Am I Real?

One of the realest videos I've ever made
0:00 - GPTZero VS my videos
3:02 - GPTZero VS AI generated content
6:04 - Conclusion?
9:20 - Conclusion about GPTZero
12:01 - Humans over-relying on AI
14:00 - Me VS AI me
15:31 - Artists VS AI art
17:01 - Whose side are they on?
17:41 - Am I real?

[@existentialselkath1264]

I feel really sorry for anyone who's had essays and the like disqualified due to awful 'AI detection' tools when they're written by real people in specific writing styles.

[@memegazer]

Seen several of these stories in the gtp sub on reddit. Sometimes the even showing the version history of their work doesn't change the accusers mind.

[@lowcostfish]

These tools unsurprisingly have more false positives when the writer is a non-native speaker.

[@xponen]

​@@lowcostfish In that case, I think they program the tools incorrectly, because it makes no sense that an AI trained on English language will write like a non-native speaker.

[@Cheese-Goblin]

@@xponen i'd guess its because it would be written clunkier

[@lowcostfish]

@@xponen As Cheese-Goblin says, I think it's about the clunkiness.

[@eichen97]

would be funny to have a "ZEROkliksphilip" with a couple videos, 100% AI made as an april fools joke.

[@beamboy07]

that would be hilarious

[@MBkufel]

nokliksphilip

[@zwenkwiel816]

10 and 11kliksphilip

[@kez963]

24/7 AI generated zerokliksphilip livestream :D

[@Kreuzrippengewoelbe]

nullkliksphilip

[@DeepWeeb]

How to tell if kliksphilip is an AI or not: *only the real Philip would say "hella-la-la-lot"*

[@PrepareToDie0]

AI be like "write that down!"

[@eccomi21]

@@PrepareToDie0 now it will

[@sketch8988]

at this point i wonder how old this joke is by now, ive been watching for several years and i think this joke has spanned over a couple of years at least

[@Pawlakov1]

I can't believe an AI generated NPC from a video game has passed the Turing test damn you Josh

[@slenderpaul]

So?

[@MsZsc]

more like a dumb captcha

[@Nimroth]

I guess I owe you an AI-pology, 2kliks. I now understand that it was your natural voice all along in that kliksphillip video, and I deeply regret spreading misinformation. To be fair, your voice did sound quite different in that video and as if it was recorded on a podcast mic. I apologize for jumping to conclusions. Keep up the great content! Thank you to ChatGPT for writing this comment for me.

[@tomsko863]

I never thought 2kliksphilip was real. 2kliksphilip has always been a force of nature.

[@vibaj16]

2kliksphilip is a shotgun confirmed

[@gabrieltiberius1076]

THIS is why I love this channel! This is why I think the klik empire stands out to me so much! What an awesome video! Thanks Philip!!!

[@jumbledfox2098]

Hello Gordon Freeman!!

[@gabrieltiberius1076]

@@jumbledfox2098 It's good to see you

[@michaelFoucheGodIsGood]

Phillip I've moved to England recently from South Africa 🇿🇦 always thought I'd love to meet you the more I watch your kliksphilp videos the more (which I've always thought) you my favourite KZbinr for 7yrs Down to earth funny , British humor , existentialism , CSGO , witty clever videos from the heart . You inspire me , to eat beans , 0 noodles 🍜 and to be more careful with my money, if I am ever in Plymouth I sure hope to bump into you. Keep being you and know you help and entertain just a massive amount of people , you're much appreciated 🎉

[@SomeRandomPiggo]

I live very close to Plymouth and visit often, sadly not yet

[@512TheWolf512]

Ah, a refugee! Welcome to Britain

[@sexydog]

@@512TheWolf512 how is a south african a refugee

[@michaelFoucheGodIsGood]

@@512TheWolf512 actually here under Ancestoral Visa , just because I've moved country doesn't mean I'm a refugee

[@Twizted86]

@@michaelFoucheGodIsGood British humour, is dry and something you'll have to get used to as a newcomer. We share a similar style of humour here in Australia. We also bring in a lot of refugees here, so I had a laugh at the comment even if it was at your expense. Immigrants/Refugees are interchangeable words here.. and obviously there too :)

[@mrtaufner]

"Everyone likes tech until it comes for their jobs/hobbies" is an argument a friend of mine does all the time. As a part of the IT industry that's something that is definitely growing on me too

[@starlight_garden]

Simple. Don't have a job or hobby.

[@Bonelord69]

You're making my head spin Philip. Soon we will have AIs writing, recording, publishing, then scrutinizing their own videos.

[@bjk0norway0bjk]

Kliksphilip, Your explanations are so clear and engaging, and I love how you bring your passion to every topic. The visuals are awesome funny too! Can't wait for your next video. Keep up the amazing work!

[@TheDiymovies]

I so so so look forward to re watching and reflecting on these videos in ten years time haha

[@mymanzach8265]

Love the completely non-ai created and voiced videos Phillip! Keep em up!

[@smartguy8219]

this type of comment gonna be on a lot of videos very soon

[@villus9384]

Phillip, I have been watching your videos for years. I've gone through all of high school and almost all of college now while watching your videos, and truthfully, an AI could never replace you to me. I've come to love not just your videos, your editing, your style, and your attention to detail, but you yourself. Call it parasocial, but all your little jokes and how much you share on kliksphillip has had me come to know a lot about you over the years. I truly hope the day that AI puts you out of work is very far ahead, but I'll watch your videos as long as you make them. Cheers, Phillip

[@eccomi21]

i think this is peak definition of parasocial relationship and could hurt you in the long term depending on how serious you are about it right now

[@vile8366]

This ran on a bit, but my point is, that I don't think that being loyal to and following Philip is necessarily a bad thing, provided you still have a life outside KZbin. I think of him like a writer I admire. Just his is video and spoken instead of written word. He brings up profound concepts along with humour and further insights into life I've yet to come across. A source of wisdom for my ''coming 20's'' if you will. Been following more of the klik empire than just the 3 klicketeers for the last several years, and I'd say I repeatedly draw parallels to my own life out of the blue that I'm then reminded of came from one of Phils more introspective segments. Phil isn't my friend, but he's one of several beams of light illuminating his part of the path of life that I have yet to traverse. And I appreciate him for that. Just as I appreciate some of my teachers, a few older relatives, my parents and anyone else who has taken the time to talk about their younger years and what choices got them to where they are now. (Of course I admire my parents more than some guy on KZbin, but he's definitely more useful as a guide than most of my relatives, seeing as the worst of them have passed 30 and are fully able to, but choose to have no job. Solely due to laziness and a curling mother.)

[@villus9384]

@@eccomi21 what are you on about? I just mean that I feel like an AI couldn't replace Phillip and I really enjoy his sense of humor

[@villus9384]

@@MasterVSB you're an interesting critter bro but I like it

[@0ia]

I like hearing your voice. It is comforting given the time I've spent growing up along it.

[@Daimaiju]

Hope you had a great weekend Philip & have a great week everyone. 😁

[@uncle-keg]

This was very good video Philip. I've had similar ideas on using AI tools as just that, a tool to help develop your own skills, since it started to get good enough on replicating some human tendencies. It's nice to hear the same sentiment from a content creator and one who has done a lot of work talking about and with AI. Thank you for the reassurance.

[@sgas]

Your data visualization is great as always

[@Cryzeu]

Remarkably thought-provoking video. I haven't had too much first-hand experience with using any AI tools myself as I haven't been able to identify a meaningful utility for them in anything I do. I personally consider this to be a lack of creativity on my part more than anything else, and I often feel a bit left behind by those who indeed do embrace the options that AI tools can make available in order to further their goals. My only experience with the concept at all is via videos like this, and I greatly appreciate being able to learn more about what can and cannot be done, and where things are heading. It is relieving to hear such a well-adjusted point of view on such a complicated topic. I can't say that I've ever felt suspicious that any content that I've recently consumed has been AI-generated, but maybe that just makes me a gullible fool. Or maybe I don't hang around in the correct places. I've still been more or less under the impression that fully AI-generated media is more akin to the example at 17:50 where it is obvious beyond a reasonable doubt that a human hasn't been involved in the production to a meaningful degree, or rather at all. Watching some of your videos, especially videos like this one, sometimes make me feel inadequate as a person as I feel such a strong appreciation for the way you conduct yourself and how you approach hostility and change. It seems that you often face a bit of backlash from all kinds of individuals for all kinds of bizarre reasons, but I personally can only wish that I might one day have such a humble and yet so densely rich perspective. Thank you for your content.

[@kingkurtis]

As a kid I dreamed of machines replacing work not fit for humans, dangerous, dull repetitive work to allow us to find better ways to use our creativity and intelligence.

[@teensuicide9103]

I really like ur vids Philip they make me smile everytime, for years ❤❤❤

[@schto173]

You would be a good teacher :) I am a teacher in IT. We also suffered through the chatgpt surge but certain students prefer their own words to those of gpt. Some don't care at all. As teacher I believe students should always have the best workflow possible and use state of the art tools to score high. If ai can help with this they need to be able to use it otherwise someone else might outperform them later in the market. So I teach with chatgpt but also explain the difference of human content and ai generated one. Your video actually helps explain this nicely

[@zenolord2242]

I haven't watched your videos in ages, but last night I had a dream where I met you and chatted with you for ages, it was pretty cool.

[@ArtemisWasHere]

Man I wish Philip was real

[@Povilaz]

2kliksphillip is powered by AI confirmed.

[@halfsine]

hes been an ai the whole time...

[@kozakalex222222]

no he's achieved with sorse

[@IAhmadGT]

​@@kozakalex2222222

[@BigJo3]

I have long been debating to myself about how I feel about this AI stuff, thank you for helping me reach a more solid conclusion about it. This is a scary topic, especially for the world of creatives. A truly unprecedented time we live in man. Who knows what the future has in stock for us. Final thing probably don't mean much but with the amount of hours of klik empire content I have consumed, I can easily tell AI you apart from the real you (voice wise), at least for now. Edit: Forgot to say directly but if it was not clear I loved the video. As with a lot of your videos quite thought provoking stuff. However much AI you used in this video if any, does not bother me. I put all my faith in you that you will remain true to your standards and that it will used to further improve your videos.

[@homeyworkey]

@@2kliksphilip You say ai-generated stuff is going to change things in 5-10 years. I think you're underestimating its progress, even 1-2 years there will be absolutely massive improvements. Gemini's coming out soon which is supposed to better then gpt4, openai said there always ready to create gpt5 (they're waiting for other competitors to catch up it seems), anthropic (i think it was them?) were talking recently about in a years time a possibility of them creating an AI model which is 10-100x the scale of gpt 4 (that would be so ridiculously expensive, you would think they would only throw this much money away if the payoff is worth it). the amount of money & attention getting pumped into this is going to cause an extremely explosive trajectory in the near term. chatgpt isn't a year old yet, and the fruits of this product has caused every researcher and their brother to get involved in AI. we're going to be seeing the impact of all this attention very soon. btw, i think u should really watch "the a.i. dilemma" on YT, very interesting, best AI discussion ive watched. its crazy (to me) that chatgpt is so mainstream, to the point boomers are talking about it, but there seems to be an extremely small market for people actually interested in the technology and watching videos on it. Ai Explained is like the best ai channel, but is only at like 200k subscribers.

[@homeyworkey]

@@2kliksphilip LMAO fair

[@R3ddyyg]

Idk why but i read "Am i feral?" as the title and instantly clicked

[@crazydave6787]

This channel is a goldmine

[@Corei-eg1tb]

Looking fabulous👌

[@sjorsdewit9394]

Damn I was just hangover laying in my bed when I found this video. Now my head hurts even more lol🤣

[@bilbobaggins5752]

I think people might mistake your videos as AI generated because of the careful way you read your scripts. You are methodical with pacing and probably have tricks which would lend your recorded audio to be easily edited. Like coming to a complete stop between sentences, which is not natural for regular speech, but makes it easier for you to reread misspoken lines and edit them out.

[@stinkypete9070]

"If I can't tell the difference why should I care" be careful with that - goes both ways. If eventually, someone can just feed your voice to a machine, get it spit out scripts and copy your music, why should they care about your videos? Especially if it got so low of a barrier I just type it on their start bar without even having to come to KZbin. No matter how much you pump out more content, no matter how much it improves, Writing scripts or even the very act of creating videos is just you "resisting change" too. This isn't so much as a new branch of exploration adding to the tree of options so much as it is taking a big chainsaw to the main branch. Cultural/Contextual value should matter, it's kind of the point of entertainment/creativity. It's a reflection of you, and your abilities and your personal filter at the time. Its why I've watched pretty much all your videos on all your channels over the years, I particularly love the video on other channel called "the simpler things in life." I get people have to min/max to stay relevant on YTs algo, but that seems like it's treating a symptom to me what's the point of putting out more content faster? Less is more and all that.

[@Telecrusher]

I always come back to this video to listen to the AI rap with Philip's narrator voice.

[@Alexmagno7]

Gracias Felipe 😘

[@gorgetted5265]

Fascinating video! It's incredible to see AI's potential and how it impacts us, but we should be cautious of AI detectors' limitations when it comes to privacy and security.

[@RaichuKFM]

I do think that the argument against AI-generated essays is that in a lot of subjects the point of assigning essays is to do the analysis and develop the reasoning skills and the essay is a product to show that work had been done, rather than the desired product itself. So all AI generation could be doing is shortcutting that analysis and reasoning (the whole point of the essay) and thus deprive the student of their learning, or not do that and waste the student's time as they have to edit all that reasoning in anyways. I think there are edge cases and some people may well benefit from being handed a skeleton even if they rip up everything originally in that skeleton, and I'd be very worried about false positives impacting students on the testing... But I do think it *is* more nuanced than being able to finagle and edit an AI into producing a good essay proving you have sufficient skill at essay-creation as writing a good essay yourself, as often the essay itself is meant as a means to an end as a learning experience which builds its own assessment tool, rather than being bulk practice for writing essays, where I think that "if it's good then it shows they can produce a good essay, alternative means of doing so is fine" is kinda aimed at.

[@quacktape]

Here's a taster of the start of Bing writing a 3kliksphilip video: "Hello and welcome to another episode of 3kliksphilip, where I go into extreme detail about something small but substantial about Counter-Strike 2. Today we're going to look at the new map, Dust 3, which was released in the latest update. Dust 3 is a remake of the classic Dust 2 map from CS:GO, but with some changes and improvements. Let's see what they are..."

[@floor.mp3]

Good one Philip . Keep on clicking !

[@dessertion]

LMAO I love the fact that you added in the AI-generated conclusion you showed in the video as the conclusion

[@Summanis]

3:48 the AI script read "Stable Diffusion, DALL-E 2, and Midjourney are not real AI tools, they are made-up names that I used as placeholders". So close and yet so far.

[@Surepeacooler]

The title of this video sounds like the name of one of steventhedreamer's songs

[@Captions-Eric]

Be ready for the basilisk

[@GoatedWithTheSause]

I was one of the people who thought that you used an AI voice generator trained on your voice for that video. Not because you're lame like that or to insult you, but I expected it to be an experiment, and that later there would be a video that would reveal that one of your previous videos had an AI voice over and that it was so good that people didn't even notice. In reality, you were just under the weather, so your voice sounded off.

[@chrxstt]

i have had no computer and only game streamed counter strike through geforce now and watched your videos for the past about 5 years and this past weekend an old family friend by the name of nico gave me a 2013 computer and i can finally play the game (albeit at a low resolution to run it over 28 fps) i know this video isn’t counter strike related but i enjoyed it and thought i should say this somewhere, much love❤

[@readmore8974]

5:29 Tip for future use, you can just respond with "continue" and it'll keep going

[@Surepeacooler]

You really made me think you were gonna reveal steventhedreamer's next album cover.

[@benjaminaburns]

What a trip, second guessing what part of the script was "real." IMO it's important to understand AI is a tool that will help augment existing skills (but not always replace); and it takes effort to successfully learn how to use a tool; and "society" needs to come to terms with the fact that these AI tools will be widely available (like in essay writing). This video touched on pretty much all of my current thoughts on AI (and related philosophical questions), thanks for making this (and CSGO content!).

[@TheDiner50]

What? You believe that you are going to carry a calculator with you all the time? You better learn how to do math without it. No. AI be dammed. People with the skills to do stupid math in there head always surprise me. There is no freaking way people should be expected to rely on AI. And do not confuse AI with the marketing term of automation and programs. I mean the AI that needs a 24/7 connection to some server in the clouds. That stuff can not be relied on to augment our skills. Since pocket calculators barely is a thing anymore. They are just a tiny bit of code running on pocket super computers. And so the skill of the avrage person has gone from being able to do basic multiplication in the head to rely on the pocket tool. Where the calculator makes up less then a dollar in cost of the device/tool. Augment or life with a reliance on artificial intelligence that we do not control? And are just going to trust to not be used agenst us? No. It is in the name. Artificial. The calculator is a tool with scientific and truth at the core of being useful. 1+3=4. Some real "AI" is a bunch of made up truths and rules. By someone. For some reason that augments our life after what it deems right and true. To accept that people willingly gives up all privacy and self reliance? Like hell no. And not to mention the shit possible with faking stuff.

[@higorhi72]

@@TheDiner50I found the first paragraph of your comment a nice reflection. Thank you for this. But the second one is a directionless rant. No, AI isn't "rules made up by someone", and in many ways the "artificial" you emphasized is based on the same exact way we learn. Don't take the shortcomings of the first 6 months of commercial GenAI to mean these things can not really take human-level comprehension and performance. It's almost futile to try and estimate how we'll use these things in the mid- and long-term future, but be sure these things will change our lives in ways the calculator just couldn't. Calculators do things we already know how to do, but faster - AI might be able to do things we maybe didn't even know were possible (such as efficient protein folding). They will become way more than just our "text calculator" - they might just be what either takes humanity to its peak or to its downfall, it's really up to us (as in humanity, not "us" as peasant users of this tech).

[@flappybird_fpv8370]

Yoooo lets go

[@woodbyte]

I'm not surprised. AI Josh was the most human.

[@andrewhamel1]

Philip.. are you wearing a pebble smartwatch in that glam photo at 7:34??

[@troutsy6897]

Philip, this video is your masterpiece.

[@seaturtleman]

I am inclined to believe this is an extremely meta meta video, where its 100% ai, with a monologue about how much better you as a human are at making the content.

[@HeathDiegert]

Got a great voice and good content

[@Tatsuki09]

I'm here so early now I'm questioning it myself

[@Sopel997]

WOW! AI is really getting incredibly good. I was almost fooled by this video!

[@friendofp.24]

Holy hell do I hate AI now. I can't trust anything. I'm just gonna go outside.

[@richardhunter9779]

People are going to use AI outside too.

[@SubjektDelta]

you should think about what trust actually means and what you even can trust. the internet has always been full of misinformation..

[@beetheimmortal]

Why? It's so incredibly obvious when something is and isn't AI.

[@marharols5793]

In Norwegian highschools they say that AI is a good tool that can help you with tips and help you get started on writing or getting summaries. Which i feel is quite accurate. Therefore its allowed as long as its used as a source.

[@zwenkwiel816]

Using AI as a primary source is dangerous though. They tend to hallucinate quite a lot (basically just make stuff up) There was this lawyer who was using previous cases to argue his case but all those examples were just made up by chatgpt... Should always ask the AI to cite its sources.

[@marharols5793]

@@zwenkwiel816 Yeah, no thats what i meant. Like if you ask it to mention some important factors for the economy in Europe around the 1920s, then research the things it says somewhere else to get a more acurate depiction. And of course its in addition to everything else. Its meant to just be a starting point for further research.

[@poipoi300]

Hey Phillip, sharing this cause I think you might find it interesting if you don't already know. These AI models have a "temperature" parameter between 0 and 1. The higher the value, the more unlikely the responses are (simplified). As far as I know, models from big tech all use very low temperatures in their chatbots like 0.10, while I use 0.65 on the models I use for fun. This costs coherency, but the results are more interesting and still very coherent most of the time. This can be combined with Generating only parts of the text that's asked at a time, with different temperatures. You could even split up the prompt. On the first batch, you would ask "in a serious tone". For the second you might ask "lighthearted". You could ask it in the style of a completely different person, too. I suspect all of these together would boost both perplexity and burstiness. You (generally speaking) could even train text embeddings so that you have many kinds of "serious" which you could alternate between when a person prompts for serious. We're still in the infancy of the usefulness of these technologies, and small model-specific stuff like that will only really be mastered much later, if by then there's not some new and better technique lol. Still, these detection systems are already not good enough to rely on. Logically, they will never be either. That would require them to understand the subject matter (language) better than AI models that are trained specifically for that purpose.

[@AbedIsBatman]

This is wrinkling my brain

[@Splarkszter]

i tohugh the video was about to end at the 7:40 mark but there was over half of the video left o.0

[@GGPlex_]

Woah.

[@mfnbpwnz]

a shame knowing most won't make it to that excellent conclusion. great video.

[@JonasTisell]

Sorry for blaming you for using AI. My comment reply was in the video haha

[@PabloPerroPerro]

Thia vid reminds me of the old "Humans have no apply" video response

[@Flow86767]

Beautiful thumbnail

[@rwlkr]

Was confident the conclusion was human generated through and through, but then started questioning if the section after it (explaining that it was written by a human) was AI or not. For now (as long as no one makes me second guess), I feel confident I can differentiate pure/lazy AI and non-AI for now (or maybe those using pure AI are too good at it with personally trained models and such), but I'll probably get more paranoid as AI gets better as I value the authenticity of the work a lot, rather than just the merit of the end product. I also feel that once AI reaches a certain breaking point, its content will become popular on most platforms (as the internet is already shifting to valuing things such as short-form content, quantity > quality, rage baiting, lies that people don't fact check, etc.), and while existing content creators will continue to do well, new people will increasingly not want to start (AI can't be a better 2kiksphillip, but what about some new guy with 10 subs using bandicam?) and existing creators will increasingly lose motivation or burnout when there are other people using AI that do 1% of the work for the same success. There will always be people who value care, effort and authenticity, such as a lot of us here, but N-kliksphilip is out here doing no gambling sponsorship on principle while other creators are making millions with little repercussion. It's not that adhering to morals can't be done, but it's tempting to not when there is a much easier and more lucrative alternative. Also, I'm gonna stick to not calling out AI unless it's blatant (though if it is blatant, I probably won't have to say anything anyways). Maybe it is AI, but it's easier for me to just ignore and move along. I've already seen multiple artists suffer from being accused of AI because they have a lot of small detail, are just really fast, or just unfortunately have an "AI style". Now I see art works coming with time lapses just to prove it's not AI. I don't want to contribute to that.

[@zalievac]

I love your content

[@e3.14c4]

Just give offline AI tools, and also the tools to proof the inputs of the ouput for business. Then you won't have thieves operating so easily, but on the other half of the coin they also don't care if they'd have the heirarchal power.

[@Hollowtriangles]

I know it’s not exactly a controversial opinion but I really hate the trajectory of AI today. It’s far too focussed on automating art and culture, the things that humans enjoy doing. It needs to be focussed on the automating labour (alongside governments introducing legislation to protect the income of those lost jobs). What’s the plan here? We automate all of the leisure activities so that we can all spend more time at work? I’ve flip flopped on my opinions of these LLMs for a few months now and I’m coming to the conclusion that overall they’re a net negative for everyone. In a lesser important scenario we’re just creating a bunch of extra crap for search engines to trawl through, bad AI images, bad AI text. It’s like unleashing a bunch of bots except the bots are people who think they’re adding value to the world with their “prompts”. I find it hard to believe that anybody at all has interest in viewing AI art, isn’t the whole point of art the human behind it. The skill involved, the passion, the history and the life of the person creating it. AI content can never have any of that, you’re consuming an algorithm. It might be technically impressive, but AI content can never be artistically impressive. TLDR: Let us do art, automate jobs, pay people

[@xntumrfo9ivrnwf]

Could you do video about soopaconducter?

[@Kilgorio]

wow

[@strangelf]

I saw an ad where I am pretty sure they used AI to copy (what sounds like the only) voice actor for those types of videos. It was eery hearing it from a well respected company.

[@NormAlTheEnd]

very good video

[@shivanshagarwal12]

im just amazed how good 2klikphilip is

[@lowcostfish]

I think I know exactly why AI generates those patronising summaries. Have you ever Googled a simple question that has a one line answer and clicked on any of the generic sites that have the answer? They all pad out the article with such patronising explanations, before finally giving you the answer at the bottom of the page, when you've scrolled past all the ads. The internet is full of these so I would bet the training data is too.

[@emberytp]

Yeah, there's a satisfaction I have that although AI can get better, the information I'm looking for already exists somewhere. Improving AI could only speed up the experience of scrolling to the bottom of stackoverflow. Otherwise it cannot give me an answer that doesn't already exist. It'll have to crawl alongside the rest of humanity's search for answers like "what is the least painful way to setup arch linux"

[@sergio9026]

But if you can't tell, then, does it even matter? - that's some good Westworld shit right there. I miss that series.

[@KaaptnIglo]

the issue with essays is that the process matters as much as the result. I suspect that if you let the AI write the essential of your essay, and you only verify that it isn't crap, the grade may be good, but the retention / learning will be much lower. And ultimately, learning is what school is about, not grades :-) Anyway, cool video as always, thanks much!

[@vibaj16]

If I could be sure I wouldn't get caught, and that the grade would be good, I couldn't care less about "lower retention/learning". I'm not gonna be a writer and have no interest in essays. The only effects using an AI to write my essay would have on me is that my grades would improve, my mental health would improve, and that I'd have more time to spend doing anything other than writing an essay. It's a win-win-win.

[@user-bt8xr5si9y]

If you use AI to write the most essential parts of your essay and then double check it to make sure it got everything right that means you are already familiar with the topic and already know exactly what to look for but you were either too lazy to write or are simply bad at writing in the first place. Some people may simply just be bad writers that can't write a proper essay correctly and use AI to do this instead, double checking afterwards that it got the information correct. Also it would be nice if school would be about learning but grades are valued in our school system above anything else.

[@kieranhurst8543]

@@vibaj16 so basically you're a lazy asshole who wants everything handed to you on a platter?

[@zombieguy]

@@vibaj16 I think it highlights the larger issue with education in general. I went to uni for game programming, yet my final grade was based on solely me writing a research paper on a topic I hated, due to me needing to get it approved. My previous years and classes never once even touched on how to write a paper, so I learned nothing while getting stressed out in-between strikes and COVID shutting down my accommodation. I would use AI in a heartbeat to pad out a paper that required a min of 10k words, like I literally don't give a fuck about how animation IK became a thing, If I have the data then it actually is just glorified padding to write about it.

[@zwenkwiel816]

​@@vibaj16the essay usually has a subject you're learning about though (unless it's just English or literature class or something)

[@sgas]

Genuinely excited to hear you use ai philip

[@oskark7676]

Love u, and u are real.

[@Veptis]

Almost a year later and we do have great interpretability research. But the only working method is invisible benchmarking. Or surprisal/binocular

[@EdgarRodriguez-gi9mw]

Damn you must be Ryan Gosling on drive cause you're a real human being

[@potts995]

I can't help but feel there needs to be some detection for formality of language. At least most popular AI models tend to avoid swearing, lewd and off-putting commentaries, etc. Real people don't hold back like ChatGPT does when it comes to sharing shocking opinions.

[@zombieguy]

It's the final part that scares me the most. Results from these AI tools are not creations, they are a copy and paste of many people's work with some fluff, yet we have large communities of people that genuinely believe they are the creators of what is output from the AI, and that's only going to get worse as they improve. Often these groups are filled with people who don't have the skill, passion or creativity needed to make something so they become aggressive when this fact is highlighted. I've seen it more recently with programming, where non-programmers are spamming their broken AI code on help forums with no intention to learn, improve or fix issues. They just want results for free so they can make money off of others hard work and with no credit to them. Ironically with web scrapers this will in turn likely improve the AI tools.

[@Goldy01]

Just adding to the second paragraph - I think AI programming is where the scary part starts; when a program can write code, it's just a matter of time... what have we done D:

[@meatisomalley]

If you explain what burstiness and perplexity is and ask ChatGPT to rewrite an output and increase those factors, it will more often fool the detectors.

[@nixel1324]

AI tools: in case your impostor syndrome wasn't strong enough yet.

[@starlight_garden]

This video doesn't have closed captions. Only auto.

[@nobodynoone2500]

I kinda miss GPT. I was an early subscriber, but had to cancel it after the results got so bad it became frustrating and worthless.

[@Tarkov.]

Wait a second, those twitter posts.... I had assumed you were 3kliksphilip's twin brother for some reason. Now you've proven to me that you're less real that I thought you were before, lmao.

[@Teal_.]

fun

[@emctwoo]

Course the real solution to a lot of this is decoupling affording to live/eat from working. We wouldn’t care so much about AI replacing our jobs if those jobs didn’t also support our lives.

[@bukeschmenzell]

You should have a look at the music people are making with AI, they’re doing lots of AI covers of songs, insane stuff

[@Xerox0073]

Conclusion at 7:30 sounds AI af, I'm getting paranoid while kliksPhilip is laughing making his videos with AI generated parts

[@StraightGangsta]

Philip is just a very advanced Ai to post videos and is so good in fact,that it can replicate a human behaivor with ease! It can also create a funni bri'ish person to fake a "host"...impressive? Innt?

[@Derpynewb]

Maybe the future is people making personalised data sets for themselves they train their own personal AI tools on?

[@DanielBerthellemy]

ChatGPT wrote thesis of diploma for me lmao xD

[@MrAgLi]

Phillip, have you seen the Austin McConnell video about AI?

[@mikethedriver5673]

I remember looking at one of these Ai detection tools and putting some CHATGPT text in it to test it which it could detect but I very quickly lost interest after discovering that all you need to do is tweak some settings in your AI using the CHATGPT API and it would be completely fooled.

[@SianaGearz]

The burstiest boi.

[@hoebare]

Douglas Hofstadter has entered the chat.

[@PowerPointGaming]

Mind-blowing experiment! 3kliksphilip's creativity and AI's capabilities collide, leaving us questioning reality and the future of content creation