Midjourney v6.1 & Leonardo.AI Acquisition!

Поделиться
HTML-код
  • Опубликовано: 8 сен 2024

Комментарии • 116

  • @net_cap
    @net_cap Месяц назад +25

    I've never seen such an "organic" RUclips storyteller. Have you worked on TV before, or something? It's always interesting to watch regardless of the topic. Thank you very much!

    • @AndyHTu
      @AndyHTu Месяц назад

      He's George RR Martin's son.

  • @High-Tech-Geek
    @High-Tech-Geek Месяц назад +6

    I think Canva, like Runway (and others), is trying to create a full-service, one-stop-shop for creating images. I occasionally use the free Canva to create flyers, illustrations and graphic art type stuff in a pinch. It's pretty good for mocking up stuff in the traditional way with clip art and fonts.

  • @TanChoonHong
    @TanChoonHong Месяц назад +4

    Hi Tim, I'm among those who didn't see this coming - Canva acquiring Leonardo AI. Very unexpected. Hope it'll lead to better things for users. Waiting with both trepidation and hope. Thanks for this alert and update.

  • @zahreel3103
    @zahreel3103 Месяц назад +8

    Great info as usual! Can't go wrong with Keanu under any circumstances

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +5

      If there's one "celeb" I've always wanted to hang with, it's Keanu. Seems like such a chill guy, and I'd love to talk Phillip K Dick with him, while he plays bass, of course! Haha, jam session with Keanu. That's the life goal.

  • @WillowsAIAcademy
    @WillowsAIAcademy Месяц назад +3

    Thanks for the really interesting update. I hadn’t heard about affinity also become part of “the canva family”. As a very keen user of Leonardo’s Canvas editor, which hasn’t been updated at all since it’s released, I’d love to see that updated with some of the more human-lead image editing capabilities we see in affinity. But I’m sure they’d actually be happier with us paying our subscription fees for each of the tools in their suite separately.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +2

      Haha, that's a really good point at the end. Yeah, I think what we'll be seeing is a neutered version of Leo's model in Canva, while the "big toys" will still be in Leo proper.
      But, I do still suspect that Leo will be making an appearance in Affinity. It just makes sense...

  • @giri.goyo_yt
    @giri.goyo_yt Месяц назад +3

    5:31 Hey! It's Pirate Anne Bonny! Thanks for the book recommendation. Looks fun. Insider Neo knowledge. Also: thanks for checking Runway prices.

  • @magejoshplays
    @magejoshplays Месяц назад +2

    This is why i use krita with the stable Diffusion plug-in to replace gen fill, it lets me choose the model and parameters myself for full creative control.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      I'll check it out! Also, really curious about Flux-- gonna play with that this weekend.

  • @zrakonthekrakon494
    @zrakonthekrakon494 Месяц назад +5

    Congrats on 100k been watching since like 20k

  • @Jade8613-ex9kb
    @Jade8613-ex9kb Месяц назад +1

    Great videos - I love your presentation style. re: Midjourney "Tools" option on the web, I don't have this option. Looks like you have special powers.

  • @kadenickel
    @kadenickel Месяц назад +6

    one quick note - for the Midjourney personalization feature, they say you have to rate at least 200 images for it to work, not 2000. great video thanks for the updates!

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +3

      Did they reduce it? I thought it was 2000! Haha, well- that just gets you that much more fine tuned I guess!
      Seriously though, it’s kind of zen rating images!

    • @kadenickel
      @kadenickel Месяц назад +1

      @@TheoreticallyMedia oh totally haha I love it! I have like 10k 😂 - I think it's if you're one of the top 2000 raters you get a fast hour, but for the --p it's a minimum of 200 ratings. but yeah doesn't matter too much I rate for fun xD

  • @OrinThomas
    @OrinThomas Месяц назад +2

    TIm's Bar and Grill definitely has the vibe of the real world Bora Bora Beach Club Bar/restaurant on the south west side of Bora Bora

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      Bora is still one of those places I haven't visited yet! I think this might have just put it on the list!

  • @deborahbelding2413
    @deborahbelding2413 Месяц назад +1

    The statute ala heresy syrup was not Michelangelo's, actually the "David" would be a good choice because it has a worried look, which is how I look when they pour syrup on my head.

  • @BinaryFrameProductions
    @BinaryFrameProductions Месяц назад +4

    Great video as always! Thanks Tim!!! Have a great weekend! 👏🏻👏🏻👏🏻🎥🎬

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +2

      Thank you! You too! (Actually, it looks like an AWESOME weekend, since for the first time this summer: NO PLANS!!)

    • @BinaryFrameProductions
      @BinaryFrameProductions Месяц назад

      @@TheoreticallyMedia Great! Enjoy the NO PLANS weekend to recharge! Sometimes we need those! 😁👍🏻

  • @suzannecarter445
    @suzannecarter445 Месяц назад +1

    I wanted to see if the new, improved Midjourney would accept an image of a dead movie star so I uploaded a high-res photo of Lauren Bacall at the height of her days with Bogie with a prompt that exactly described the reference photo. Midjourney gave me a buxom fifteen year old with big pouty lips. Bacall's ash blond medium length hair was changed to very long dark curly hair and her piercing blue eyes now looked like she was on serious downer drugs. This was the creme de la creme image generator! I feel like I'm going crazy (or it is).

  • @chariots8x230
    @chariots8x230 Месяц назад +4

    4:33 I’d hate to discard a good image just because Midjourney spelled something incorrectly. I wish there was an easy way to fix the spelling mistake, while still keeping the image.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      I know, it kills me. You can Inpaint it out-- and try again, but usually the best bet at that point is to take it into Photoshop, remove the MJ text and just do your own.
      AI is never one-button push, as I always say!

    • @OgoMaduewesi
      @OgoMaduewesi Месяц назад

      Heheheheh, I struggled yesterday and added a 'You', felt a bit relieved 😂

  • @circelink
    @circelink Месяц назад +1

    I'm an IN TO TM Book Club! Also I think Tim's Bar and Grill has no Wifi. Great stuff as always Tim, thank ya!

  • @LouisGedo
    @LouisGedo Месяц назад +4

    😲 Canva purchased Leonardo? That's news to me!
    Fascinating

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +2

      Crazy right?! I did NOT see that coming!

    • @Sinful_Josh
      @Sinful_Josh Месяц назад

      @@TheoreticallyMediacurious if they plan on incorporating it into their Afinity graphic softwares

    • @876rr
      @876rr Месяц назад +1

      @@TheoreticallyMedia Hopefully they dont censor like Midjourney does. Cant use celebrities as character reference

  • @ScientificGentlemen
    @ScientificGentlemen Месяц назад +40

    How is everything "Game changing" and every week "A big week" for AI? Those terms have lost all meaning. Something I've noticed. I like your videos, however.

    • @Dutchruler
      @Dutchruler Месяц назад +6

      That's what they call exponential growth. Think 2 years back when there was nothing, and imagine 2 years ahead when the news now seems child's play

    • @dannyellis971
      @dannyellis971 Месяц назад +5

      You have to have this style of title to show in RUclips algorithm. Otherwise your video doesn’t show up.

    • @Buckdawg
      @Buckdawg Месяц назад +6

      I hear ya. But right now, everything IS 😂 The growth of AI is crazy right now.

    • @sierrawander7108
      @sierrawander7108 Месяц назад +2

      By reporting the most influential improvements/changes to a rapidly advancing technology.

    • @sierrawander7108
      @sierrawander7108 Месяц назад +3

      ​@@Buckdawgyup, it's leaping forward. There are new tools so often it seems hourly.

  • @cdc1138
    @cdc1138 Месяц назад +2

    Thanks for the heads-up about a Keanu book. I'm curious about it as he uses to say he's not really a "man of words". About Affinity, I've just downloaded it. I lean heavily into LR for processing my photos but if they did manage to get a decent RAW editor paired with a PS equivalent, I'll be SO HAPPY to ditch Adobe. And, as you said, Adobe's "AI" is incredibly bad. Let's see how it goes... Tks!

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +2

      Curious as well-- haven't cracked into it until vacation officially starts, but I tend to think he was the "ideas guy" while his co-author was the "words on page" guy.
      As far as Affinity-- yeah, I think (as a the version I used awhile back) they had a Raw editor built into it. In fact, yup: Just double checked, it's in there!

    • @cdc1138
      @cdc1138 Месяц назад +1

      @@TheoreticallyMedia Thanks for checking. I haven't had the time to unpack Affinity as I'm trying to get my Linux dev machine going. About Keanu, yes, I thought the other guy I saw when you quickly showed the book was "The Wordsmith", notorious companion to John Wick! PS - Although I think you kind of a "digital Chaplin", if you ever need a script for the noir movie I've been meaning to get some visuals to what I write. Latest story is about a building in an interesting place in Rio where some black market pharma thing is going on. It also involves a character who's been "cancelled" from all systems. (If Chandler met Gibson and they were both watching Death & Robots in Rio that would come close!!) 😁

  • @LouisGedo
    @LouisGedo Месяц назад +2

    4:48
    The old Describe wasn't great. I hope the new Describe fares much better because it can be a very useful tool. 👍

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      The old Discord Describe I thought was pretty good-- but the newer Keyword Describe? Yeah, I can see that. What I liked about it was that it would link terms and artists to Wiki pages, where you could read up on various styles. It was a nice little "art school" hidden in there.

  • @NoCodeFilmmaker
    @NoCodeFilmmaker Месяц назад +2

    The thing that really sucks about Adobe's generative fill is that even if you use it to SOLELY remove an object from the background, it embeds it with a made by Ai metadata tag. I don't think that's really fair because simply removing an object doesn't suddenly negate an entire project built from scratch without AI assets. I get tagging it if you use actual AI generated elements such as pictures or components. That's really my only gripe with it

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      That's actually a good point. Hadn't thought of that-- It does lead to a really interesting quandary: Is removing something, generating? I don't think it is.
      Every once in awhile, I get to talk to some Adobe folks-- I'll bring that up next time. Might just be one of those overlooked things. From what I've seen, Adobe is a pretty big company and the left hand often doesn't know what the right is doing.

  • @FriscoFatseas
    @FriscoFatseas Месяц назад +3

    I hope This month is hype because i think ill still be waiting for advanced voice for "a couple weeks"

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +3

      The voice thing is KILLLING me-- I want to play with it so much. I wake every morning, but still no love from OpenAI.

    • @Doctor_Sex_0
      @Doctor_Sex_0 Месяц назад

      @@TheoreticallyMedia fr lol

  • @user-ti4tf4kn3o
    @user-ti4tf4kn3o Месяц назад +3

    Canva esp. will eventually kill the market for graphic artists. That company doesn't even try to make the user put some decent effort into getting good results... If any average person can access pro features so easily, why would anyone hire you as a professional? I think the best thing for any artist to do is to actually learn to create his/her own AI network and saturate the market just these companies are doing with their software and level the stress out. This whole scene is crazy! With all that subscription we would end up not owning our tools of trade. Great channel btw...Don't say I didn't tell ya!

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      I think at some point we may see a monetization strategy akin to musicians/spotify, in which you will see a small profit every time an artist's data is used in an output. Maybe, at least?
      That said, as someone who has music on Spotify-- I also know the rev split is NOT going to be in the artists favor.
      Ultimately-- at least in a professional context: most "higher ups" (those that pay the bills) aren't going to be doing design work on their own. They'll still be hiring artists to do that work. Some will be lazy and use Canva and other tools-- but the ones that will have long lasting careers will be utilizing their talents along with tools like Leo (proper), MJ, Photoshop, and others...

    • @user-ti4tf4kn3o
      @user-ti4tf4kn3o Месяц назад +1

      The idea of creating with diminishing tokens in mind goes against my artistic grain. There has got to be a better way lol artistic people come from all different sorts of backgrounds, that includes students. Why provide a service/product with all these features and then make it out of reach to the most artistic sector of folks? there are more creative people on the globe that cannot afford a basic subscription a opposed to the ones who can in this present day world economic situation. Thanks to companies like Synthesizer v, there is a flicker of hope...Knock on wood! ​ @TheoreticallyMedia Thanks for your response... I hope your idea pans out.

  • @gu9838
    @gu9838 Месяц назад +2

    midjourney is great for art and sci fi but i still think it lags a bit behind with people and still feels "ai" i find for people images stable diffusion xl lightning + realvis 4.0 or crea checkpoint really is the best workflow for me but HOPEFULLY 7.0 will add better generation of people plus improved lighting etc and something to look into! flux is not bad but i still find in my test for realism and lighting SDXL with the right checkpoint does better but thats just the base model so im sure in time with user made checkpoints and updates it will only get better! and gen 3 alpha image to video rocks!

  • @StijnHommes
    @StijnHommes Месяц назад +4

    Great! So now we don't have a viable Photoshop alternative anymore and Leonardo will probably become too expensive to use as well...
    It was great while it lasted.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      oh, I don't. know-- the plus side w/ Affinity is you can buy it now, (for like $39!) and you've got it for life. I still have a copy of CS6 on a DVD somewhere! Haha-- I'll never surrender it!

    • @Sinful_Josh
      @Sinful_Josh Месяц назад +1

      @@TheoreticallyMediaplus I think they made the trial software extended for like 6 months

  • @majesticmythology
    @majesticmythology Месяц назад +2

    I still find the leonardo ai upscaler to be the best but it’s heavy on credits. Anyone have any other suggestions for an upscaler? I find best results with Midjourney + Leonardo upscaler + kling for animation

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад

      Check out the Krea upscaler for a 3rd option. The one thing Krea has that I haven't seen anywhere else is a creative video upscaler. So, a bit like Magnific, but video? It's pretty wild. Note: It can go a bit crazy, so you've really got to play with those sliders, but still: A really neat feature.

    • @majesticmythology
      @majesticmythology Месяц назад

      @@TheoreticallyMedia thanks so much!

  • @seanmaynard7417
    @seanmaynard7417 Месяц назад +1

    Great job as always.

  • @PhilAndersonOutside
    @PhilAndersonOutside Месяц назад +4

    It's also been a big week for music and audio AI. Wait, that was last week. ...I can't keep up anymore!

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      Haha, I know! I'm behind on a Suno/Udio update video! Hopefully next week?

  • @TheBlessingReport
    @TheBlessingReport Месяц назад +3

    Go Tim

  • @jymcaballero5748
    @jymcaballero5748 Месяц назад +1

    flux has arrive!!! every other competition is nuked!

  • @JurgenAlan
    @JurgenAlan Месяц назад +2

    Rock star

  • @spacekitt.n
    @spacekitt.n Месяц назад

    i really hope the canva thing doesnt make leonardo come down harder with censorship

  • @LadiesFS2024
    @LadiesFS2024 Месяц назад

    Amazing!!!

  • @TPCDAZ
    @TPCDAZ Месяц назад +1

    0:53 Zark Muckerberg

  • @adrienbiosseduplan9666
    @adrienbiosseduplan9666 Месяц назад

    noting down, "a free drink and a free burger."

  • @ECHSBACHS
    @ECHSBACHS Месяц назад +2

    Cool

  • @sachindatt4045
    @sachindatt4045 Месяц назад +1

    Before runway wakes up to the cost barrier...kling would have already controlled the market. It's dead cheap and simply awesome....runway has no chance to compete with kling's performance and price. I never thought till like a week ago that I will ever buy a AI video subscription as it's a lossy proposition considering the price and that you also pay for the failed output which is ridiculous.
    But here comes KLINg and I instantly subscribed to the PRO version without any argument. It's that good.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      I'm thinking that Kling forced their hand on the price reduction. Those prices were tough to ignore! I picked up the 6 month pro plan-- mostly, since at the speed AI moves, 6 months is a massive amount of time and who knows what we'll be seeing at that point!

  • @Edit9-dj9vc
    @Edit9-dj9vc Месяц назад +1

    And how does one acquire these vouchers for a free beer and burger at Tim’s Beach Bar?

  • @JuneStarling
    @JuneStarling Месяц назад

    Leonardo is great! Has anyone compared it to CGDream? Would love to hear some opinions

  • @kurtisleeTV
    @kurtisleeTV Месяц назад +1

    Runway pricing structure is shameful. I was one of the first users and still am paying $15 for editing 5 videos

  • @user-TheConscientiousnesscow
    @user-TheConscientiousnesscow 24 дня назад

    I need help.
    I'm looking for the best ai video generator. I've tested them all but b4 paying I need advice. I'm getting pushed towards luma but I've tested kling also and haiper.
    I need help

  • @supermario_ai
    @supermario_ai Месяц назад +1

  • @MartinZanichelli
    @MartinZanichelli Месяц назад +2

    Anyway, the non existent Sora is still cheaper than Runway. :)

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      Ha! I heard a rumor that Sora was supposed to release in July...uh, here we are on Aug 2nd...so, I guess not! Geez...those guys.

    • @MartinZanichelli
      @MartinZanichelli Месяц назад

      @@TheoreticallyMedia LOL

  • @MONTY-YTNOM
    @MONTY-YTNOM Месяц назад +2

    Always America though :)

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад

      To be fair, I do shout out a certain Woman's Dutch Footballer quite often! Ha! Fun side note: Did you know if you prompt Man or Woman, followed by an emoji flag of whatever country in Midjourney, you get someone from that country?
      Kinda cool!

    • @MONTY-YTNOM
      @MONTY-YTNOM Месяц назад

      @@TheoreticallyMedia I tried it with Wales where I live, I got a yank :)

  • @noobicorn_gamer
    @noobicorn_gamer Месяц назад +2

    Acquisition lol the winner is Leonardo imo grats to them

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад

      Seriously! Big congrats to the Leo crew! I’ve met a few of them and they’re great people!

  • @brentglittle
    @brentglittle Месяц назад +2

    Leonardo would be great if it wasn't for the wretched, schizophrenic Content Moderation System. Inconsistent nightmare, that for a year they have refused to fix.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад

      Well, hopefully now they've gotten an infusion of cash to solve some of those problems! Fingers crossed!

  • @Wolframandheart
    @Wolframandheart Месяц назад +1

    We've come to the point where we're getting crumbs from the AI industry. I think AI has reached its peak.

  • @Jorsten
    @Jorsten Месяц назад

    Flux.1 pro/dev is insanely good, but it doesn't take in images for references. I don't like Leonardo AI. Its generated images are underwhelming and it burns through tokens like crazy. Ideogram is among the best.

  • @jasonshere
    @jasonshere Месяц назад

    We can't see what you're talking about without having side-by-sides or by showing images right after another, instead of cutting back to you between images.

  • @kensummers7757
    @kensummers7757 Месяц назад

    I love Gen 3, but (!) I figure it’s costing me ten dollars an hour. Too rich for just playing around.

  • @dockaiju
    @dockaiju Месяц назад +1

    Nooooooo!!!! F*** Canvas!

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад

      Haha, not a fan?

    • @dockaiju
      @dockaiju Месяц назад

      @@TheoreticallyMedia I hadn't watched all the way through when I commented. Thought they were pulling the Leonardo site and giving its tools to Canva

  • @collinpotratz750
    @collinpotratz750 Месяц назад +1

    Leonardo has been going downhill for awhile - and just in the last week or two, it's really gone to hell. Really a shame for what *was* my go to free image generator for realism.
    Hopefully Canva can get it back on track.

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад

      I think they've been dialing in the new Phoenix model. I've for sure noticed that as well-- Phoenix is their own "owned" model, so I think we're basically looking at a v.1. Leo has always been rock solid, so I'm happy to give them some time to get their ducks in order.
      Plus, even if you ignore the image gen part: You've still got upscaling, real time, and motion there. They do offer a lot!

    • @collinpotratz750
      @collinpotratz750 Месяц назад

      @@TheoreticallyMedia The biggest issue I have with Leonardo, other than the occasional Cronenberg monstrosity being generated, is the content filter. Everything gets blocked these days...and I mean **EVERYTHING** - The most innocuous/innocent/vanilla prompts get flagged. It's pretty much unusable for character creation at this point. It might still be good for scenery...maybe?

  • @piteshbhanushali1140
    @piteshbhanushali1140 Месяц назад +1

    After watch flux ai..this look like so shitty..

  • @liviere1
    @liviere1 Месяц назад

    Canva goes over Adobe............

  • @holdthetruthhostage
    @holdthetruthhostage Месяц назад +1

    Mid journey is just to lazy man, they are focusing on video, all this ither BS instead of working on making images the best ever, Mid journey is about to be surpassed, they can't even focus on the website

  • @weaponizedtv5228
    @weaponizedtv5228 Месяц назад +1

    Hey I enjoy your content. Do you have a Discord?

    • @TheoreticallyMedia
      @TheoreticallyMedia  Месяц назад +1

      I did, but it kind of went to trash as I didn't have time to moderate it. Working on a new community hub, but building it out slowly so I can get it right next time! I'll make some announcements when its ready!

    • @weaponizedtv5228
      @weaponizedtv5228 Месяц назад

      @@TheoreticallyMedia well if you did have a Discord I was going to ask you what are your current too three image to video generators. Is Pika in your top three and if not why?