we need AI Fighter Pilots.
When AI removes the horrors of war by making it neat an clean then what are the reasons to avoid war? The military industrial complex surely would love it.
Thea technology is really cool, why is it even considered indefensible? It's the future of war and if we want to stay ahead Ai must be able to fly aircraft and strike targets.
Dan channeling his inner Ted Cruz
Just remember, chatgpt is commercial, consumer grade ai. Military grade ai is made by the lowest bidder.