No Waitlist! Sora Competitor with FREE ACCESS! | Dream Machine First Look

[@MattVidPro]

Finally AI Video that comes close enough to SORA that we can access for FREE! Don't forget to join my *LIVE* tomorrow to request YOUR Dream Machine Gen Ideas as a viewer! kzbin.info/www/bejne/jnbcnYyEbtFlnbc&ab_channel=MattVidProAI We will also review my gens from this video in that stream. (they are STILL in queue)

[@NicVandEmZ]

It’s awful slow now

[@LouisGedo]

👋

[@davidcapuzzo]

was this trained on movies?

[@mobster5626]

Had the same problem, you just have to click the videos. That works for me at least.

[@Ovyron]

I expected this by November 2024, so it's hard to know what to expect in 1 or two years because my expectations are being outrun!

[@pon1]

bigger companies means bigger risk if things goes sideways, I think that's why we see these smaller companies always bringing the models to the public before the bigger ones. This way the smaller ones take all the risk while the bigger companies learn from it. It's not fair but it's the name of the game.

[@LaughSleepMeme]

Generations take 120 seconds. However, due to extremely high demand, your request will be queued.....FOREVER...

[@armondtanz]

And ever... amen

[@HCforLife1]

it's fine now

[@EnterPlayMode]

Hey Matt, you've probably noticed this, but most of your generations are done. To start, click on the video-it won't begin automatically. When you see "dreaming" in the top left corner, it's still generating. Once that's gone, it's ready. Just click to play.

[@nevergonnagiveyouupnevergo3263]

Artifical Theorists

[@shavedsunglasses]

Do any of you find it curious as to how much AI generations look like our own dreams while sleeping...???🤔🤔🤔

[@MattVidPro]

i have also noticed this

[@yvandanny911]

Or nightmare 😂

[@justinwescott8125]

These aren't what my dreams look like

[@Scn64]

Somewhat yeah, especially when it has trouble keeping a consistent character. I'll be dreaming about a dog and all of a sudden it will morph into my second grade teacher, lol

[@24-7gpts]

@@Scn64 Lol exactly!

[@kevnar]

Don't bother. I just went to this site, typed in one prompt, and gave up after a 2-hour wait. "Images are generated in120 seconds, but due to high demand, requests are placed into a queue..." Maybe I'll check back at 3 am.

[@noyu8]

Facebook you can get a couple frame photo. It's not perfect but infinite times better than everything else

[@armondtanz]

I'm up to 4 hours. Stoopid them doing this and not giving any info? Wot would happen if u paid? Even a 10th of the time would be 25 min each generation. 😢

[@DangerManQ]

@@armondtanz for me it takes like 4 minutes to generate

[@armondtanz]

@@DangerManQ are u on the free plan?

[@kevnar]

@@armondtanz They probably weren't expecting millions of people to all show up in the same day.

[@lokitsar5799]

Good luck getting anything. I've had a generation waiting over an hour now. They were prepared about as much as SD3 for the launch today.

[@americanjesusshow4087]

me 2 but several hours

[@bloxyman22]

3-4 hours here and still in queue..

[@armondtanz]

I've just topped 4 hours. Is there any info on paying? And wot does it speed up by?

[@timer4times2]

@@armondtanz I've got several cool generations in under 20 minutes....

[@armondtanz]

@@timer4times2 are u on the free plan?

[@DragonBallDetroit]

Well sora is for sure gonna drop soon for competition

[@24-7gpts]

Probably, 10 generations per month on plus plan

[@DragonBallDetroit]

@@24-7gpts i think it's probably going to go with the open ai playground approach. You put your card in and pay per generation

[@1sava]

Unlikely. They probably won’t drop before the US election has passed. OpenAI has had a lot of bad press lately, so i doubt they’d risk releasing it ahead of the presidential election.

[@PaulArnold-d4b]

@@1sava i dont understand the election Argument - because a few years later we have the next election - so i dont understand which difference it makes

[@puikplan]

​@@PaulArnold-d4b two reasons, as of the last update with the WSJ interview Mira claimed that 1. the current cost is too high (they want to bring it down to the cost of one Dall-E generation) and 2. The generation time is too long, it took 5 minutes to generate one video, which is too long for commercial use. I hope Matt also sees this comment, so he also understands why they are not releasing Sora yet. As this is publicly available information

[@Atreyuwu]

At least this channel doesn't try to sell it as "BETTER THAN SORA!!!! BEST AI VIDEO GENERATOR YET!!!!".

[@MattVidPro]

We don’t have access to Sora, but both kling and sora seem to be better than this. However this is accessible

[@xAgentVFX]

OMG The anime... We goona see a burst of new anime boiiiiiiiii

[@TombstoneDaDeadman]

I tried it. There are just way too many people trying to use it now. The queue seems very long. Ill give it another try later

[@Krysanthia]

Same problem ! 😅

[@MemePopperz]

14 mins ago Runway announced their Best model yet Gen 3 alpha...Probably the best video generator...im so excited! Waiting for your new video, Matt!

[@levonkenney]

suno just came out with there audio input feature now you will have to compare it to udio to see if it is any better

[@tokyobobcat]

I've been playing with it, and I think suno does it a little better because the songs are more complete. However Udio still does it better with music quality.

[@jeffwads]

They are getting hammered. What a surprise.

[@NakedSageAstrology]

Except you can't actually use it. 3 hours later and I'm still waiting in que.

[@wtflolomg]

I was able to see my generations clicking on the blank thumbnail. If it's not "dreaming" it's probably done.

[@atxmix2238]

4am central time, 1min queue, 1min creation

[@dommyshan]

Thanks! Trying it out now. Queue is working but you gotta wait about 2 minutes. Got two weird vids that didn't really follow my prompt (about 30% of it). One was ok. And I'm uploading a pic I drew and see if it can animate it haha :D

[@ayaneagano6059]

Did it do it? That’s what I want to do too

[@phen-themoogle7651]

1 year from now will probably be free for doing these kind of generations. It's not worth paying for them nowadays, unless you're actually already in the industry using the technology and it actually helps with your workflow idk I'll be back in a year.

[@ragemax8852]

It's working way better now, as my generations are taking between 2-10 minutes now which is cool. It's pretty cool so far, I just need them to make the videos longer than 5 seconds.

[@AIWithMohammed]

I tried it and gave a pic of my kids in a mini car and man surprisingly it did a decent job getting the car to move in the frame. I thin we about to see some crazy stuff soon

[@ChandravijayAgrawal]

I also got this stuck generation, its just a bug, i clicked it and it started playing

[@mtallan]

If you see a blank thumbnail after the “Dreaming” text stops showing, just click on the thumbnail to see the video. The thumbnail generates after the video has been watched.

[@VinMan-ql1yu]

It's quite fast right now, results don't always perfectly follow the prompt, but the few concepts in the prompt do end up influencing the video. This is the best model I've been able to test so far.

[@InsideYouTubeMinds]

I love the new editing style 🔥

[@agnesslovehealz]

This is fire it's funny we all waiting on sora meanwhile luma just blew it out of the water and who better to cover this than matt lol this is def the most wild capabilities the transitions is insane mind blowing and huge shifts and the ability to prompt 5 scenes etc and the Pixar style no morphing love and image to motion and quality is wowza

[@nightmisterio]

Video to video visual effects would be great 👍

[@wasfiakab]

Application error: a client-side exception has occurred (see the browser console for more information).

[@austinlevine]

bro this has such a long wait to generate

[@micallbeard923]

barely managed to generate after like 3 hours

[@jonplaud]

I made a prompt at 4:00pm today...it's 5:19pm now.

[@armondtanz]

8:40pm Now Midnight😢

[@HCforLife1]

they had server issues. Now generates in about 2 minutes

[@Meeeeeeepppp]

Works well ngl. Usually averages around 2-4 minutes

[@BillMill]

Sora and Kling is in a different league. This is in par with Krea probably slightly better.

[@SyntheticaShow]

I think it’s insane. I’ve been playing with it on free account and it’s been a lot of queuing but still… come on, openAI need to do something, because Luma is impressive AF

[@ItWasntAPhase]

This is amazing. The photo to video finally allows for consistency in AI video. Just made a little short with a sketch I made

[@Brenden-H]

generation only took ~6 minutes for me. It looked grayed out like it wasnt ready, but it was just the thumbnail not loading. I clicked on the empty box because the "dreaming..." text went away, and sure enough it loaded the video it generated (prompt adherence is a bit iffy)

[@JPeoplzOIII]

Am I the only one anticipating the day one of these AI news channel hosts starts their video with a realistic AI rendered version of themselves opening the video, only to then reveal that it was AI like 5 minutes into the video? Matt's newish lighting had me think min it was AI at first.

[@FSK2]

I finally got the results after overnight. I tried the image to video. They were very fast in motion. I tried text to video for pixar style and it was almost perfect. The motion was just fine. I tried at least 8 or 9 prompts and i got a message from the app you have reached maximum. If they add some sort of controls it will be epic. 😊

[@NeoKailthas]

Video needs another 6 months to a year it seems

[@hipjoeroflmto4764]

Give it a few more than a single year

[@mycelia_ow]

For what?

[@jimmysrandomness]

😳😳 Expensive!

[@FriscoFatseas]

i'll have to try this when the servers stabilize or whatever

[@IndigoBabies]

We just gunna skip over ass Pikachu! (another great video Matt)

[@luuketaylor]

Been playing around with this a bit just now and had some good and some weird results! I just wish they could make the interface more complex (allowing for negative prompts would be HUGE) and specific. I had a character just walk off screen despite saying they were leaning against a wall.

[@Jeremy-sk9qh]

you think stable diffusion will have an open source super good model like sora at some point?

[@emotedeeeznuts]

No. Stable Diffusion 3 was Terrible. Stable Diffusion is done imo

[@Klokinator]

Stable Diffusion is falling off a cliff. Their lead devs have jumped ship, and the new SD3 is looking like garbage. I expect nothing good from them anymore. It's a miracle if they can produce anything people want to use.

[@bazookaman1353]

​@@KlokinatorSad but true :( I'm starting to think there was moat all along.

[@anigah]

​@@Klokinatorare they the people that got access to one of my throw away google drive account? I forgot to remove it

[@bazookaman1353]

@@anigah Doubt, SD is open source and locally ran. You may have used a middleman service that runs it for you tho, there are a lot of them.

[@micbab-vg2mu]

it takes 1h to generate an animation :) - the quality is better - but still you need at least 3-5 generations to chose the good shot:) i hope soon we will have access to SORA

[@aexetan2769]

An average Google search uses 0.3 watt-hours of electricity. An average text-to-image prompt uses approximately 3 watt-hours of electricity. An average text-to-video prompt uses around 100 watt-hours of electricity.

[@hipjoeroflmto4764]

Zeus has some work to do

[@luizpaes7307]

Energy and computing power are going to be the new money

[@ChristianIce]

Worth it.

[@markmuller7962]

Less censorship than OpenAI, I welcome the progress

[@SoulSeeker]

actually if you click on the blank images they will pop up

[@24-7gpts]

So cool, waiting for the livestream!

[@vermadheeraj29]

I tried this from India and I'm getting results in less than five minutes.

[@jonmichaelgalindo]

Yeah I don't think this will have calmed down by tomorrow.

[@LordJonRay]

Just posted my first attempt using the Free version at making a short film with Luma AI Dream Engine. Had more to go but it stopped output on me on my 3rd clip.

[@millionize]

I just made the trailer video for my channel with Dream Machine. It was super fun!!!

[@CaroleBaskin]

Nothing like competition to provide the best tools for the users.

[@user-pw4ri3gc8c]

I have tried it, but it feels like the technology is not very mature yet. It needs to be further strengthened.

[@cesarsantos854]

Funny how video is advancing so fast but audio is not on Elevenlabs level yet or song generation on par with the heavily censored commercial ones.

[@JRBodds]

I got it to work for me straight away. One shot, gemerated in a bout 3 Min's! I tend to wake up early though, it's 2:45 AM here in austin...

[@makulVR]

A similar model like this offline usable would be a game changer

[@The_Questionaut]

I was thinking for text to video being easy and intuitive to pose and animate characters You could use something like controlnet with key frames, so you can use it for videos posing and movements of a character in the text to video generation. I don't think anyone has done this, I would love to see it done. Imagine posing your character just by dragging a stickman around then hitting the generate image button. Easy posing if someone pulls this off.

[@scottiewardle]

btw if you just refresh the page after 30secs , your video is usually done

[@titolindj]

Very nice! Hopefully the server load drops a little bit soon

[@onidaaitsubasa4177]

Hmnm, I was generating the prompts you were saying out loud, in my brain, and it was generated very fast and was very realistic, unfortunately the video was only a few seconds, and since it was generated in my brain it makes it really hard to share, which is a definite positive point for AI video, however if they had a live brain neuron system that uses an organic neural network, it would definitely be fast and easy to share with others.

[@alteredalley]

Luma is unbelievable! ❤👏👏👏👏

[@ucfufd]

Matt, u can click on the name of the generated video and it will work

[@MojaveHigh]

I come from the future. Generations are working great now!

[@gu9838]

videos are not bad but some camera control could help and wait times are pretty long right now

[@KingofHeroes-MLBB-CQC]

🎉 This is what I’ve been waiting for! 🎉

[@sammyjohnson5596]

I’m blown away by the title.

[@JeffreyWongOfficial]

Well, I could upload some images and tried some simple prompts on them and the outcome looked horrible

[@justinwhite2725]

Openai making the same mistake google did with dalle. Showcase neat AI but fail to release it before competitors eat up the market share

[@aisonikamusic]

Matt, you are the bomb again! Man this is awesome!

[@SaidThaher]

Pay for 5sec videos 😂😂😂😂

[@Skuuurt-Sama]

Some pay for pic generation services for one by one pictures which is litteraly one frame of a video so...

[@akanalytic]

“Due to high demand requests will be queued.”

[@jameshughes3014]

I want to know how they do it. the temporal coherence is astounding

[@koen.mortier_fitchen]

Finally something that I can use NOW. Thanks Matt

[@koen.mortier_fitchen]

aaand it sucks. The shown clips here are useless and I couldn't generate a vid bc of non-stop waiting

[@dniilii]

​@@koen.mortier_fitchengive them some time

[@ItWasntAPhase]

You only get 30 generations for free so make them count. The rush is over and servers working great now

[@apersonlikeanyother6895]

I tried it. its really good! Took about 15 minutes. But I am awake when Americans are asleep.

[@WillyFlowerz]

what's the name of the song from 0:10 to 0:45 during the DreamMachine trailer ?

[@OneLeggedDiver]

mine are generating in under 2 minute but they don't look as great as the examples in the video. Seems like they accommodated the influx of users at the expense of the quality maybe? IDK how that stuff works though

[@hipjoeroflmto4764]

More than likely

[@mycelia_ow]

The example videos used comprehensive prompts and used iteration, and probably had access to tools we don't (because closed-source). You simply cannot expect the same level of quality and detail with one, simple prompt. This applies to all generative services. Their promo example generations will always be better than what most users could create.

[@DadzisLoco]

2 mos later, I am on a waiting list, even some weeks ago I was generating vids, those took some 4 to 12 hrs to make.

[@DataJuggler]

17:35 One in the hand is better than two in the bush (that probably won't make it past censors)

[@TheSuperStroker]

We all dreamed about giving ourselves crazy hair in videos when we were kids

[@robinrichardson6495]

Not too good with running horses at the moment, looking forward to playing more with it

[@Shyeep]

People need to stop using it.... because i can't get my generations done :P

[@Typical-H12]

Same lol

[@FusionDeveloper]

4:40 movie clip of a horror movie where someone is using mind control to force him to eat a over-ripe lemon, against his will.

[@mycelia_ow]

This will improve quickly.

[@vustation]

No surprise the site is nearly crashed at this point, I'll tune in and try again later, wow

[@davemorgan6013]

You'll be stuck in a queue for hours unless you go online at 7-9 a.m. CET (1-3 a.m. EST) 😄

[@nickgirdwood3082]

I haven't really tried any of the video generators. What I have tried hasn't been overly impressive. It seems with a lot of these generators, the quality depends on the use case. DALL-E 3 is the best for images for me, and Suno is better than Udio for music. I don't where people are getting Suno sounding like singing into a fan from, I've generated tons of singing that have been fantastico! I've generated intro and outro music for my KZbin channel (Anti-HyperLink] and they sound great. Are these people only generating rap and hip-hop music?

[@DiceDecides]

welp, that's what happens when a good video model is released for free to the public... the servers can't handle the demand, hopefully we get another competitor that can off-load some demand

[@DiceDecides]

actually nevermind, it didn't take too long for me, 4 mins top

[@theneverwas2835]

what is the best open source option for video so far?

[@you-are-the-tharpsters]

17:14 reminds me of battlefield v

[@agnesslovehealz]

Haha this is so real Tim showed this one clip of person walking out and smoothly entering car it's mind blowing runway and pika we now consider old straight up cuz that's ai and ability to extend using last shot as image to then extend the things we can do with this just blows my mind the best is yet to come and if this is infancy stage oh boy let's gon

[@Alice8000]

Quick question how do i find diamonds in Minecraft really fast???

[@bertplatini2249]

Can you download the generated video.?

[@r0drig096]

So you cant create 3d models anymore ??

[@IntellectCorner]

it's working guys

[@Aimerxd]

What are the prompts in the video at minute 10:14 ?

[@Art-AI-and-beyond]

Tried an image. Its way better quality than stable diffusion and tooncrafter but very difficult to get it to do what you want. Not easily steerable with just a prompt.

[@IDJOT]

bro, you just had to click the empty box, they were loaded.. jesus

[@TheoCahill]

Wait times have now dropped to around 5 minutes

[@SuMiTMeshram25]

I think your videos have already generated you just need to open them in another tab, that's just visual bug I guess.

[@anawakoff]

I just finished a 3m30 AI short film with Haiper and Pixverse and Dream Machine come now ! My work is already outdated ! 😂😂