MattVidPro AI
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: https://www.futurepedia.io/tool/dalle-2
✩ Midjourney V4: https://www.futurepedia.io/tool/midjourney
✩ Openaidalle Instagram: https://www.instagram.com/openaidalle/
► MattVidPro Website: https://MattVidPro.com
► MattVidPro Discord: https://discord.gg/bQgcbjs2Sg
► Follow Me on Twitter: https://twitter.com/MattVidPro
————————————————-
▼ Extra Links of Interest:
✩ AI LINKS MASTER LIST: https://www.futurepedia.io/
✩ Lemon Gallery: https://MattVidPro.com/Gallery
✩ General AI Playlist: https://www.youtube.com/playlist?list=PLrfI66qWYbW3acrBQ4qltDBsjxaoGSl3I
✩ AI I use to edit videos: https://www.descript.com/?lmref=nA4fDg
✩ Second Channel: https://www.youtube.com/channel/UC06GdmaEdCdCFwR3NvszloQ
————————————————-
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: MattVidProSecond@gmail.com
Love the engagement!
Dope Content 🤘🏼😎💯💧
Say midjourney again. I dare, I double dare motherfucker. Say midjourney one more goddamn time!
I personally think this is a step in the right direction!
dalle is far too restrictive when it comes to adult content. even things that should not be "adult" like lingerie which could be for fashion is censored. also if you type too many prompts it forbids you can actually get banned. WHY? its not like you can generate those images. dalle1/mini was more fun even at lower res honestly with all the updates and user models i think stable defusion which is free to run on your own pc is the best.
9:46 they have GAY BIOS?!💀💀
yeah but also DALL-E 2 is performing worse nowadays than it used to. I generated some pretty decent images in the fist months when I got access, but lately just stopped using it because everything it generates now looks very chunky or lacking definition. They definitely nerfed it.
So can you feed any of them a lot of pictures of someone and they create fictional images of that person?
If you think about it any update is a leap
Mid journey still lookin like they in the lead 😅
Guys are there some ai generators that enable me to generate something into existing picture? Like a mix between generated and nongenerated picture?
It is not better then MidJourney, even Dalle 2
I was so excited to finally get access to Dalle2 Beta…. And then so quickly disappointed. I don't know that I've ever gotten a useful result from them…. It'll be interesting seeing them step it up after all this time. Thought it was strange that they were being so stagnant, when MJ & SD were accelerating practically weekly.
where is url for deepfloyd: want spelling
Those horribly burnt pizza slices look exactly like what is served at the middle school I work at 🤣
Dall-e 2 deserves to be better, they were the pioneers.
dalle 2.5(3) for sure has been trained on midjourney & alikes
If DALLe tries to sensor everything they'll get left behind. Anyone could have told them that.
1:43 that last pictute reminds me of that knock off toothbrush that combined both Sonic and Tails on the package.
POWERRRRRR
Dali is ass cheese
All of them are really…the censorship is fucking horrible…nudes are art…fucj censorship un all forms…fuck all censorship
The "Midjourney look" might just be the result of offset noise training – now implemented for SD as well (epi_NoiseOffset).
I feel like DALLE does a fantastic job with prompts that are more out there. I also have noticed that it’s images look a lot more realistic, in some instances, and look more like they were shot by average Joes, as opposed to professional photographers – which I really like, at times. That being said, I rarely use DALLE, and I really like how amazing pictures generated by Midjourney look.
you said artist like yourself–what is an artist ?
At 17:53 if I had the new engine I would do this prompt:a drawing of a purple block holding a knife looking at a mouse in a house
They are so far behind Midjourney and Stable Diffusion they might as well switch to another avenue of A.I.
Is it possible to create images with a transparent background with DALL:E? And how do I get rid of the stupid coloured boxes at the bottom right?
Excellent
HI MATT , please SUGGEST ANY AI SERVICE WHICH, DOES NOT INCLUDE AUTO SUBSCRIPTION,
AND ALLOWS TO SELL AI GENERATED IMAGES..
Why is the link for DALLE 2 on the web take you to a vanilla page with only "DALL E" inscribed in the GUI?
well, do we have to see your face all the time? i do not think so, let's be honest your ego is too large for your charisma.
it needs to have an option to search for the web and gather images to learn style instantly
DALL-e3 is in beta you can text to see video now but is not open to the public for now
I was about to ask why you wouldn't compare the new DALL-E model results against those from Midjourney v5, but I guess you made this video a couple of weeks before v5 came out 😉 Definitely worth doing an updated comparison whenever the new DALL-E version comes out (hopefully frikkin SOON).
As for the new version being only a model update without any additional features… I think the pure simplicity of DALL-E's UX/interface remains one of its greatest strengths. As much investigation as I've done into Midjourney and various flavors of Stable Diffusion, none have come close to DALL-E'S ease of use and seamless results when it comes to outpainting. Midjourney doesn't even do inpainting/outpainting. SB can do inpainting well enough, but even then it takes a lot more know-how to get it right than it does with DALL-E. Their major weakness thus far has been the quality of their image output, so I'm glad that's what they are choosing to concentrate on improving first (albeit still way overdue, considering Midjourney's gone up 2 major versions over the same time span).
I agree wit most of, what you say, but I have to disagree with you about Midjourney following the prompt better on the one with "Hanging Gardens of Babylon" than Dall-E 2.
Midjourney's pictures might be better, but there is nothing Hanging garden in any of its 4 images, and I think, that is a big issue with text to AI generators in their current state.
They do make awesome pictures, but that they have tendency to more or less ignore many of the instructions they are given.
Turn off your Discord notifications while you're recording. Sheesh.
clayster's twin?
Would it be possible to make realistic AI pics of Chrome balls? Like if the balls were in a Lab or professorially taken. What are your thoughts?