ChatGPT: Should Creators be Worried?
Вставка
- Опубліковано 15 тра 2024
- We're back! Marques and Andrew sit down to discuss everything from weird Huawei smartwatch/earbuds leaks to how artists should be credited when an AI trained on their dataset creates art. This week there is a nice mix of weird gadgets, EV stuff, and philosophical debates about the future of technology. Then we wrap it up with a recap of all of the hot takes video that Marques posted last week before going over trivia answers. It's a classic episode of Waveform!
Chapters:
00:00 Intro
02:29 Huawei smartwatch/earbuds
10:09 Manual electric cars?
19:07 Trivia question 1
20:04 Canva (Sponsored)
21:22 ChatGPT and AI art conversation
29:59 Andrew reads a story about MKBHD Santa
32:29 AI Avatars being made and creator credit
53:00 Trivia question 2
53:21 ZocDoc (Sponsored)
54:06 Capital One (Sponsored)
54:33 Grading Hot Takes
01:11:05 Trivia answers
01:15:02 Outro
Shop products mentioned:
Beats Fit Pro at geni.us/nFLIb0
Apple AirPods Pro (Gen 2) at geni.us/B61pMn
Apple Watch Ultra at geni.us/hqmrUx
Samsung Galaxy Watch4 at geni.us/b2y5
Garmin Epix Gen 2 Smartwatch at geni.us/HMDaL
Links:
Verge Huawei leak: bit.ly/vergesmartwatchleak
Marques Hot Takes video: bit.ly/mkbhdhottakes
Shop the merch:
shop.mkbhd.com
Twitters:
Waveform: / wvfrm
Marques: / mkbhd
Andrew: / andymanganelli
Adam: / adamlukas17
Ellis: / ellisrovin
Instagram:
/ wvfrmpodcast
Join the Discord:
/ discord
Music by 20syl:
bit.ly/2S53xlC
Music licensed from Epidemic Sound:
www.epidemicsound.com/campaig...
Waveform is part of the Vox Media Podcast Network. - Наука та технологія
I can now say I was here for the first explicit WVFRM podcast. It’s a great feeling.
Indeed
Whats explicit in this podcast?
15:14
@@abheekmalhotra 😂😭😭😭
I don’t know why Andrew dropping the f bomb was so strange to me
No way haha, time stamp?
@@prathavjuloori7876 15:21
It is probably because he is a perfect, soft-spoken angelic personification of all things innocent and good.
@@galekeoma7276 that's literally what they think priests are.. that's more weird actually
Then you should watch the FLAGRANT podcast with Marques because he curses on there too. It makes them more relatable. They’re human lol
Don’t know if the MKBHD team noticed but a little before 17:09 mark there’s an unbleeped f-bomb.
there goes the monetization for the video!
The MKBHD Santa story was way better than I expected, but AI (based on the story and the images I've seen from AI Art) will by design be completely derivative. To me, this raises a lot of intellectual property issues as artists who's images are used to "feed" the AI programs receive no compensation, or acknowledgment, and more importantly their copyright is infringed. I think what would be interesting to see/test, is whether an AI program would be able to generate a story using prompts from popular movies and tv shows without plagiarising. So, if given a prompt to create a story about a team that plans a heist inside a dream within a dream - would it be so close to Inception that it plagiarized the script?
I have a daily driver SRT10 Viper, and an ACR, both in manual I am also a SCCA driver. There is no, no chance I can out drive a comparable car with a electrical mechanical pdk (style) 10 speed. Just stop, everyone. Marquee nailed it, you can’t put move the computer.
I have loved my garmin epix gen2 so far. I'll leave for a weekend snowboard trip with 85% and use the GPS and activity tracking for 2 full days on the slopes and track sleep and come back with 25%. Very usable and the garmin connect data is really easy to use and look at
I feel like EVs should be able to have a ‘sim’ mode where you can learn how to drive manual without the repercussions of missing a proper shift. I never learned and would love to, always have, but also not trying to destroy any engines learning 😂😅
You’re more likely to wreck a transmission than an engine. Also yes.
Just get a rental
I want to be able to flash different car profiles to my EV, I’d love to for an hour have the simulated drive train/experience of a GT40 or other car legacy car I’d never get a chance to drive. Like for fun in a parking lot or something trundling along like your car was a Model T, and maxing out the 20 horse powers or however much they made.
Also a third pedal, that just enables 100% coast and cancels all regenerative braking
As a manual daily driver, the chances of you wrecking a transmission are pretty low. If you have someone teaching you, it’s much more mental than actually difficult mechanically. Highly recommend learning, the connection you feel with the mechanics of a car are greatly appreciated when you’re changing the gears yourself!
I also used AvatarAI (also using many wedding photos resulting in suit pictures), and the driver photos turned out to look like a copy-paste of one of my images as well. (same driving suit and all)
I just want to say I Respect you both for seeing this with AI and agree with you.Where do you draw the line.Some of it was nice but you need to know what you really experience in life.And what you don’t.Love the Podcast guy’s and so Awesome episode.Deb ✌️
Adam forgot to censor 15:22 LOL
15:22 lmao, I think this is the first time I heard someone curse on any MKBHD channel ever 😄
For AI there's the difference between legality and ethics. There's also monetization vs research.
My personal issue with AI is that artists are given no voice or option to not have their works excluded from AI learning, in fact in some cases, people have purposely fed their art against their wishes into these databases. If this was copyright free items and works that are voluntarily submitted or paid to have submitted, I think the conversation would be vastly different.
The conversation artists are currently having are:
1) having their works submitted for AI learning without their consent (and in many cases against their specific wishes)
2) the inability for AI to unlearn something that it shouldn't have learned/fed
3) corporations using data ostensibly gathered under 'research' and 'nonprofit' fair use and then commercializing it (Stable Diffusion and Laion - one case of a corporation using a loophole of taking nonprofit research to circumvent copyright (and make a lot of money from it). Laion itself has taken from numerous sources that normally wouldn't be available for corporate entities.
4) The argument that Machine learning =/= inspiration (from both artists and AI experts - some talk on plagiarism and overfitting)
5) AI art databases vs what AI takes from other genres (e.g. music and dance which has large corporate interest already in place in which AI has limited only to open source/copyright free data)
6) AI 'Artists' (not artists who are using it as a tool to enhance their works but people plugging in phrases and otherwise not a part of the creative process) are already taking these pieces, claiming them as their own and advertising themselves as regular artists (without informing that they are using pure AI with maybe a sharpening or colour adjustment filters)
7) AI 'Artists' stealing works in progress (from streams and such) and 'finishing' the pieces with AI, posting and claiming the art as theirs and then bullying the original artist to claim refence work because they posted it earlier (before the artist could finish)
Here are a bunch of links to helpful information that may shed light on why this is such a difficult topic for most:
www.kortizblog.com/blog/why-ai-models-are-not-inspired-like-humans?fbclid=IwAR27VVogdwWYRySGKMAeBPjr6Q0Ku7_o7e21p-MjYnQUGsgUYOVAfbqSXI4
twitter.com/svltart/status/1592220369599045633
www.upworthy.com/lensa-app-ai-profiles-privacy-ethics
twitter.com/PeachiieShop/status/1599858286307987456?s=20&t=CQRUdXCrNOQtiN0IYp5yow
twitter.com/kortizart/status/1593723972314882048
kotaku.com/genshin-impact-fanart-ai-generated-stolen-twitch-1849655704
Here's a link to a Town Hall with Artists, US Copyright office and an AI Ethics expert discussing AI Art. It's long but goes to show how complex things like this are since style and artist names are not by and large considered copyright-able.
ua-cam.com/video/7u1CeiSHqwY/v-deo.html
Already have a lot of links here...but I have a bunch more lol, won't post them here since you can find a lot more out there (e.g. look up Samdoesart, Kim Jung Gi who died and had his art uploaded the day after his death, Greg Rutkowsk and many more)
Wonderful. I would also add that if it's true how they are "learning" this stuff it is violation of human rights. Their defence is actually making it worse. When they "train" ai to copy and then generate "new art" they are appropriating humans without their consent to steal their talent- it can be ability to sing so they mimic your voice, ability to create different kind of strokes with your hand etc.- that is all part of a individual talent like a fingertips. Talent is a part of individual and makes part of their identity. If you are creating robotis to first copy and then futher generate products by unauthorized talent appropriation you are violation someones identity and you are outsourcing their work from them self and therefore we have violation of human right to work and identity theft. I know it's a bit philosophical but it does go under ilegal activities in humans rights category. This has opened a new sets of violations in many forms, both moral and legal.
@@leogirl6695 You can whine and moan all you want but the truth is people complaining like you are nobody's, you don't get to decide where the future is headed for better or for worse. So stop acting all high and mighty with all those buzzwords you've learned from people equally annoying as you.
@@typeflawed9130 your name says everything.
As an artist, These AI Art Programs scare me a lot!
But I'm pretty sure the human generated art would become more valuable since everyone can make Art with any of those AIs out there, and so would be considered less valuable... especially because there's less creative control and the fact that you can't always get exactly what you imagined when making the prompt.
To quote on one of the UA-cam channels which usually create videos about AI stuff: let's see this two papers down the line. This area is still in heavy development, and some of the tools are already producing crazy good results. If development doesn't stop suddenly caused by some kind of limit, I think the future will be interesting. But I feel you, for me GPT-3 is scary as a programmer :D
We’ll have to more and more prove that your drawing was made by you and I mean at some point a.I. won’t just generate but could draw from start to finish and also seem legit
I NEED YOUR PERSPECTIVE KOBBY. i made a video that should hopefully help you relax on why chatgpt wont take your job away. its on my channel, let me know what you think!
I resonate with what you said Andrew, I don't feel anything knowing that this art was made my a "machine". It's really good art but I don't emotionally connect to it in the same way.
would love to hear your perspective on my video where i argue why ai wont take people's jobs away. i think we may agree on a lot of things
About the shifting gears: F1 still has it along with most Pro motorsports around the world because of midrace changes.
-When the safety car comes out you need to be able to manually slow down to avoid a wreck and also get back up to racing speeds. Automatic does not allow for dynamic changes like that in speed.
Also for pit stop which happens multiple times in races. You need to reduce your speed from race speed (while avoiding race traffic) and enter a pit lane, which also has its own speed limit. Then stop in your pit. Then you have to do the whole thing over again to join the race live. Automatic can not account for all the changes happening in front of you in real time so manual is still needed
Loving the cohesion of your sweaters and the branding.
Hey hey, love your work! I think your wide camera needs to be levelled a touch. I know a bit pedantic, but as a photographer horizons are my thing.
Have a good holiday season.
You guys definitely should consider Matt Farah, or some other car journalist as a guest to ask them about EVs. Could make for a really interesting episode. (I know yall had one with Doug and props for that I really liked that episode)
I am the one person who wants an EV with a manual. Heel-toe downshifts for life.
Guys, you should be talking with people like Karla Ortiz or Steven Zapata when talking about not only Art AI but also the ethical and law side of things. As these artists have done research or are doing something about this whole issue.
You should just ask the AI chat bot to write the pro and con arguments for it's own use of source material in relation to giving credit to the creators of the source material.
oh course now it will reference this podcast ;)
I wonder if you could ask it to site sources :)
Man, Andrews hair is like PEAK right now, not too short or long. Perfectly medium with that cute morning messy look
you might not like my long ass messy hair in my videos then xD i talk about chatgpt too let me know what you think!
@@cutieapplepie xD worth a try :3
crypto hot take is something i have always agreed with, thank you Marques for so eloquently explaining why !!!
I agree with Marques on the idea of putting two devices on one thing you have to pick up. I feel like a less unwieldy way of putting your headphones on your watch would be to put the headphone case like 90 degrees around the band with some sort of connection to share power, so it faces up when your hands are at a natural position, and some sort of latch so it definitely stays closed. If the case is small enough and rounded it might not be too distracting
Thats awesome to hear you guys talk about F1 : D hope you guys bring it up again some time in the future!
I love the talk about AI and inspiration and all that. It reminds me of the movie iRobot and when Will Smith's character asked the robot Sunny if a robot (AI) can write a symphony or create a work of art. And then the part where the robot draws the picture of the bridge.
My question to you guys would be; do you think AI will eventually get to a point where it doesn't need to pull from the internet at all? Will AI be able to create originals not found anywhere else?
I'm pretty certain that we will NEVER have ""AI"" that can create "works of art" without having "references". How ever to "create (synthesise) new art", sure, i.m.h.o. ""AI"" has more or less already reached that state. Though that (as always) depends on where You "draw the line" for something being considered "new"/unique (not to mention being "art").
Regarding my certainty that ""AI"" always (and forever) will need "references" in order to ""create art"". I base this on my belief that in order for the "created art" to "make sense"to the "human interpreter"(audience. The "art" HAS TO HAVE cross references with the "audience", and these "references/parameters" for that HAS TO to be "embedded into the AI" either "hard coded" or by training before it will be able to then ""create meaningful/interpretable art""...
AND I would argue that exactly the same same goes for humans!! If You were to take a human baby and somehow manage to extract all the innate inherited biological references, from it, and then managed to raise it without in any way imbuing it with any "references". Then that "human" would also be unable to create anything that would constitute "MEANINGFUL ART" to a normal average human...
I believe that what ever they would create would be the equivalence of "white noise". And what ever meaning/value the "normal human audience" found in that "art", would be what ever that they themselves contributed and created, basically from what they (the audience) "brought to the table".
Best regards.
@@onlyeyeno I like this take on it but I must say you may have confused my saying; "doesn't need to pull from the internet." As referencing things. Of course AI will need inspiration from something just like, as you stated, humans do too. I simply am curious if we will get to a point where a program or code can be written that is all about learning\info-gathering but from surroundings. From the audio the AI takes in. The images "seen" and behaviors witnessed. And without being connected to the internet in any way, could AI learn like that?
But I do share your view. Thanks for the input. This is very interesting and we live in such an amazing time when it comes to technology!
Guys what an amazing podcast! Congratulations!
Just wanna add regarding how specific ChatGPT can get when you prompt it to make a script for a specific UA-camr.
I've asked it for Vsauce video from Micheal Stevens regarding a certain topic, and it literally did what generally Micheal does in Vsauce videos, start with a weird problem and expand it to the place where your brain explodes and you are left with questioning life.... So yeah it does more than just add a UA-camr like intro and outro 😅
As for the signatures found in AI generated art, the AI has seen a LOT of images it knows to be portraits that have signatures on them. So to the AI, portraits often include a little squiggle on them and it adds that in. It's not just trying to recreate specific artists' signatures.
If we were talking about dall-e, I’d agree. But they were talking about the Avatar AI that seems to be literally splicing images together and leaving signatures in there, basically like if a human photoshopped 2 images together, rather than creating a digital drawing from scratch while looking at reference material. I’d argue that’s far less impressive and is most likely less morally justifiable than dall-e and other diffusion-based methods
@@jmunt All of these apps are just using Stable Diffusion. It works the same as Dall-e
I hope this pod keeps going forever like jre 🤌
Adam asked a good question about AI being limiting in regards to depending on them to create something. We know that these kinds of tools in both AI and machine learning have problems with getting stuck in local minima. They try to find the "best path" but because there are inherent limitations in the training sets, it could only ever find this best path within the scope of its training data. It's very possible that outside of their scope are tons of minima, or best paths, that are much better. Humans also have this issue - we have limited knowledge and experiences of many things. Theoretically these new tools have a bigger expanse to report from, the cool part will be how well the AI can reach out to the fringes of that expanse to return something really cool and unique.
great points Daniel! would love to hear what you think about the video i made on the topic of why chatgpt wont take people's jobs. it's on my channel, would love some candid feedback from you!
Transition at 34:40 was so pleasantly done
AI art scrapes images under copyright and these AI art companies are profiting off artists’ life work without their consent. Some creators have even found what look like artists’ signatures inside images generated on Lensa. How is this ethical or legal?
I agree. While one can use open source and royalty free are prompts, any current artist can be ripped off, copied and never receive a penny of profit.
It is no different than a human taking inspiration from artists and then making their own art. This is the future like it or not
@@djjean85 that's not how it works. using art as a reference for your own art and trying to imitate art through a computer bot are 2 COMPLETLY different things
@@djjean85 i’m pretty sure copying someone’s style and adding their signature to it is not “inspiration”
@@viruscat385 can you definitively say that we aren't living in a simulation and our humanness isn't anything more than a row of computers somewhere? The issue isn't that artists are being "stolen" from it's that we are commodifying something that doesn't need to be commodified. Capitalism forces the concept of ownership for the sake of profit. It's bad that artists may lose money as a result of new technology but they are hardly the only ones who have encountered this problem. Adapt or evolve.
This might've been the best podcast episode yet. The producercam is a great addition - more Ellis and Adam (and David)!
a case of AI Art use i seen that was amazing and went viral was Shaun Harrison's use of Midjourney, After effects and Photoshop to make concept art for "Bashenga: The first Black Panther" it looked so cinematic and epic everyone was fooled thought it was really produced by Marvel and would stream on Disney+ 🤣🤣
Glad to see you back!
I am definitely in the small group of manual drivers that love ev. Like, I'm getting an ev as soon as it's affordable to have one that is as reliable as my car is, but I will also miss driving stick. I think it is much more fun, even at normal speeds. But, I also use it to save gas, which I won't need to even think about with an ev
Manual EVs will probably be popular in Europe for driving lesson cars
I love driving manual! It feels like you are more directly in contact with the engine and the road, and it's durable and cheap to maintain. Still, even though it makes sense in a gas car, it seems kind of pointless in an ev.
I just wanna say two things:
1. ChatGPT does not have access to internet. Everything it "knows" is contained within the trained machine learning model. So it doesn't have access to, say, a Wikipedia article. It may still output something very similar to a Wikipedia article, since it was probably trained on all of Wikipedia, among other things.
2. Dall-E doesn't have a database either, its a generative machine learning model trained on tons of images. The model might again generate very similar images to something it was trained on, but it doesn't have access to any of the images it was trained on.
The MKBHD Chevron hoodie looks really dope🔥🔥
I wish there was a way to get it in India without the high shipping charges😐
Beleive me in under 2 years, you can literally use ai to do that for you (if we are talking just about the print here, but who knows, maybe we will have ai in manufacturing of textile tooo, who knows!)
Interesting! good discussion. ✌️
Lmao is this the first F-word on the podcast? I love it!
Andrew dropping the F bomb at 15:13 is not the Christmas present I was expecting this holiday season family friendly programming be damned
I love the idea. How many times have we adapted to technology advancements. If apple did the same thing, this conversation would be different.
I loved this watch concept. May not be perfect at this time but it solves a lot of problems
On the subject of stick shifting. In Europe, most people drive with a stick shift and many don't actually have any experience using an automatic transmission. So even so I do consider having a manual shifting experience a bit odd on an electric car, I do think that there must be a reason they're building this.
It comes down to how much you trust a CVT or an automatic. Manuals are more satisfying, gives you more control over the car, and manual gearboxes can still handle way more power than any other gearbox built, which is why semi trucks still use manual transmissions.
You know, talking about the Huawei Earbuds/Smartwatch, it feels like the same discussion when smartphone companies started to put a stylus in a smartphone which now is in many phones. Maybe this really is the future of smartwatches/earbuds?
31:39 maybe I missed it, but how much time does it take before the AI kicks the story back to you? Does it take a few minutes of processing, only a few seconds?… Two or three hours? How instantaneous is it to create a full story from just a few prompts?
For a short story like that it varies a bit based on how many people are using chatgpt, but from my experience it might take around a minute.
I agree with Andrew's take on gears, it's just another mechanism of fine tuning control, engine braking is a thing even with CVT fake shifters, I have a triptronic gear box also and I'd often switch to manual since its faster to let the ECU know that I am about to floor it instead of pressing the accelerator harder.
Now I do it to get out of traffic or if someone blocks me and I need to cut them sometimes, no racing lol.
But I can imagine actual enthusiasts finding this very cool as it's just another knob to play with to get the most optimal performance out of your vehicle
People be playing forza and hrna Turismo with 10K usd kits to get racing experience, I don't see this as much different.
The physical control always feels nice even if below the hood everything is electrical, this is common in other industries also where products still like a feedback mechanism.
Another example I can give is the electric steering wheel and electric brakes, we all know its all connected by wires now, but there is still physical feedback for you to fine tune your adjustments when steering or braking - so instead of being full throttle, you can be full throttle to a certain point in a curve while no limits on a straight, kinda limiting the power to match your driving behavior.
Hope this makes sense lol
Not a car expert
But just 2 cents
No censorship on 15:23? I respect it😄
It would be cool to pay an actual artist to illustrate the MKBHD Christmas story! It would make a cute comic.
There should be a survey of people who go to casinos and if they own crypto as well, and other avenues that would be consider "gambling" and see if there's a correlation
I hate that I have watched too much Big Bang Theory, so I knew the answer to the Higgs-Boson question😅
you guys are getting better and better at these podcasts.
Agree that the smartwatch with earbuds in it, is soo useful. Something new as Marques said💯⌚️
Now, if the new battery technology in future makes the battery so slim then. Maintaining some thickness to fit in maybe foldable earbuds in the watch, wouldn’t be that bad, right? 😋
on the AI front a youtuber called Jazza (he's an art youtuber) just posted a really interesting video on the subject that's worth a watch
I would totally get a full electric car with a manual gearbox, the torque and acceleration of electric are some things I want but I love to drive so much and I drive manual everyday for the past 13 years and every time I drive an automatic car, it’s so boring and I hate it.
Yes most people just have a car to go from point A to point B or as a tool. But for me it must be fun to drive because I really like it and I’m sure I’m not the only one like this.
I don’t like the argument of just get a gas car, we want to move away from gas car and we have to find ways to convert people like this one way or another and if those feature are it, then I think it’s a great thing. I really hope this will become a thing, I don’t care if it reduce the efficiency, I just want to keep the fun of driving while doing my part and switching to electric sooner than later.
I agree 100%. I feel like a key demographic was missed. How about the people that like driving a manual car because it's WAY FUNNER, but don't want to have to spend a ton of money on gas every week?
If you didnt know LuLu AR is a company thats setup to make avatars for people. They did the show Alter Ego
5G is bad in NYC as well. I constantly have connection issues now. Once it’s connected and I’m not moving great. But transitioning from my apartment to my car I lose connection for 30 seconds to 1 minute.
Before you come for me this is just one issue. Happens more often than just that.
already listened on spotify but im here to see the AI pics of Andrew 🤣
TURN AROUND
@@Manganellia1 NEVER 🤩
Love the content, just wanted to help correct GPT, rather than GTP @Waveform
huge difference between sacrificing speed on an off road vehicle, and sacrificing basic build quality on a high-end vehicle 😅
I got goosebumps listening to the AI generated MKBHD saves Christmas story. Amazing!
55:25 "this is a cold take" hahahaha
While Teslas have a single speed gear box attached to the drive motor, don't the dual motor variants have different gear ratios for the front and rear?
That is one dope ass Christmas sweater
15:23 Whoops 😅
Manual in ev makes sense for some specific senario
Thought about it if you want to go 100mph lets say motor is running at 10k rpm consume more power and now if we add a transmission we can go 100mph at 6k to 7k rpm saves battery
31:08
If it said, "built-in RED camera" that would have been terrifying.
14:12 13:45 15:33 16:43 36:25 36:35 38:11 44:30 44:47 45:12 I like motorsport talk
41:45 Maybe it's still using stable diffusion but thinks the signature is part of a certain thing as it can't differentiate that the signature isn't part of the piece of art itself. Kinda like to the outro of UA-cam videos (44:05).
47:48 That was a big point in the main channel video
50:36 I don't know why Marques could not tell Brandon directly. He also understands English & Marques can also guide Brandon during the process to what he wants. I think this technology is good for inspiration, but not to show people your ideas as humans can also imagine something from what you tell them.
55:23 That's what I said
In F1 automatic shifting is banned, its not that drivers are better than the machine its just prohibited under the formula. This has actually been the case with active suspension, traction control and automatic shifting since the nineties when electronics came into the cars in a big way. Even the invention of the paddle shift by Ferrari in 1989 was almost banned.
You should try the huawei watch for once to know the battery, they are great.
I asked CHATGPT what was the most prominent difference within MKBHD and LTT:
There are a number of differences between Marques Brownlee (MKBHD) and Linus Sebastian, as they are two distinct individuals with different backgrounds, styles, and approaches to creating content about technology. Here are a few of the most prominent differences between these two UA-camrs:
Content focus: While both Brownlee and Sebastian create content about technology, they often focus on different aspects of the tech world. Brownlee tends to focus more on product reviews, unboxings, and tech news, while Sebastian's content often covers a wider range of topics, including computer hardware, software, and gaming.
Style: Brownlee and Sebastian have distinct styles when it comes to creating content. Brownlee is known for his concise and informative approach, while Sebastian has a more laid-back and humorous style.
Background: Brownlee and Sebastian have different backgrounds and experiences in the tech industry. Brownlee is a self-taught tech expert and journalist, while Sebastian has a background in computer science and engineering.
Following: Both Brownlee and Sebastian have large followings on UA-cam, but Sebastian's channel has a slightly larger subscriber base and overall viewership.
Overall, these are just a few of the differences between these two UA-camrs, and there are likely many more. While they are both known for creating high-quality content about technology, they have their own unique approaches and styles that set them apart.
Watching the Croatia vs. Brazil World Cup game rn. Coming back later
It's half time so just Tuned in lol
@@chhitijpahari1011same 😂
What a great game to watch
The penalties are going to be crazy 🤪
That was an amazing game!!
I think some of us are curious about the gym ones Andrew
That whole discussion about ownership of AI generated pictures reminded me of Aaron Swartz. Had he been alive I wonder what his opinion on this would have been!?
I don’t know how many times I have lost my AirPod cases so I think that watch case way cool
I believe huawei has a smartwatch with 14 days of battery as well, so they could pull the earbuds + watch thing off
I only get audio on my right speaker with surround settings
My fear with AI art and stories is that, If the only art moving forward is interpolations of previously created art. Then creativity will die.
I mean that’s all creativity is, it might feel a bit less human though.
this is a very interesting and insightful perspective chad. would love to hear what you think of my video on the topic
LMAO I WANNA SEE THE GYM PICS
missed opportunity - Creative AI: Should I be worried?
I’ve lived in Calgary and Toronto in Canada and the 5G giving me 280-~500 Mb/s instead of 100 or 50 is definitely handy, and worth having I would say.
There is no 'database'.
LensaAI, uses stable diffusion. They fine tune the model with the images you provide.
The file they use, is 2gb. It does not contain 'links' to anything. It contains learnt weights that allow it to understand the relationship between images and text.
The 'watermarks' you see, are not watermarks of artists, its watermarks of the AI.
The AI has simply learnt, some art has watermarks, and thus tries to make its own.
15:20 missed the censor?
3:38 The product market fit for this device is probably runners and athletes.
This watch feels like that Ben10 thing
I LOVE my EV but I miss shifting gears 🤷🏾♂️
Huawei smartwatches actually have a one week battery. The size would be ridiculous but, I think the battery would last enough 🔋
They are worth testing even just for comparison.
I personally like EV’s, but I love manual. Seen this manual EV seems cool to me, but he’s right, in this economy I would never afford it ):
If Garmin did a Collab with Harmon. It would probably be AKG or some kind of galaxy buds in there.
Finally after 2 weeks.
I wanna know everyone’s end game moment for what they believe peak AI will be like. I personally think AI will get to the peak point of something in the vein of Altered Carbon. Downloaded consciousness into synthetic bodies.
Efficiency in mind would gearboxes help electric motorcycles finally get rolling? Or are they just doomed?
I think these models tend to have a problem with hands because they’re so expressive jsut as a a Face is. Not to mention the many degrees of usability the ai now can’t see the hands being in because the programmer didn’t give as many anchor points to hand recognition than to face recognition.
15:20 WOAHHHHH WOAHHH WOAHHHH WOAHHHHHHHHHHH
10:12 Is there battery tech right now that does not heavily rely on mining earth's rare metals?
Imagine MagSafe power bank housing the AirPods!
F1s and f-bombs. All this and more on todays episode of the waveform podcast.