# MidJourney v5 OpenAI Supercharged DALL E 2! Is it Better than Midjourney V4? Let’s Find Out!

## Introduction

## What are Leads?

### Organic Leads vs Paid Leads

## Why are Leads Important?

## How to Generate Organic Leads for Your Business

### Optimizing Your Website for Search Engines

#### Keyword Research

#### On-Page SEO

#### Off-Page SEO

### Creating Engaging Content

#### Understanding Your Audience

#### Using Visuals

#### Storytelling

### Using Social Media to Generate Leads

#### Identifying the Right Channels

#### Creating Shareable Content

#### Engaging with Your Audience

### Building a List of Email Subscribers

#### Creating Lead Magnets

#### Using Opt-In Forms

#### Creating Email Campaigns

## Conclusion

## FAQs

1. What is the difference between organic and paid leads?
2. How do I optimize my website for search engines?
3. What kind of content should I create to engage my audience?
4. How can I effectively use social media to generate leads?
5. How do I build a list of email subscribers?

br>In this video, we’ll be checking out the Dall-E 2.5 update! What’s new? What’s improved? Let’s find out! We’re also going to see whether or not the Dall-E 2.5 update is actually better than the Midjourney V4.

▼ Link(s) From Today’s Video:

✩ DALL-E 2:

✩ Midjourney V4:

✩ Openaidalle Instagram:

► MattVidPro Website:

► MattVidPro Discord:

► Follow Me on Twitter:

————————————————-

▼ Extra Links of Interest:

✩ AI LINKS MASTER LIST:

✩ Lemon Gallery:

✩ General AI Playlist:

✩ AI I use to edit videos:

✩ Second Channel:

————————————————-

Thanks for watching Matt Video Productions! I make all sorts of videos here on Youtube! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!

All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.

————————————————-

► Business Contact: [email protected]

Este vídeo foi indexado através do Youtube link da fonte
MidJourney v5 ,

dall e 2,dall e 2 update,ai,dall e 2.5,dall e 2 demo,dall e mini,dall e 2 tutorial,dall e 2 openai,dall e 3,ai art,dall-e 2,dalle 2,dall-e 2 ai,openai dall-e,openai,dall e ai,text to image ai,dall e,openai dalle-2,dall-e 2 explained,dall-e 2 openai,use dall e 2 image generator,dall-e 2 demo,dall e 2 try,ai art generator,dall e text to image,dall-e 2 access,openai dalle,text to image,dall-e 2 tutorial,open ai,dall e demo ,

https://www.youtubepp.com/watch?v=sCGKtIeKqPg ,

Honestly viewers this is a video that I didn’t think I would ever make but I’m really glad that I’m making it you see when Dolly 2 first hit the scene almost a year ago now it was seen as complete magic it’s still a pretty decent AI art image generator and its results can

Still be absolutely fantastic and at the time it was mind-blowing but almost a year here later mid-journey V4 is absolutely wiping the floor with Dolly 2’s results and quite honestly stable diffusion’s versatility is seen as much higher value than Dolly 2’s limited features and ability to tweak Generations however Dolly 2’s editor

Without painting and in painting still is fantastic and it still can generate really nice images as you can see from these examples but mostly everyone uses mid Journey now for that added extra creativity the colors pop more it’s really vibrant and interesting to be honest the pricing model in mid Journey

V4 is a launch cheaper than Dolly and it’s a little less egregious it’s like a subscription rather than Credits based however there’s a new Dolly model in the works this is very very exciting I heard rumors about a new Dolly model months and months ago even emad the creator of

Stable diffusion was tweeting about a new Dolly 3 Model so somehow he was able to predict that they were working on an upgrade to Dolly too the reason I’m showing you this rather hilarious image is because it actually highlights a lot of what Dolly 2 originally was missing

The characters were trash in Dolly too so if we go into original Dolly 2 here and do a very similar prompt to what this is again it definitely looks kind of like Sonic the Hedgehog but it’s not very coherent at all the creators of Dolly openai made sure to try to get all

Of the famous characters out of the database as much as possible so while this definitely resembles Sonic it’s a little off-brand and weird feeling this Sonic feels a lot lot more on par with what you’re used to be seeing from the character and also you’ll notice in

Comparison to a lot of these generations with which look a little bit sloppy they’re not as detailed this one is is very high detail high resolution the burgers are very coherent Sonics very coherent he’s actually rather chubby and large because he’s been eating so many cheeseburgers and it just looks like a

Dolly three rather than a dolly too to be honest although people are calling this dolly 2.5 now if we head over to the dolly to Instagram you can see here is a post the first post that I can find that they made about this new experimental Dolly 2 model I’m thrilled

To share with everyone some work I’ve been generating with the at open AI Dolly Next Generation experimental models these are still in the testing phase and openai is using works like this to push and test the model in each and every way before launching worldwide so we will be seeing this model

Worldwide confirmed the goal is to have artists like like myself offer Early Access feedback reaction suggestions and any concerns you know for this experimental model so yeah it’s it’s on lockdown right now it’s it’s like this very mysterious email that’s being pushed out people actually in my Discord

Server have been getting access to it so far which is very exciting the email looks a little something like this hello from the doll E2 team you were an early contributor and you gave us good feedback on the original Dolly model so it seems like if you gave good feedback

And useful feedback early on in the dolly model you are granted access to this new experimental model and you get unlimited credits and they’re only running it for about 10 days while they collect some feedback on it and yes the person who did receive this email in my

Discord server was the one who created this Sonic Image anyways here is our first generation here in the example of the new model this one I think is a little bit more telling than the Sonic example we first looked at again pay attention guys to the amount of detail

That you can see in this image it is very very detailed more so than you get with typical Dolly too it looks a lot more realistic the Grain and the texture I think looks a lot better on the spacesuit it still reminisces of the dolly model which is very interesting it

Just seems like the dolly we all hoped for in the beginning and then we have up here some text at the top that I think looks more coherent than text and numbers that you would typically see from Dolly this looks more like the kind of text that you would see from

Mid-journey for example however there actually is an AI art model out there known as deep Floyd that does text and can actually completely write sentences I did a whole video about it before you can see here are some results on Twitter from the deep Floyd if model that does

Text to image and it can actually completely spell and I’ve tried this out and I have access to this model and I gotta say it’s pretty pretty awesome so that’s one thing to note here they don’t seem to have Dolly spelling out sentences or anything like that just yet

And I think that’s actually very very very crucial for the model but it’s getting better than it was and I think it’s on par with mid Journey as far as I could tell in terms of spelling definitely quite a lot more detail than we typically get from Dolly too I will

Say that here’s a rather terrifying image from this new Dolly model here honestly reminds me a lot more of mid-journey in some ways especially with this image here and the details and stuff you can see but yeah it’s a gross image it’s very coherent though and I

Think they’re doing a pretty good job I do want to compare this model to Mid Journey we just need prompted images so this prompt was Basset Hound portrait 8K as Roman tile mosaic pretty good results here from the new Dolly model let’s take a look at the dolly model we have access

To right now and the current Dolly model clearly leaves a lot to be desired these results are really poor honestly I don’t think any of these really could be considered usable so yeah the new model definitely is improving quite a lot on current Dolly but that’s not a shock is

It Improv proving on what mid-journey can do right now I gotta say guys this is a little tricky because these mid-journey results for the same exact prompt look pretty darn good especially this one down in the corner I think it reminds me a lot actually of that

Original image so again here’s the new Dolly here’s what Dolly currently generates and here is what we got out of mid-journey so I think that this new Dolly update is sort of putting it almost on par with mid-journey although I I gotta say mid journey is looking very creative here especially with this

One this one’s really really nice here’s some more results with the experimental model again don’t this does this kind of like actually remind you guys of mid Journey a little bit I feel like there’s a lot more color and Detail in these results and like the way that the shapes

Are formed they remind me more of mid Journey pretty cool results nonetheless definitely more detail I really really liked this image that was generated with the new model of course we have a prompt for this one as well so we’re going to be doing a comparison but yeah very very

Detailed image here you can see the texture in the skin of the goldfish is absolutely fantastic it’s a realistic photo but it’s still artistic in some way the color is really nice and yeah the whole thing is just very coherent I have a feeling that dolly is not going

To do great with this prompt so let’s see what current Dolly does and honestly I gotta say current Dolly did better than I anticipated but definitely all of these are less detailed and they’re just less coherent in general than the new model this one looks pretty similar to

That original image but uh yeah I mean it definitely looks like a grumpy old fish but not nearly as detailed or as coherent as this model which just looks and feels a lot more cleaner I think you guys can agree this is like a really blurry image this one’s not so great

Either and this is just bam like sharp clean kind of like mid-journey again mid Journey has like the sharp clean color aspect to it it’s more of an interesting image I think and it’s more of a settled image a more complete image so I think this new model is really going after

What mid Journey V4 can already do here are our results from mid Journey before I gotta say Dolly 2 is a new model or or the dolly 2.5 model stomped all over mid Journey V4 for this prompt these images are they’re Bland they’re kind of boring

That the fins are sticking out of the Goldfish Bowl he doesn’t look particularly old the detail really isn’t there and neither is the camera angle so again Dolly 2 has actually always been pretty good at following prompts it’s always just been like a coherency and resolution problem so this new model is

Definitely really capturing this prompt a lot better than I think mid-journey V4 is I think a lot of what is being shown off with this new experimental Dolly 2 model is the extra detail that we’re getting out of a lot of these again this looks like a nice photoshopped image

This almost looks like I couldn’t tell it was generated by Ai and yeah again you can see a lot of detail no way this image would be generated from the original Dolly model but I could see it being generated in um mid-journey that’s for sure maybe not with hands that look

This good though gotta say the hand came out pretty good in this this result something that reminds me a little bit of mid-journey I think wow these actually all remind me a lot of mid-journey I have to say it’s very very weird though it’s like these really subtle little details like you could

Just tell them like the numbers and stuff and the way it’s doing lines it just looks more like a typical mid-journey diffusion model to me than it does the previous Dolly model which I guess is a good thing here we go we have another prompt we can try we actually

Have multiple results for this one so let’s try this in original Dolly and mid-journey hanging gardens of Babylon as a travel poster from the 1960s this is pretty good nice sharp detailed results I mean it could be a little bit better in a lot of areas but I would

Consider this above average results for AI image generation here’s what we got from regular Dolly to this is all close enough to the problems I say but there’s like a lot of just gross text on there that doesn’t actually make any sense at all I mean the letters are kind of

Correct at the end of the day but it’s starting to make up letters and it’s not actually writing the letters correctly though and the results look mushy we’ll say that they look mushy this one’s pretty decent this is super mushy and gross and this is pretty mushy as well

Original result again looks pretty darn good pretty high resolution better than what we were used to be seeing and here is what we got from mid-journey and mid-journey honestly is doing a better job I think with this problem mid Journey V4 I think is beating the new

Dolly 2 model in this one at least this looks more like a poster I actually think it’s following the prompt better with the new mid-journey model because yes it’s still putting these ugly not spelled correctly words in here but the Babylon Gardens look better they look

More like a 1960s poster and they just fit better as a coherent poster image I think I just think mid Journey’s following the prompt better for sure this is a different kind of test this is thought bubble pop art comic book art so it’s really you know they’re after a

Coherent image here the background on this honestly isn’t that great and it’s not super coherent but this is a tough test for AI art generators to pull off those perfect dots in the background seamlessly blending together and all of that that’s not easy to resolve as a a

High art image diffusion model let’s see what the original Dolly model thinks of this and honestly this is not super impressive you know looking back at the original one the old Dolly model barely did worse than it and you know these aren’t too bad I suppose but they’re not

Really great either and you can see it’s just super hard to resolve that stuff in the background there and the new model is not really doing it that much better in my opinion and this is what we got out of mid-journey I don’t really like these results at all they’re adding

Extra things in that we didn’t want and there’s actually people in this result this time around and while the thought bubbles actually can end up pretty coherent it’s not just the thought bubble we’re getting extra stuff and we’re getting words in there we’re getting bubbles just bubbling around in

A woman’s face and they look like cool images but it’s not as versatile as you get with this dolly too result it’s not as usable as an actual image so yeah honestly don’t like the mid-journey result at all this image honestly doesn’t look super impressive to me but

It’s simply a woman studying at a desk next to a window overlooking a busy City below headphones on pencil in hand style of anime looks better than what I would guess typical Dolly produces but let’s see what typical Dolly gets us and we’ll try this with mid-journey I have a

Feeling mid-journey’s gonna gonna Stomp all over with this one to be honest so yeah again here is the results from the new Dolly image model and here’s what we get from current Dolly I’m not gonna lie these are atrocious I don’t think I like any of these images you can see it’s

Trying its best but it just looks like this weird gross collage I don’t like it at all like this is just horrific so yeah there’s actually a huge improvement with this image in comparison to the old one mid Journey did all black and white imagery for some reason I’m not sure

Where it picked that up because we didn’t really ask for black and white image but you know these these are all pretty pretty good Generations I would say actually pretty close to the original Dolly 2 experimental model generation in terms of I think coherency and all of that but yeah completely

Black and white for some reason some more just Alpha Testing results from this new model these are non-existent creatures they look like they could be real creatures I will say that Dolly 2 has always been pretty good at its photorealistic results it all depends really on what you were trying to

Generate but these look very very realistic they almost look like real photos if you showed me this and said oh this is some deep sea creature I’d be like oh yeah okay that seems legit these actually look really really good very very photorealistic oh this one’s lots

Of detail on there it almost Reliance the mid Journey with the colors and stuff it’s really how I feel about this model is it it reminds me a lot of mid-journey and it’s like the mid Journey V4 and Dolly 2 had a baby here is another result this results actually

Pretty impressive I think this is field full of flowers grass wind blowing Blue Sky thick brush Strokes new impressionism so you can actually really see those thick brush Strokes in this image it looks like someone took a picture of a high detail art painting or something like that the original Dolly

Model just kind of makes this Mushy Mushy mess you can see there is kind of some detail in there it definitely looks like a painting but it’s really broad it’s not that detailed at all and you definitely don’t really see those thick brush Strokes like you do with this

Result which is just way more coherent I think this results quite a bit better and captures a lot more of the prompt I think Than This original Dolly model is mid Journey going to capture these insane brush Strokes the answer is kind of yeah actually mid Journey did a

Really good job here I think you know definitely see all the flowers the grass blowing the whole prompt really is here and you are starting to see some of these brush Strokes again I haven’t upscaled any of these results but I think you can see more of the brush

Strokes than you do in the original Dolly to model and it’s a little bit more coherent too especially this result this is a result was really nice and so was the first one so again I think this is almost on par with this model I think

This this is a little bit better though I really think you can see those brush Strokes in there and I think it’s I think it’s a better result this looks so much like a real painting it really really does and here are some more insane results very reminiscent of my

Journey I keep saying that just the color and the detail and the way things generate is reminiscent of mid-journey in some ways like especially these types of images this is what you expect to see when you’re looking at mid-journey stuff people seem to be pretty impressed with

This new model here is clearly a DeLorean image not the most coherent thing in the world I will say this image needs quite a lot of work but I really like the sun rays at least in this one this image is pretty darn good as well

And you can see a lot of the color and everything and these results this is such a mid-journey image it’s a weird thing to say but like really it does remind me a lot of mid-journey some more results here this is actually a fantastic result I’m glad I went on

Twitter to look for some stuff to show you guys because man there’s actually some really cool stuff look at the UFO image this is really really cool again I wouldn’t put this past mid-journey though I really think mid-journey could do this maybe maybe not this one though this one’s pretty surprising this is

Interesting because we actually got text full text spelled out so look at that it might actually be able to do text every once in a while clearly not coherently and efficiently like let’s say deep Floyd i f but it might actually be able to do some some text again this is just

Experimental stuff so this is where we’re at right now things could get much better we don’t know but yeah definitely more detail in these results there’s definitely a coherency bump for sure not past mid-journey in my opinion and a lot of the results it seems like mid cherny

In this go head to head more often whereas the original Dolly 2 is just clearly worse than mid Journey V4 they got some space girl results in here which look pretty good the eyes and the faces still need some work though like this nose is pretty screwed up but the

Rest of it looks really really cool these prompts are very well done oh these are some really really cool Generations but yeah this is pretty good results this is a lot closer to what we’re used to nowadays here is the default Dolly 2 results and here are the

New ones definitely a lot more color and detail and creativity tossed in there they’re taking notes out of the mid journey book for sure beautiful images really really like these it’s cool stuff really really awesome stuff love to love to see it I’ve messed around with this

Uh style prompt before but it’s just really cool to show all the different pieces and stuff around it apparently it’s also confirmed there’s no extra resolutions or aspect ratios or anything like that it’s just a new model there are no added features some people really aren’t that impressed though with the

New model this is all just a bunch of generated pizzas from it and yeah these do look like pretty generic doll E2 results rather unexpected the new Dolly models a little bit more into Gore and blood than the previous one they might try to remove this though because you

Know they’re testing things out but yeah you can see this one’s a little bit more habitants willing to generate some of the stuff that you know you weren’t allowed to generate at all in the previous Dolly model and that’s one of the things that really held the original

Dolly 2 model back because you weren’t allowed to generate any Gore you weren’t allowed to generate very suggestive things let’s say whereas you have a lot more freedom with mid-journey or stable diffusion and I mean that is just that is a horrific disgusting gory image but it’s nice to see that they’re actually

Allowing it to do stuff like that because it does it very very well these are good images I think and maybe they’re learning that it’s okay to have a little bit of fun every once in a while in terms of the blood and the the creepy stuff the the PG-13 I think

Mid-journey is pretty good in terms of their safety system I think their safety system is is the way to go although a lot of people seem to think this is going to be a huge leap for Dolly too I mean it’s still in the experimental phages so you expect it to get better

This is clearly better than original dolly was I I think it’s still a little less coherent though in some areas like man what is going on here but it’s a pretty cool model here we go some more actual copyrighted properties here this is more of like a spider girl thing in

Comparison to the first one which wouldn’t let you generate any uh you know copyrighted properties at all this new model seems to be allowing you to get away with it I don’t know guys tell me you think down in the comments below is this not really a leap at all for

Dolly too is this you know just bringing it up to mid-journey standards is it better than mid-journey again they’re actually going to be updating this in the future it’s going to be better apparently when they fully release it so this could be mid-journey V4 out but of

Course mid-journey is working on a V5 model I don’t know guys it’s really hard to tell it looks like it’s a more free-formed model it’s going to be better which is good it might actually bring us to start using Dolly too again we kind of all stopped using Dolly to uh

To be honest now let me know what you guys think should they leave the the actual copyrighted properties stuff in there should they leave the gore in there in the blood I think that if they just kind of stick to what mid Journey’s path has been so far it’s been working

Out pretty good so yeah thanks so much for watching guys and thank you for tuning in please check out some of my other videos let’s chat about this on my Discord server and again where I originally heard about this news was on my Discord server that’s where I get a

Lot of my AI news from is the community viewers like you and my Discord so check it out I’ll see you guys next time and goodbye

,00:00 honestly viewers this is a video that I
00:02 didn’t think I would ever make but I’m
00:05 really glad that I’m making it you see
00:07 when Dolly 2 first hit the scene almost
00:10 a year ago now it was seen as complete
00:13 magic it’s still a pretty decent AI art
00:16 image generator and its results can
00:18 still be absolutely fantastic and at the
00:20 time it was mind-blowing but almost a
00:22 year here later mid-journey V4 is
00:24 absolutely wiping the floor with Dolly
00:27 2’s results and quite honestly stable
00:29 diffusion’s versatility is seen as much
00:33 higher value than Dolly 2’s limited
00:35 features and ability to tweak
00:38 Generations however Dolly 2’s editor
00:40 without painting and in painting still
00:42 is fantastic and it still can generate
00:45 really nice images as you can see from
00:46 these examples but mostly everyone uses
00:49 mid Journey now for that added extra
00:51 creativity the colors pop more it’s
00:54 really vibrant and interesting to be
00:56 honest the pricing model in mid Journey
00:58 V4 is a launch cheaper than Dolly and
01:01 it’s a little less egregious it’s like a
01:03 subscription rather than Credits based
01:06 however there’s a new Dolly model in the
01:09 works this is very very exciting I heard
01:12 rumors about a new Dolly model months
01:14 and months ago even emad the creator of
01:16 stable diffusion was tweeting about a
01:19 new Dolly 3 Model so somehow he was able
01:22 to predict that they were working on an
01:24 upgrade to Dolly too the reason I’m
01:26 showing you this rather hilarious image
01:28 is because it actually highlights a lot
01:30 of what Dolly 2 originally was missing
01:33 the characters were trash in Dolly too
01:34 so if we go into original Dolly 2 here
01:37 and do a very similar prompt to what
01:39 this is again it definitely looks kind
01:42 of like Sonic the Hedgehog but it’s not
01:44 very coherent at all the creators of
01:46 Dolly openai made sure to try to get all
01:49 of the famous characters out of the
01:51 database as much as possible so while
01:53 this definitely resembles Sonic it’s a
01:56 little off-brand and weird feeling this
01:58 Sonic feels a lot lot more on par with
02:01 what you’re used to be seeing from the
02:03 character and also you’ll notice in
02:05 comparison to a lot of these generations
02:07 with which look a little bit sloppy
02:09 they’re not as detailed this one is is
02:12 very high detail high resolution the
02:15 burgers are very coherent Sonics very
02:17 coherent he’s actually rather chubby and
02:19 large because he’s been eating so many
02:20 cheeseburgers and it just looks like a
02:23 dolly three rather than a dolly too to
02:25 be honest although people are calling
02:27 this dolly 2.5 now if we head over to
02:30 the dolly to Instagram you can see here
02:32 is a post the first post that I can find
02:34 that they made about this new
02:35 experimental Dolly 2 model I’m thrilled
02:39 to share with everyone some work I’ve
02:40 been generating with the at open AI
02:42 Dolly Next Generation experimental
02:44 models these are still in the testing
02:47 phase and openai is using works like
02:49 this to push and test the model in each
02:52 and every way before launching worldwide
02:54 so we will be seeing this model
02:56 worldwide confirmed the goal is to have
02:59 artists like like myself offer Early
03:01 Access feedback reaction suggestions and
03:03 any concerns you know for this
03:05 experimental model so yeah it’s it’s on
03:07 lockdown right now it’s it’s like this
03:08 very mysterious email that’s being
03:10 pushed out people actually in my Discord
03:12 server have been getting access to it so
03:14 far which is very exciting the email
03:15 looks a little something like this hello
03:17 from the doll E2 team you were an early
03:19 contributor and you gave us good
03:21 feedback on the original Dolly model so
03:23 it seems like if you gave good feedback
03:25 and useful feedback early on in the
03:27 dolly model you are granted access to
03:30 this new experimental model and you get
03:32 unlimited credits and they’re only
03:33 running it for about 10 days while they
03:36 collect some feedback on it and yes the
03:38 person who did receive this email in my
03:41 Discord server was the one who created
03:43 this Sonic Image anyways here is our
03:46 first generation here in the example of
03:49 the new model this one I think is a
03:51 little bit more telling than the Sonic
03:52 example we first looked at again pay
03:55 attention guys to the amount of detail
03:57 that you can see in this image it is
03:59 very very detailed more so than you get
04:02 with typical Dolly too it looks a lot
04:04 more realistic the Grain and the texture
04:06 I think looks a lot better on the
04:08 spacesuit it still reminisces of the
04:10 dolly model which is very interesting it
04:12 just seems like the dolly we all hoped
04:15 for in the beginning and then we have up
04:17 here some text at the top that I think
04:19 looks more coherent than text and
04:21 numbers that you would typically see
04:23 from Dolly this looks more like the kind
04:26 of text that you would see from
04:27 mid-journey for example however there
04:30 actually is an AI art model out there
04:32 known as deep Floyd that does text and
04:35 can actually completely write sentences
04:37 I did a whole video about it before you
04:40 can see here are some results on Twitter
04:41 from the deep Floyd if model that does
04:45 text to image and it can actually
04:47 completely spell and I’ve tried this out
04:49 and I have access to this model and I
04:50 gotta say it’s pretty pretty awesome so
04:53 that’s one thing to note here they don’t
04:54 seem to have Dolly spelling out
04:56 sentences or anything like that just yet
04:58 and I think that’s actually very very
05:00 very crucial for the model but it’s
05:02 getting better than it was and I think
05:03 it’s on par with mid Journey as far as I
05:06 could tell in terms of spelling
05:07 definitely quite a lot more detail than
05:09 we typically get from Dolly too I will
05:13 say that here’s a rather terrifying
05:15 image from this new Dolly model here
05:18 honestly reminds me a lot more of
05:20 mid-journey in some ways especially with
05:22 this image here and the details and
05:24 stuff you can see but yeah it’s a gross
05:26 image it’s very coherent though and I
05:27 think they’re doing a pretty good job I
05:29 do want to compare this model to Mid
05:31 Journey we just need prompted images so
05:34 this prompt was Basset Hound portrait 8K
05:36 as Roman tile mosaic pretty good results
05:39 here from the new Dolly model let’s take
05:41 a look at the dolly model we have access
05:43 to right now and the current Dolly model
05:45 clearly leaves a lot to be desired these
05:48 results are really poor honestly I don’t
05:50 think any of these really could be
05:52 considered usable so yeah the new model
05:54 definitely is improving quite a lot on
05:57 current Dolly but that’s not a shock is
05:59 it Improv proving on what mid-journey
06:01 can do right now I gotta say guys this
06:04 is a little tricky because these
06:06 mid-journey results for the same exact
06:08 prompt look pretty darn good especially
06:10 this one down in the corner I think it
06:11 reminds me a lot actually of that
06:13 original image so again here’s the new
06:15 Dolly here’s what Dolly currently
06:17 generates and here is what we got out of
06:20 mid-journey so I think that this new
06:22 Dolly update is sort of putting it
06:24 almost on par with mid-journey although
06:25 I I gotta say mid journey is looking
06:27 very creative here especially with this
06:29 one this one’s really really nice here’s
06:31 some more results with the experimental
06:33 model again don’t this does this kind of
06:35 like actually remind you guys of mid
06:36 Journey a little bit I feel like there’s
06:38 a lot more color and Detail in these
06:40 results and like the way that the shapes
06:42 are formed they remind me more of mid
06:44 Journey pretty cool results nonetheless
06:46 definitely more detail I really really
06:48 liked this image that was generated with
06:50 the new model of course we have a prompt
06:51 for this one as well so we’re going to
06:53 be doing a comparison but yeah very very
06:55 detailed image here you can see the
06:57 texture in the skin of the goldfish is
06:59 absolutely fantastic it’s a realistic
07:01 photo but it’s still artistic in some
07:02 way the color is really nice and yeah
07:05 the whole thing is just very coherent I
07:07 have a feeling that dolly is not going
07:09 to do great with this prompt so let’s
07:11 see what current Dolly does and honestly
07:13 I gotta say current Dolly did better
07:15 than I anticipated but definitely all of
07:17 these are less detailed and they’re just
07:20 less coherent in general than the new
07:22 model this one looks pretty similar to
07:24 that original image but uh yeah I mean
07:26 it definitely looks like a grumpy old
07:27 fish but not nearly as detailed or as
07:30 coherent as this model which just looks
07:32 and feels a lot more cleaner I think you
07:33 guys can agree this is like a really
07:35 blurry image this one’s not so great
07:37 either and this is just bam like sharp
07:39 clean kind of like mid-journey again mid
07:42 Journey has like the sharp clean color
07:44 aspect to it it’s more of an interesting
07:46 image I think and it’s more of a settled
07:48 image a more complete image so I think
07:50 this new model is really going after
07:52 what mid Journey V4 can already do here
07:55 are our results from mid Journey before
07:57 I gotta say Dolly 2 is a new model or or
07:59 the dolly 2.5 model stomped all over mid
08:02 Journey V4 for this prompt these images
08:04 are they’re Bland they’re kind of boring
08:07 that the fins are sticking out of the
08:09 Goldfish Bowl he doesn’t look
08:10 particularly old the detail really isn’t
08:13 there and neither is the camera angle so
08:15 again Dolly 2 has actually always been
08:16 pretty good at following prompts it’s
08:18 always just been like a coherency and
08:20 resolution problem so this new model is
08:22 definitely really capturing this prompt
08:24 a lot better than I think mid-journey V4
08:28 is I think a lot of what is being shown
08:30 off with this new experimental Dolly 2
08:32 model is the extra detail that we’re
08:35 getting out of a lot of these again this
08:36 looks like a nice photoshopped image
08:38 this almost looks like I couldn’t tell
08:40 it was generated by Ai and yeah again
08:43 you can see a lot of detail no way this
08:45 image would be generated from the
08:46 original Dolly model but I could see it
08:48 being generated in um mid-journey that’s
08:50 for sure maybe not with hands that look
08:51 this good though gotta say the hand came
08:53 out pretty good in this this result
08:54 something that reminds me a little bit
08:56 of mid-journey I think wow these
08:58 actually all remind me a lot of
09:01 mid-journey I have to say it’s very very
09:03 weird though it’s like these really
09:05 subtle little details like you could
09:07 just tell them like the numbers and
09:08 stuff and the way it’s doing lines it
09:10 just looks more like a typical
09:12 mid-journey diffusion model to me than
09:14 it does the previous Dolly model which I
09:17 guess is a good thing here we go we have
09:19 another prompt we can try we actually
09:20 have multiple results for this one so
09:23 let’s try this in original Dolly and
09:25 mid-journey hanging gardens of Babylon
09:27 as a travel poster from the 1960s this
09:29 is pretty good nice sharp detailed
09:32 results I mean it could be a little bit
09:33 better in a lot of areas but I would
09:35 consider this above average results for
09:38 AI image generation here’s what we got
09:40 from regular Dolly to this is all close
09:44 enough to the problems I say but there’s
09:45 like a lot of just gross text on there
09:47 that doesn’t actually make any sense at
09:50 all I mean the letters are kind of
09:52 correct at the end of the day but it’s
09:54 starting to make up letters and it’s not
09:56 actually writing the letters correctly
09:58 though and the results look mushy we’ll
10:00 say that they look mushy this one’s
10:01 pretty decent this is super mushy and
10:04 gross and this is pretty mushy as well
10:06 original result again looks pretty darn
10:09 good pretty high resolution better than
10:11 what we were used to be seeing and here
10:13 is what we got from mid-journey and
10:15 mid-journey honestly is doing a better
10:18 job I think with this problem mid
10:19 Journey V4 I think is beating the new
10:21 Dolly 2 model in this one at least this
10:23 looks more like a poster I actually
10:24 think it’s following the prompt better
10:26 with the new mid-journey model because
10:28 yes it’s still putting these ugly not
10:31 spelled correctly words in here but the
10:34 Babylon Gardens look better they look
10:36 more like a 1960s poster and they just
10:38 fit better as a coherent poster image I
10:42 think I just think mid Journey’s
10:44 following the prompt better for sure
10:45 this is a different kind of test this is
10:48 thought bubble pop art comic book art so
10:51 it’s really you know they’re after a
10:53 coherent image here the background on
10:55 this honestly isn’t that great and it’s
10:57 not super coherent but this is a tough
10:59 test for AI art generators to pull off
11:02 those perfect dots in the background
11:04 seamlessly blending together and all of
11:06 that that’s not easy to resolve as a a
11:09 high art image diffusion model let’s see
11:11 what the original Dolly model thinks of
11:13 this and honestly this is not super
11:16 impressive you know looking back at the
11:19 original one the old Dolly model barely
11:22 did worse than it and you know these
11:24 aren’t too bad I suppose but they’re not
11:26 really great either and you can see it’s
11:27 just super hard to resolve that stuff in
11:29 the background there and the new model
11:30 is not really doing it that much better
11:32 in my opinion and this is what we got
11:35 out of mid-journey I don’t really like
11:37 these results at all they’re adding
11:39 extra things in that we didn’t want and
11:41 there’s actually people in this result
11:43 this time around and while the thought
11:45 bubbles actually can end up pretty
11:46 coherent it’s not just the thought
11:48 bubble we’re getting extra stuff and
11:49 we’re getting words in there we’re
11:50 getting bubbles just bubbling around in
11:52 a woman’s face and they look like cool
11:54 images but it’s not as versatile as you
11:58 get with this dolly too result it’s not
12:00 as usable as an actual image so yeah
12:03 honestly don’t like the mid-journey
12:05 result at all this image honestly
12:07 doesn’t look super impressive to me but
12:09 it’s simply a woman studying at a desk
12:10 next to a window overlooking a busy City
12:12 below headphones on pencil in hand style
12:15 of anime looks better than what I would
12:17 guess typical Dolly produces but let’s
12:20 see what typical Dolly gets us and we’ll
12:22 try this with mid-journey I have a
12:23 feeling mid-journey’s gonna gonna Stomp
12:25 all over with this one to be honest so
12:27 yeah again here is the results from the
12:29 new Dolly image model and here’s what we
12:32 get from current Dolly I’m not gonna lie
12:34 these are atrocious I don’t think I like
12:36 any of these images you can see it’s
12:38 trying its best but it just looks like
12:40 this weird gross collage I don’t like it
12:44 at all like this is just horrific so
12:46 yeah there’s actually a huge improvement
12:47 with this image in comparison to the old
12:50 one mid Journey did all black and white
12:52 imagery for some reason I’m not sure
12:54 where it picked that up because we
12:56 didn’t really ask for black and white
12:57 image but you know these these are all
13:00 pretty pretty good Generations I would
13:02 say actually pretty close to the
13:04 original Dolly 2 experimental model
13:06 generation in terms of I think coherency
13:09 and all of that but yeah completely
13:11 black and white for some reason some
13:13 more just Alpha Testing results from
13:15 this new model these are non-existent
13:17 creatures they look like they could be
13:19 real creatures I will say that Dolly 2
13:21 has always been pretty good at its
13:22 photorealistic results it all depends
13:24 really on what you were trying to
13:25 generate but these look very very
13:27 realistic they almost look like real
13:29 photos if you showed me this and said oh
13:31 this is some deep sea creature I’d be
13:32 like oh yeah okay that seems legit these
13:35 actually look really really good very
13:37 very photorealistic oh this one’s lots
13:39 of detail on there it almost Reliance
13:41 the mid Journey with the colors and
13:43 stuff it’s really how I feel about this
13:44 model is it it reminds me a lot of
13:47 mid-journey and it’s like the mid
13:48 Journey V4 and Dolly 2 had a baby here
13:51 is another result this results actually
13:53 pretty impressive I think this is field
13:54 full of flowers grass wind blowing Blue
13:57 Sky thick brush Strokes new
13:59 impressionism so you can actually really
14:01 see those thick brush Strokes in this
14:03 image it looks like someone took a
14:04 picture of a high detail art painting or
14:07 something like that the original Dolly
14:09 model just kind of makes this Mushy
14:11 Mushy mess you can see there is kind of
14:14 some detail in there it definitely looks
14:15 like a painting but it’s really broad
14:18 it’s not that detailed at all and you
14:20 definitely don’t really see those thick
14:22 brush Strokes like you do with this
14:23 result which is just way more coherent I
14:26 think this results quite a bit better
14:28 and captures a lot more of the prompt I
14:29 think Than This original Dolly model is
14:32 mid Journey going to capture these
14:34 insane brush Strokes the answer is kind
14:37 of yeah actually mid Journey did a
14:38 really good job here I think you know
14:40 definitely see all the flowers the grass
14:42 blowing the whole prompt really is here
14:44 and you are starting to see some of
14:46 these brush Strokes again I haven’t
14:47 upscaled any of these results but I
14:50 think you can see more of the brush
14:51 Strokes than you do in the original
14:53 Dolly to model and it’s a little bit
14:56 more coherent too especially this result
14:58 this is a result was really nice and so
15:00 was the first one so again I think this
15:02 is almost on par with this model I think
15:04 this this is a little bit better though
15:06 I really think you can see those brush
15:08 Strokes in there and I think it’s I
15:09 think it’s a better result this looks so
15:12 much like a real painting it really
15:13 really does and here are some more
15:15 insane results very reminiscent of my
15:18 journey I keep saying that just the
15:19 color and the detail and the way things
15:21 generate is reminiscent of mid-journey
15:23 in some ways like especially these types
15:25 of images this is what you expect to see
15:27 when you’re looking at mid-journey stuff
15:29 people seem to be pretty impressed with
15:31 this new model here is clearly a
15:33 DeLorean image not the most coherent
15:35 thing in the world I will say this image
15:37 needs quite a lot of work but I really
15:39 like the sun rays at least in this one
15:41 this image is pretty darn good as well
15:43 and you can see a lot of the color and
15:45 everything and these results this is
15:47 such a mid-journey image it’s a weird
15:49 thing to say but like really it does
15:50 remind me a lot of mid-journey some more
15:53 results here this is actually a
15:55 fantastic result I’m glad I went on
15:57 Twitter to look for some stuff to show
15:58 you guys because man there’s actually
16:00 some really cool stuff look at the UFO
16:02 image this is really really cool again I
16:05 wouldn’t put this past mid-journey
16:06 though I really think mid-journey could
16:08 do this maybe maybe not this one though
16:10 this one’s pretty surprising this is
16:12 interesting because we actually got text
16:14 full text spelled out so look at that it
16:16 might actually be able to do text every
16:18 once in a while clearly not coherently
16:20 and efficiently like let’s say deep
16:22 Floyd i f but it might actually be able
16:24 to do some some text again this is just
16:26 experimental stuff so this is where
16:28 we’re at right now things could get much
16:30 better we don’t know but yeah definitely
16:32 more detail in these results there’s
16:34 definitely a coherency bump for sure not
16:37 past mid-journey in my opinion and a lot
16:39 of the results it seems like mid cherny
16:41 in this go head to head more often
16:43 whereas the original Dolly 2 is just
16:45 clearly worse than mid Journey V4 they
16:47 got some space girl results in here
16:49 which look pretty good the eyes and the
16:51 faces still need some work though like
16:53 this nose is pretty screwed up but the
16:54 rest of it looks really really cool
16:56 these prompts are very well done oh
16:58 these are some really really cool
17:00 Generations but yeah this is pretty good
17:02 results this is a lot closer to what
17:04 we’re used to nowadays here is the
17:07 default Dolly 2 results and here are the
17:09 new ones definitely a lot more color and
17:11 detail and creativity tossed in there
17:14 they’re taking notes out of the mid
17:15 journey book for sure beautiful images
17:17 really really like these it’s cool stuff
17:20 really really awesome stuff love to love
17:23 to see it I’ve messed around with this
17:25 uh style prompt before but it’s just
17:26 really cool to show all the different
17:28 pieces and stuff around it apparently
17:30 it’s also confirmed there’s no extra
17:32 resolutions or aspect ratios or anything
17:34 like that it’s just a new model there
17:36 are no added features some people really
17:39 aren’t that impressed though with the
17:41 new model this is all just a bunch of
17:43 generated pizzas from it and yeah these
17:45 do look like pretty generic doll E2
17:47 results rather unexpected the new Dolly
17:50 models a little bit more into Gore and
17:52 blood than the previous one they might
17:54 try to remove this though because you
17:55 know they’re testing things out but yeah
17:57 you can see this one’s a little bit more
17:59 habitants willing to generate some of
18:00 the stuff that you know you weren’t
18:02 allowed to generate at all in the
18:04 previous Dolly model and that’s one of
18:05 the things that really held the original
18:07 Dolly 2 model back because you weren’t
18:09 allowed to generate any Gore you weren’t
18:11 allowed to generate very suggestive
18:13 things let’s say whereas you have a lot
18:15 more freedom with mid-journey or stable
18:17 diffusion and I mean that is just that
18:19 is a horrific disgusting gory image but
18:21 it’s nice to see that they’re actually
18:23 allowing it to do stuff like that
18:24 because it does it very very well these
18:26 are good images I think and maybe
18:28 they’re learning that it’s okay to have
18:30 a little bit of fun every once in a
18:32 while in terms of the blood and the the
18:34 creepy stuff the the PG-13 I think
18:36 mid-journey is pretty good in terms of
18:38 their safety system I think their safety
18:40 system is is the way to go although a
18:42 lot of people seem to think this is
18:44 going to be a huge leap for Dolly too I
18:46 mean it’s still in the experimental
18:47 phages so you expect it to get better
18:49 this is clearly better than original
18:51 dolly was I I think it’s still a little
18:54 less coherent though in some areas like
18:56 man what is going on here but it’s a
18:59 pretty cool model here we go some more
19:01 actual copyrighted properties here this
19:04 is more of like a spider girl thing in
19:06 comparison to the first one which
19:07 wouldn’t let you generate any uh you
19:10 know copyrighted properties at all this
19:12 new model seems to be allowing you to
19:14 get away with it I don’t know guys tell
19:16 me you think down in the comments below
19:17 is this not really a leap at all for
19:19 Dolly too is this you know just bringing
19:22 it up to mid-journey standards is it
19:24 better than mid-journey again they’re
19:26 actually going to be updating this in
19:28 the future it’s going to be better
19:29 apparently when they fully release it so
19:31 this could be mid-journey V4 out but of
19:33 course mid-journey is working on a V5
19:35 model I don’t know guys it’s really hard
19:38 to tell it looks like it’s a more
19:39 free-formed model it’s going to be
19:41 better which is good it might actually
19:42 bring us to start using Dolly too again
19:44 we kind of all stopped using Dolly to uh
19:48 to be honest now let me know what you
19:49 guys think should they leave the the
19:51 actual copyrighted properties stuff in
19:54 there should they leave the gore in
19:55 there in the blood I think that if they
19:58 just kind of stick to what mid Journey’s
20:00 path has been so far it’s been working
20:01 out pretty good so yeah thanks so much
20:03 for watching guys and thank you for
20:05 tuning in please check out some of my
20:06 other videos let’s chat about this on my
20:08 Discord server and again where I
20:10 originally heard about this news was on
20:11 my Discord server that’s where I get a
20:13 lot of my AI news from is the community
20:15 viewers like you and my Discord so check
20:18 it out I’ll see you guys next time and
20:19 goodbye
, , , #OpenAI #Supercharged #DALL #Midjourney #Lets #Find , [agora]