Haha, I tried to remain as neutral and objective as possible, but at the end of the day: whatever your takeaway, I’m happy to have taken the hit for everyone!
@@TheoreticallyMediaI could have avoided wasting time and money, by watching your video first!!! I even commented a few weeks ago that you are the most reliable source of info when it comes to MUSIC/IMAGES/VIDEO generative AI. I was too excited and curious... Despite one very good innovation, the Storyboard, SORA is a complete disaster! But at least I know now that OpenAI SORA will never make it... WHY? Keep in mind the time they took to get to THAT! Then the fact that we get errors in almost all generations... 3 hands humans! Objects disappearing! All kinds of inconsistency during short videos! Improper object movements... Now, just extrapolate the time and money it will take to correct all the flaws! During the same period of time with a fraction of the budget, KLING will be KING! OpenAI was already facing bankruptcy due to their precarious financial situation. It may be the last nail on the coffin!
For sure that delay did them no favors. But, I will say: pre-election, the media would have been ALL over them for releasing something so “dangerous” …even though there are plenty of other tools that do the same thing, and we barely saw anything on that front. In some ways, I do feel bad for OpenAI. ….uh, apparently enough that I gave them $200? Sigh.
@@zrakonthekrakon494 dunno, I tried it yesterday and honestly waiting 10 minutes for a base request just to end up with something in best case that doesn't do what I wrote at all, in worst case is weird and scary... Nah. It's like a hit or miss, like 5% hit 95% miss
Wow! Unless you find the 'we're just kidding' button, I won't be paying for this. I get much better results from Kling and Minimax! I do, however, appreciate you taking the bullet for us.
I’ve been saying for awhile, Sora is not going to be what everyone thought it was. Not to say it isn’t good- I’ll spend more time with it and I’m sure I’ll get some great results out of it- but “better” than anything else? That’s totally subjective
@@TheoreticallyMedia IMO when you use image2video, the image you provided should be the 1st frame. Not "incluenced by". That's the only way to get any accuracy to what you want (for example trying to make a short film, consisting of small clips), in the current AI video scene. Everything else is just a guessing machine. This makes Minimax superior imo. Because AI images can be handcrafted to exactly what you want (with inpainting). However what I've noticed with Minimax, is that image2video works best with real photos. If the AI image has even small errors in anatomy or other weirdness to it, it will multiply in video, and throw it off even more. Probably because Minimax was trained on real videos, so it understands real photos as reference points better. Or, AI images that are flawlessly realistic
I was tired of the channels who are just trying to ride a wave...this one really feels good and actually feels like did not waste my time, thank you for Good authentic info
In other words.... Open AI spent almost a whole year to come up with a tool that's basically still barebones and honestly subpar to any other AI out there now... great
To be totally honest, I was expecting it to be a lot more nerfed. So, in some ways this is kind of a win. I really thought public Sora was going to be DallE but Video.
@Screamer99REVIEWS Sora is more advanced in many areas of physics, but subpar in some other areas, but yes, 200 is far too much for what they are offering now. The Sora Team do not live in reality.
Thanks for taking the hit, Tim. Sora is way late to the game. The Chinese are killing it. Sora adding copyright control and $200.00, forget about it. Love OpenAI, but they dropped the ball on this. I don't see this being the dominant gen.
I can't use OpenAI without a cell phone, but yeah $200 is where companies get disconnected with their worth. I just paid $60 for a year of Kling in their half price sale, and it a far better value.
A greedz, 200 is waaaaay too much, Runway isn't as good but you do have more tools and an unlimited subscription for less then half that cost. I often make 300 scenes of 5 to 10 seconds. That would kind of mean half my credits would be gone on just one video... My credits would be gone after just 2 music videos. Yes the end quality would be better but I have much more stories to tell with 4 to 8 videos a month
the only youtuber who actually spends for proper hands on examples. tired of these other youtubers just riding this wave and reviewing examples thats displayed on the website lol
Appreciate that! That's kind of a guiding principal on the channel, and something of a point of pride for me. And look, I've been guilty of it here and there-- mostly, when the tools aren't actually available. But, if I CAN generate on it, I'll do it. It's the only real way to know if a product is good or not.
had to support you Tim, for being willing to pay out the sora subscription cost in order to do this video - was already subscribed of course, but now i've joined as a member
lol. Its paid at the end of the month and youtube payment covers it. Youre a lemming. Hes not hurting from buying something that alloows him to make more videos and money. Smh. Subscribe because you like content. Not because someone signed up for a subscription. Weird af
I think Minimax is doing far more interesting video, at least in terms of if you want to use AI video to make an actual film. Hadyn Rushworth's recent tests of Runway, Kling, Vidu, and Minimax for portrayal of character emotions from an image to video character was very interesting. Vidu and Minimax were both doing very interesting work.
100% agree, I have often commented this, myself, as someone who is on MiniMax's Unlimited Plan. The Runway Team mostly focus on gimmicks as their upgrades, so far, yet Runway's animation physics is far behind compared to MiniMax, and Runway generally doesn't follow prompts very well. The Runway Team do not think like real filmmakers. MiniMax is far superior, for now.
Agree. None of those are super refined. But they each are capable of producing very interesting, engaging stuff in their own way. And each of them continue to push the envelope, at a reasonable price.
Appreciate that!! Yeah, it also takes a lot of the pressure off as well. When I do sponsored content (and I only do it for the stuff I genuinely use/like) I try to stay as balanced as possible as well. I never like to partner with any company that wants me to outright lie. In fact, I think if you look at every sponsored video I've done, there's always a "but...there's also this..." That said, outright paying for Sora-- well, now there's NO pressure!
@@TheoreticallyMedia It has some interesting features, but I don't feel the quality is much better than what we get from other AI video generators, especially for the 200$ price tag, but thanks for taking one for the team 😄 we will se what people creates with it the next few weeks.
Dude, thank you so much! Haha, and ads, that’s brilliant! Or Product Placement? Hey, why does my “Man in A Blue Business Suit” keep walking past a Starbucks?
In my opinion, Sora looks like a $49-$99 a month subscription. But not $200, maybe if it was 4K with a 45 second timeline with audio included.😅 Thanks for the review though.
Letterboxing has black bars on top and bottom (the image is the shape of a letter-box slot). Pillarboxing has black bars on the sides (the image is the shape of a pillar box, which is a free-standing post box in the shape of a pillar).
Thanks for testing it out! We in the EU area do not have to worry about the cost as it's not available here :D Sora is surprisingly not better at all than the current competition. The extra features that Sora has will be copied to other platforms quite fast if users find them useful.
$200 is beyond steep, especially considering all the other video tools but what I will say is... If you wanted to make a movie and you took the time to generate all the starting frames, then signed up for just 1 month... You could probably knock out your entire movie with one months worth of generations, which suddenly makes the $200 seem like a bargain.
True, although I keep running into that issue of Sora not taking the full first frame. But not always? It’s weird. It’s like sometimes it just decides to head off in its own direction. Which- might be interesting as well.
@@TheoreticallyMedia Yeah, it seems strange to me that we don't really encounter that with some of the less hyped models. I find it curious that OpenAI have included Sora's release so early in the 12 days of OpenAI... I'm wondering if this is the warm up act.
Personally, from what I have seen - I cant say I'm hugely impressed with Sora compared to other AI Video generators. Disclaimer her - That is only based on what I have seen here, on X and Discord - I personally have not used it. Also be very interesting to see the next upgrades to Luma, Kling, Minimax and Runway over next few months and then see how this compares and cost also. Great video Tim - enjoyed watching this one 👍🏻👍🏻
FWIW, I do really appreciate them going the MidJourney route of having a recent feed. You even can follow remixes back through the chain. So you can learn a lot quickly about what it does well and what it doesn't. The UI is definitely still rough though. Like you can filter by a user's generations and bookmark individual videos but can't follow the user (or even get to their other videos directly from a bookmarked video). Sometimes Esc doesn't work, it often puts you back at the top of a list, etc. Still, I agree that the creative tools have some promise.
For sure there is some flow problems. I’ve been running into that lately, where you upload a ref video and then there’s no simple way to get back to the main page or do anything, but x’ing out. Also, the blend controls are….perplexing? That, in all honestly, might be by design. Kind of to encourage experimentation. Still, I’m not going to knock them too hard, it’s a v1 of the site, so I’m sure all that will improve with time. Agreed on that MJ layout. I kind of love it.
I am sitting out on the sidelines till something coherent story-wise can be made out of all the evolving tools, but I do see it all coming together. Thanks.
I’ll test it out more over the next week or so…but yeah, it’s good. “Better” than anything else? I mean…subjective, but I don’t think so either. Shootout incoming for sure.
Sora has great physics in a number of areas, but not so much in other areas. MiniMax follows prompts much better than Sora, overall. Runway and Sora need to get their act together, to be as great as Kling and MiniMax.
When a product is first released, don't rush to try it out. Wait a few days to see others' usage experiences before deciding. After all, $200 is not cheap!
The evolution of Sora is fascinating-it’s like watching a tech protagonist go through a full story arc. From its rocky moments in the public eye to its current state, it’s clear there’s a lot of potential here, even if it’s not fully realized yet. Features like the storyboard and blending tools stand out as game-changers for video generation, offering creative flexibility that’s rare in this space. The initial outputs seem impressive but far from perfect. The tie glitch and character inconsistencies highlight the limitations, though it’s interesting how lighting and scene details can still look so polished. The ability to remix or blend inputs opens up some exciting possibilities for experimentation, even if the results sometimes veer into unpredictable territory. Charging $200 a month feels steep, especially when there are still quirks to iron out, like weird frame glitches and less-than-seamless transitions. But if Sora’s tools continue to evolve and streamline workflows, it could justify the price for professionals who value precision and control in AI video creation. Definitely curious to see how it develops.
I would really love to hear from you about your opinions regarding the tools and the control you have with Sora. Granted it is launch day, the model is the "turbo" version, and they will become better over time; However, the things I struggle the most with in regards to AI tools are not the quality of their outputs but rather the CONTROL, or more specifically, the creative control. How do you feel the Sora tools are in this regard, how much creative control and "director" influence do you feel that you have? Keep up the good work Tim!
Very interesting to see! I'm not a fan of the price set-up for image to video and because its taken so long for it to come out, I believe Kling AI for example is borderline close/slightly better (from a price perspective, consistency, overall video quality and a few other things). But of course Sora is only out today, so time will tell. Great video 😁
Glad you had a chance to play with Sora for a while today! I got home from work and I got a message that they couldn't create new Sora accounts right then, and to check back later. An hour later I was able to log in and get started. I used a Midjourney image of a Christmas themed living room I made for one of my LoFi videos, and rolled a couple 480s so I could get a (relatively speaking) faster output. Ummmm... not impressed at all. Then I rolled a 720p version - although it's much improved over the 480p, Kling 1.5 Pro mode did a more even and realistic job of animating the same image. I then used just the description it interpreted of the picture, and deleted the picture, so now it only had the description with which to generate a video. The resulting video was a different looking room, but the overall vibe was the same, and the video was extremely realistic looking (except for the gaps between the logs in the walls of the log cabin). So that was very impressive. It will be very interesting to spend the next handful of days experimenting to see what types of things work well, and what types of things aren't quite "fully baked" yet.
Kling for one is already really good by now. I can’t access sora yet cause of demand on the servers, but they messed up waiting so long. They could have made a lot more money releasing it earlier.
Nice review Tim, Not feeling $200 a month from me at the moment, I think the Chinese models are pretty damn good right now although Sora is leading the charge, I just don't know how long that will be the case. Hopefully they serve through API soon and we can explore more on a video by video basis
tried that, you still get the "still" image for a second or two, and then Sora heads off in its own direction. BUT, I've got something REALLY crazy coming up with Sora tomorrow. Kinda found something amazing.
Great first look mate, I'm spoilt atm being in several CPP so I wont be doing too much SORA atm but it's good to know its there I already use Openai for a bunch of other things
Midjourney, Magnific, Suno, Udio, and Kling. I'm paying for all of them and suddenly realised, 'This is a creative outlet, and I'm an AI enthusiast.' Long-term hype marketing can backfire, and this is a good example. Who did they think their target market was? $200 a month? Thank you for doing this, Tim. I feel bad you had to. :(
Kudos for spending the $200 Pro fee. $2,200 a year for Sora is crazy work. Sora doesn't seem as groundbreaking as it did a year ago, yet the potential seems limitless. I imagine Minimax, Kling, Runway and the rest won't take this threat lightly, that's good for the Gen AI community. Great video 🍻
Thank you for doing this so we can all see what's to be expected. But, and I'm gonna say this as somebody who didn't dive in all this but with only what I've seen from you and others videos, Sora isn't comming with something ground breaking for 200 per month. Can't wait to see your opinion after you do some more stuff with it, but from what I see it's having the same "style" of generating videos like others and from with all that you tested so far, is a battle of good prompts and finding the best one for your needs. Just my 2 cents, don't mind me :)
So we already had the best video models and just didn’t know it! The tools are intriguing and will hopefully inspire the other providers. I think a useful metric to start considering is, how many usable generations you get, as a percentage. Based on that metric, I think Vidu is leading for me.
I have lots of Ai video tools, but I rarely use any of them anymore. Good for making music videos and that's it! not willing to spend time on multiple retakes swapping between tools etc ( more hobby than work). But Sora seems to have taken what works well from other tools and placed it in one package as in a storyboard. I am sure other features will arrive taking the sting out of the 200. Look forward to future reviews.
I use a combination of locally generated AI photos by first drawing my own outlines often, then generating them in SDXL and if needed going over them or over parts in flux. I was a family photographer so I focus on making heartwarming family videos with KLING and MINIMAX Your prompting has too much pointless fluff in it "the city that has taken so much from him" this is unnecessary, just prompt for an emotion, angle, view, camera position or movement, but you don't need to describe abstract concepts like this to an AI Prompt, it doesn't affect anything other than to take some things literally and misinterpret your prompt. I've seen many people write entire poetry in their prompts and wonder why things don't turn out the way they wanted. If you look at my currently uploaded videos they show as much emotion as KLING would allow, but I never go out of my way to prompt than what's needed.
@@privateprofile3517 No, it's 200$. There are two price models. Plus for $20/month and Pro for $200/month. I copied the prices directly from their site. With Plus you get 1000 credits and with Pro 10,000 credits and can create videos in better resolution and download them without watermarks, for example.
Yeah, the fact that it SOMETIMES decides to head off in its own direction is SUPER weird. Like, there's no way those running dudes in alley should have turned into a Kpop music video, but...there it is.
I haven't heard yet-- in the meantime, I've been using Topaz. Since we're talking about cost, I mean-- I know that's pricey, but at least you buy it once and own it.
@@TheoreticallyMedia 100%!!! I bought Topaz because of a video you made a while back! LOL I use it now too, but when I gen in KLING, I don't have to because its in 1920 already. TOPAZ is amazing though.
@@TheoreticallyMedia Hi Tim, I thought I replied to this, so sorry if you see it twice. YEAP!! I bought TOPAZ because of a video you made a while back showing how well it did with upscaling. Its a great app and does a very nice job. Its just easier when the gen is already in 1920x1080. Keep up the great work!
Overall quality seems modestly better, though I've seen some very mediocre Sora generations, but when you combine that with the ability to direct the action inside of a single cohesive output and I'd say I'm impressed. Quality is going to be the main thing for the general public that wants to use this for fun and short memes but storyboarding/quasi-keyframing is huge in conjunction with the ability to be able to remix and regenerate for anyone trying to use these tools more seriously.
The new UI tools seem quite interesting, especially the storyboard - but outputs seem about as food as the rest, maybe we will see some quick improvements in the next few days though. Interestingly the chatGPT Pro is not just for Sora, it's a whole new tier of membership, presumably for professionals if all types. For examples it includes a beefier version of GPT4o-1 than the plus tier. I think this actually truly is a pro version as in the only people getting it will be those who's employer is paying for it
Haha, if you check out the latest video, I forgot to cancel. So 'ole Sam got an extra $200 out of me. Sigh, I guess I should do another Sora test on the channel, since I've still got it. I am convinced there's something pretty good here. It just isn't what OAI thinks it is.
I think SORA has some great things going for it in terms of image quality, but I'm not sure it's quite there yet in terms of stability. From what I've seen, it's still at the level it was in summer 2024. I think the SORA could use a little extra love and care. I think we can all agree that Kling and Minimax are doing a great job, but there's always room for improvement! Have a wonderful generation, folks! 😊
Based on what I saw in your video, I’d wait a few months before considering paying $200. While the functionalities are promising (when they work), the overall quality of the outputs isn’t particularly impressive yet. Thanks for saving me $200!
To make one music video ( which i do a lot of) i generate hundreds of clips. I have found only Runesy gen3 turbo being the most optimal flow of speed witb its unlimited plan. Otherwise it gets either expensive or boring to wait. You can compromise the quality a bot or substitute some clips with a side kling generation. Works for me.
@TheoreticallyMedia if you need some example shots in different levels of rendered quality to compare how realistic a 3d render must be to work with subtle let me know
Its wild how so many people are saying there are other better options than this! HAHA, this is literally some of the best video generators we have. And its just the start. Its seriously amazing tech. 10/10 - not worth $200 just yet but it will get there for sure.
That’s the thing that keeps me from dogging on it too hard. It’s a bad launch, but it’ll get there…eventually? But man, black eye for AI video for sure
The best ai video generator, with extend function, that I've tried so far is aitubo (model C1). I'm gonna try pika ai 1.5 next with the extend function. I think Runway is still very glitchy as well as Sora. Extend function is by far the most important feature for me, for an ai video generator, because I want consistency so I eventually can make quality short films and music videos.
Good video. If they dropped this when they first teased Sora, I think everyone would be jaw dropping. It does look like they put a lot of polish on the site, but it seems from what you have shown the quality may have not been worth the wait and price point. Personally I think this $200/month price point with the offering is hurting them. I think what really hurts is that there are better models locked behind the $200 price point. So it's like well I can't even test them. Maybe if they gave a small taste of the pro models, I don't know, just makes me not interested.
Those glitches look very much like 'optical flow' - which is used to create slow motion - I would imagine they're running everything through a process to slow mo it as. well as the generative stuff
@@TheoreticallyMedia yes I guess to make processing faster, they're generating it at normal speed and then post processing (using something along the lines of Topaz or an internal slow motion based on optical flow) that file to slow it down - so they could theoretically slow it down by 25, 50% etc Optical flow does create artifacts when more abrupt transitions happen in the scene such as lighting changes etc
This is exciting! Thanks for reviewing. Still... No easy workflow to add my products on those videos. It is getting there with multiple tools but to insert a can of beer looking like the same can of beer... not there yet.
Adding Sora to O1 Pro mode makes the $200 more appealing, but definitely needs to get better... but my original ChatGPT sub from a year ago has o1 now so it definitely could be worth it one day soon. I am more excited for local video models though.
6:46 Your guy became Ted Lasso. By the way, did you try character consistency throughout several shots? In my opinion, the 20 seconds alone, if you can achieve consistency, worth more than $200.
I typed in versions of all the Sora examples on the website into Hailou and it pretty much created perfect versions of them. Open AI have shot themselves in the foot here. First they were late to the party, then when they did turn up they ask for double what everyone else wanted for pretty much the same thing. Not to mention half the world can't even sign up to it. Sora is dead in the water before it even began to learn to swim for me.
YES!!! MiniMax is the KING, for now, which is why I'm on their Unlimited Plan. And Runway needs to get their act together; their videos are basically slow motion and do not look very cinematic.
@@FilmSpook Spot on. I actually got a refund for the Unlimited Runway and went over to Hailou instead. After going Unlimited Runway throttled me down to just one generation which I thought for $100+ per month shouldn't be happening. The outputs had very little movement compared to Hailou. Glad I switched.
The features like timeline and such can be easily copied. I say this as a SW dev. Some of the other stuff like remix and things that actually use the generative AI as opposed to a well done UX could make it compelling for others. The $200 price point is... interesting because it (and the gp4 reasoning model) seem to be incurring a much higher operating cost thats being passed down to the customers with questionable value. The videos you showed almost strike me as nothing special considering what ive and almost certainly you, someone who actually tracks this , has seen. Its a very interesting time and I wonder when we will start to see consolidation because I can imagine there is a large market for a $200/month subscription. Maybe im taking the wrong impression away, but seeing a-la-carte 'credits' popping up as an alternate to subscription seems to indicate some at least understand that the wider consumer market is more price sensitive. Im sure someone who understand econ better will chime in lol. Thanks for the video!
I knew you were a 1 percenter by all the guitars. One of my friends left his guitar in my living room once and it looked really good. $200 is a lot of food.
Haha. Not a 1%’er at all! One of those guitars is my “nice” one and was about $700 when I bought it. The other was actually a good ole fashioned Facebook marketplace score! Admittedly, that once was a steal. I got it for $200 and it resells for around $550! Funny enough on the food part: man, my kids talked me into McDonalds the other day, eating fairly simple (Burger, they got nuggets, and fries/drink) cost $40! That’s almost one week of Sora! Haha.
Been using it for a week now. For 200 dollars, the sofware is very charasmatic. You get close to your idea, and then it randomly generates a completely original scene. Sora is very promising. I would pay 200 a month for a working product
Google has Veo out there....somewhere. But, it is Google, so they've likely forgotten they have it. I don't know if that old reality show Hoarders is still on the air, but if it is, we should send them over to Google and see if they can't find Veo and a whole slew of other forgotten projects.
Thanks for your analysis. I think that right now, AI video tools like this that rely on "projecting" each frame independently, makes their results inherently unstable and hard to direct. I wont be surprised if, in the future, the winning solution will come from professional 3D software like Cinema 4D enhanced with an AI front end that can instantly build entire 3D environments and fully rigged characters, texture maps and lighting. Instead of trying to predict each frame, you will quickly create a fully modeled scene that’s ready for traditional direction, animation, and rendering, giving the director both speed and rock-solid control. This shift means today’s frame-by-frame solutions are just a stepping stone toward more robust, AI-enhanced 3D creation. Who knows, a great time to be alive!
The thing is if it’s unlimited generations for 200 bucks and it’s just as good or better than Kling then it’s probably worth it for us perfectionists out there. Even with a Hailou subscription I still use Kling for the stuff I’d actually put my name on and those inspiration points ad up…
I don't think the video was too disjointed - I thought it was great, as usual! But gosh, spending $200 per month is quite a lot of money, so thanks for being the guinea pig for the rest of us, Tim!!
Editor Tim appreciates that!! On-Camera Tim was a total mess, but Editor Tim managed to do a LOT of cutting to make it seem like this was a lot less rambling! Editor Tim hates On-Camera Tim. They fight every day.
I think Sora is looking great. When I can bring my stories to life with it in a consistent way and throw in audio then I might be willing to pay for the 200/month.
My account says, 'Your account does not currently support creating videos with uploaded media containing people,' instead of mentioning my "plan". Does this mean I wouldn’t have the option even if I upgraded?
The reality seems to be plain and simple here: if you pay $200/month you expect a PROFESSIONAL tool, to do professional things. But instead, SORA is delivering a (very, very, I need to say) nice toy. The model looks amazing, but has a really bad ROI, so the skyrocketing cost makes it almost worthless. I'm still expecting NVIDIA to provide fixed point math on its chips, so we can stop drying rivers just to run AI and drop the costs of filmmaking in general. Maybe this day SORA can be viable. Thanks to take the hit for us, dude. You deserve a new YT plate.
Is it possible to use sora for subtile changes? Like if I would film a scene with a bad lense and sora changes the footage to look actually professional without adding any big changes of the actual scene? This are the changes that could revolutionise the market. No need to rent expensive equiptment for family events. Those hyper hollywood goals arent anything I will use in the near future.
Title should be, "I paid $200 for Sora so you don't have to".
"don't give these people your money" that doesnt sound smart business wise
My exact thoughts.
Haha, I tried to remain as neutral and objective as possible, but at the end of the day: whatever your takeaway, I’m happy to have taken the hit for everyone!
Lmao facts well this can be extremely useful for people that do content and marketing .
@@TheoreticallyMediaI could have avoided wasting time and money, by watching your video first!!!
I even commented a few weeks ago that you are the most reliable source of info when it comes to MUSIC/IMAGES/VIDEO generative AI.
I was too excited and curious... Despite one very good innovation, the Storyboard, SORA is a complete disaster!
But at least I know now that OpenAI SORA will never make it... WHY?
Keep in mind the time they took to get to THAT!
Then the fact that we get errors in almost all generations...
3 hands humans! Objects disappearing! All kinds of inconsistency during short videos! Improper object movements...
Now, just extrapolate the time and money it will take to correct all the flaws!
During the same period of time with a fraction of the budget, KLING will be KING!
OpenAI was already facing bankruptcy due to their precarious financial situation.
It may be the last nail on the coffin!
Honestly, I wouldn't pay $200 a month for something like this! There are still too many flaws!
agreed. but maybe ill try to get something out of it. runway gen 3 is a much better value at half the cost
@@gu9838 It's a shame Kling AI doesn't offer unlimited subscriptions like RunWay or Minimax...
it will be jw censored shit
Gen 3 is trash mate y u still using it, its well behind minimax and kling @@gu9838
Still If they want only $20 a month,I would not pay ...
Long time subscriber, Tim...as you know. So best i can do is a tip. 🤜🏽🤛🏽
I appreciate that so much, John!! You’re the best!
Its crazy they waited this long to just release a product that's "meh" compared to everything else we now have.
For sure that delay did them no favors. But, I will say: pre-election, the media would have been ALL over them for releasing something so “dangerous”
…even though there are plenty of other tools that do the same thing, and we barely saw anything on that front.
In some ways, I do feel bad for OpenAI.
….uh, apparently enough that I gave them $200? Sigh.
Yeah, lol, I bet they saw all the other Ais releasing, and they probably thought, "Oh no, we are getting out matched. We better release it now!" 😭😂😂
Thats probably why they waited lol
After your testing is it really only meh? To me it seems like it’s the highest quality video gen available now
@@zrakonthekrakon494 dunno, I tried it yesterday and honestly waiting 10 minutes for a base request just to end up with something in best case that doesn't do what I wrote at all, in worst case is weird and scary... Nah. It's like a hit or miss, like 5% hit 95% miss
Wow! Unless you find the 'we're just kidding' button, I won't be paying for this. I get much better results from Kling and Minimax! I do, however, appreciate you taking the bullet for us.
I’ve been saying for awhile, Sora is not going to be what everyone thought it was. Not to say it isn’t good- I’ll spend more time with it and I’m sure I’ll get some great results out of it- but “better” than anything else? That’s totally subjective
@@TheoreticallyMedia IMO when you use image2video, the image you provided should be the 1st frame. Not "incluenced by". That's the only way to get any accuracy to what you want (for example trying to make a short film, consisting of small clips), in the current AI video scene. Everything else is just a guessing machine. This makes Minimax superior imo.
Because AI images can be handcrafted to exactly what you want (with inpainting).
However what I've noticed with Minimax, is that image2video works best with real photos. If the AI image has even small errors in anatomy or other weirdness to it, it will multiply in video, and throw it off even more. Probably because Minimax was trained on real videos, so it understands real photos as reference points better. Or, AI images that are flawlessly realistic
@@TheoreticallyMedia I'll say it, it isn't good.
I was tired of the channels who are just trying to ride a wave...this one really feels good and actually feels like did not waste my time, thank you for Good authentic info
In other words.... Open AI spent almost a whole year to come up with a tool that's basically still barebones and honestly subpar to any other AI out there now... great
To be totally honest, I was expecting it to be a lot more nerfed. So, in some ways this is kind of a win.
I really thought public Sora was going to be DallE but Video.
pretty much. sighs. what a let down.
@@gu9838 and will be censored crap
Which AIs do you think are superior?
@Screamer99REVIEWS Sora is more advanced in many areas of physics, but subpar in some other areas, but yes, 200 is far too much for what they are offering now. The Sora Team do not live in reality.
Thanks for taking the hit, Tim. Sora is way late to the game. The Chinese are killing it. Sora adding copyright control and $200.00, forget about it. Love OpenAI, but they dropped the ball on this. I don't see this being the dominant gen.
I can't use OpenAI without a cell phone, but yeah $200 is where companies get disconnected with their worth. I just paid $60 for a year of Kling in their half price sale, and it a far better value.
@@DataJugglerKling and Minimax are both amazing products.
Ft. Pay up for using (c) material. Artists need to get paid. Period.
A greedz, 200 is waaaaay too much, Runway isn't as good but you do have more tools and an unlimited subscription for less then half that cost. I often make 300 scenes of 5 to 10 seconds. That would kind of mean half my credits would be gone on just one video... My credits would be gone after just 2 music videos. Yes the end quality would be better but I have much more stories to tell with 4 to 8 videos a month
100%. the price alone is ridiculous, lots of prompt limitations and copyright.
the only youtuber who actually spends for proper hands on examples. tired of these other youtubers just riding this wave and reviewing examples thats displayed on the website lol
Appreciate that! That's kind of a guiding principal on the channel, and something of a point of pride for me. And look, I've been guilty of it here and there-- mostly, when the tools aren't actually available. But, if I CAN generate on it, I'll do it. It's the only real way to know if a product is good or not.
had to support you Tim, for being willing to pay out the sora subscription cost in order to do this video - was already subscribed of course, but now i've joined as a member
Ahhh-- appreciate that, and all the support all along!! I need to do more w/ that membership tier! Bug me about that for sure!
lol. Its paid at the end of the month and youtube payment covers it. Youre a lemming. Hes not hurting from buying something that alloows him to make more videos and money. Smh. Subscribe because you like content. Not because someone signed up for a subscription. Weird af
I think Minimax is doing far more interesting video, at least in terms of if you want to use AI video to make an actual film. Hadyn Rushworth's recent tests of Runway, Kling, Vidu, and Minimax for portrayal of character emotions from an image to video character was very interesting. Vidu and Minimax were both doing very interesting work.
100% agree, I have often commented this, myself, as someone who is on MiniMax's Unlimited Plan. The Runway Team mostly focus on gimmicks as their upgrades, so far, yet Runway's animation physics is far behind compared to MiniMax, and Runway generally doesn't follow prompts very well. The Runway Team do not think like real filmmakers. MiniMax is far superior, for now.
Agree. None of those are super refined. But they each are capable of producing very interesting, engaging stuff in their own way. And each of them continue to push the envelope, at a reasonable price.
I tried looking up Hadyn Rushmore on RUclips with no success, are they on another platform? Thanks
I only use vidu now
Thanks for rundown, Tim! Sure shows your dedication to provide great content by investing the money for an honest review.
Appreciate that!! Yeah, it also takes a lot of the pressure off as well. When I do sponsored content (and I only do it for the stuff I genuinely use/like) I try to stay as balanced as possible as well. I never like to partner with any company that wants me to outright lie. In fact, I think if you look at every sponsored video I've done, there's always a "but...there's also this..."
That said, outright paying for Sora-- well, now there's NO pressure!
Man Open AI is going full corporate greedy :D
Yeah, I'll see what I can do for $200 in the next month, but likely churning this sub by January!
I mean, they're still losing a lot of money only charging $200...
@@TheoreticallyMedia It has some interesting features, but I don't feel the quality is much better than what we get from other AI video generators, especially for the 200$ price tag, but thanks for taking one for the team 😄 we will se what people creates with it the next few weeks.
Jews
It is probably not about that, but because this model is really expensive (at energy level) to run. Doesn't look optimized.
Can't believe we got Sora before GTA 6 🤣
At this point, we'll have the holodeck before we get GTA 6! (Or at least the next Elder Scrolls!)
ASI before GTA 6
heck, we might even get quantum computing CPUs before gta 6
we will have AGI before GTA6
Thanks for showing the PRO Version. 200 a month is a bit high for small indies like Me. Maybe they could do ADs to bring the price down?
Omg they could put ads in the outputs! Noooooooooo!!!
@@jac001 LOL
Dude, thank you so much! Haha, and ads, that’s brilliant! Or Product Placement? Hey, why does my “Man in A Blue Business Suit” keep walking past a Starbucks?
@@TheoreticallyMedia I imagine them reading the comment, implementing it, and then regretting this thread lol 😂😂
@@jac001 LOL
Thanks Tim - liking and commenting for your $200 - one thing to note for some of us: not available in the EU or UK : (
In my opinion, Sora looks like a $49-$99 a month subscription. But not $200, maybe if it was 4K with a 45 second timeline with audio included.😅 Thanks for the review though.
I agree with you
Letterboxing has black bars on top and bottom (the image is the shape of a letter-box slot).
Pillarboxing has black bars on the sides (the image is the shape of a pillar box, which is a free-standing post box in the shape of a pillar).
Appreciate that! Long day, and a later shoot than I normally do. Clearly the afternoon slump had kicked in!
@@TheoreticallyMedia No big deal. Appreciate your dropping the $200 so we can all see what the fuss is about.
Thanks for testing it out! We in the EU area do not have to worry about the cost as it's not available here :D Sora is surprisingly not better at all than the current competition. The extra features that Sora has will be copied to other platforms quite fast if users find them useful.
$200 is beyond steep, especially considering all the other video tools but what I will say is... If you wanted to make a movie and you took the time to generate all the starting frames, then signed up for just 1 month... You could probably knock out your entire movie with one months worth of generations, which suddenly makes the $200 seem like a bargain.
True, although I keep running into that issue of Sora not taking the full first frame. But not always? It’s weird.
It’s like sometimes it just decides to head off in its own direction. Which- might be interesting as well.
@@TheoreticallyMedia Yeah, it seems strange to me that we don't really encounter that with some of the less hyped models. I find it curious that OpenAI have included Sora's release so early in the 12 days of OpenAI... I'm wondering if this is the warm up act.
2:32 Liked the video just tor that.
Thanks for taking one for the team buddy 😂
Someone had to do it!
Personally, from what I have seen - I cant say I'm hugely impressed with Sora compared to other AI Video generators. Disclaimer her - That is only based on what I have seen here, on X and Discord - I personally have not used it.
Also be very interesting to see the next upgrades to Luma, Kling, Minimax and Runway over next few months and then see how this compares and cost also.
Great video Tim - enjoyed watching this one 👍🏻👍🏻
FWIW, I do really appreciate them going the MidJourney route of having a recent feed. You even can follow remixes back through the chain. So you can learn a lot quickly about what it does well and what it doesn't. The UI is definitely still rough though. Like you can filter by a user's generations and bookmark individual videos but can't follow the user (or even get to their other videos directly from a bookmarked video). Sometimes Esc doesn't work, it often puts you back at the top of a list, etc. Still, I agree that the creative tools have some promise.
For sure there is some flow problems. I’ve been running into that lately, where you upload a ref video and then there’s no simple way to get back to the main page or do anything, but x’ing out.
Also, the blend controls are….perplexing? That, in all honestly, might be by design. Kind of to encourage experimentation.
Still, I’m not going to knock them too hard, it’s a v1 of the site, so I’m sure all that will improve with time. Agreed on that MJ layout. I kind of love it.
I am sitting out on the sidelines till something coherent story-wise can be made out of all the evolving tools, but I do see it all coming together. Thanks.
Thanks for reviewing, Tim. 😊 $200 pm to fund Open AI's half-baked, dysfunctional software. WOW! Where do I sign up?
Thanks for the video!
Watermark and no human image input for Plus version. Like, huh?
I think you can remove watermark? I’ll double check for the full review! Pretty sure I saw that option!
@@TheoreticallyMedia Watermark can only be removed in the pro plan so far
The storyboard feature is cool, but I'm not seeing outputs that are better than Minimax.
I’ll test it out more over the next week or so…but yeah, it’s good. “Better” than anything else? I mean…subjective, but I don’t think so either.
Shootout incoming for sure.
Sora has great physics in a number of areas, but not so much in other areas. MiniMax follows prompts much better than Sora, overall. Runway and Sora need to get their act together, to be as great as Kling and MiniMax.
Love the well timed hard cut after "my name is.."
When a product is first released, don't rush to try it out. Wait a few days to see others' usage experiences before deciding. After all, $200 is not cheap!
The evolution of Sora is fascinating-it’s like watching a tech protagonist go through a full story arc. From its rocky moments in the public eye to its current state, it’s clear there’s a lot of potential here, even if it’s not fully realized yet. Features like the storyboard and blending tools stand out as game-changers for video generation, offering creative flexibility that’s rare in this space.
The initial outputs seem impressive but far from perfect. The tie glitch and character inconsistencies highlight the limitations, though it’s interesting how lighting and scene details can still look so polished. The ability to remix or blend inputs opens up some exciting possibilities for experimentation, even if the results sometimes veer into unpredictable territory.
Charging $200 a month feels steep, especially when there are still quirks to iron out, like weird frame glitches and less-than-seamless transitions. But if Sora’s tools continue to evolve and streamline workflows, it could justify the price for professionals who value precision and control in AI video creation. Definitely curious to see how it develops.
I would really love to hear from you about your opinions regarding the tools and the control you have with Sora. Granted it is launch day, the model is the "turbo" version, and they will become better over time; However, the things I struggle the most with in regards to AI tools are not the quality of their outputs but rather the CONTROL, or more specifically, the creative control.
How do you feel the Sora tools are in this regard, how much creative control and "director" influence do you feel that you have? Keep up the good work Tim!
Very interesting to see! I'm not a fan of the price set-up for image to video and because its taken so long for it to come out, I believe Kling AI for example is borderline close/slightly better (from a price perspective, consistency, overall video quality and a few other things). But of course Sora is only out today, so time will tell. Great video 😁
kling is far better bec you know already sora will be censored rubbish western lib crap
Glad you had a chance to play with Sora for a while today!
I got home from work and I got a message that they couldn't create new Sora accounts right then, and to check back later. An hour later I was able to log in and get started. I used a Midjourney image of a Christmas themed living room I made for one of my LoFi videos, and rolled a couple 480s so I could get a (relatively speaking) faster output.
Ummmm... not impressed at all. Then I rolled a 720p version - although it's much improved over the 480p, Kling 1.5 Pro mode did a more even and realistic job of animating the same image.
I then used just the description it interpreted of the picture, and deleted the picture, so now it only had the description with which to generate a video. The resulting video was a different looking room, but the overall vibe was the same, and the video was extremely realistic looking (except for the gaps between the logs in the walls of the log cabin). So that was very impressive.
It will be very interesting to spend the next handful of days experimenting to see what types of things work well, and what types of things aren't quite "fully baked" yet.
Kling for one is already really good by now. I can’t access sora yet cause of demand on the servers, but they messed up waiting so long. They could have made a lot more money releasing it earlier.
Nice review Tim,
Not feeling $200 a month from me at the moment, I think the Chinese models are pretty damn good right now although Sora is leading the charge, I just don't know how long that will be the case.
Hopefully they serve through API soon and we can explore more on a video by video basis
Still consistency issues and soo many restrictions make this not worth the price.
It’s okay to admit that Sora isn’t even close to the top 3 AI video models and self-censored to the point of being useless for sAfeTY
I believe the image to video feature is more precise if you added the image in storyboard rather than uploading as a reference image
tried that, you still get the "still" image for a second or two, and then Sora heads off in its own direction. BUT, I've got something REALLY crazy coming up with Sora tomorrow. Kinda found something amazing.
Great first look mate, I'm spoilt atm being in several CPP so I wont be doing too much SORA atm but it's good to know its there I already use Openai for a bunch of other things
It's way worse than I expected given how long it took to be out.
Did you type in the name ,"pirates of the Caribbean" when promtin in sora.
Or did you type in "a pirate" ?
So its still strict in 200 dollar mode?
I think I just said a pirate runs from a crowd. It fully recognizes the image as Capt Jack.
Thanks for the breakdown , quality as always
Midjourney, Magnific, Suno, Udio, and Kling. I'm paying for all of them and suddenly realised, 'This is a creative outlet, and I'm an AI enthusiast.' Long-term hype marketing can backfire, and this is a good example. Who did they think their target market was? $200 a month? Thank you for doing this, Tim. I feel bad you had to. :(
Kudos for spending the $200 Pro fee. $2,200 a year for Sora is crazy work. Sora doesn't seem as groundbreaking as it did a year ago, yet the potential seems limitless. I imagine Minimax, Kling, Runway and the rest won't take this threat lightly, that's good for the Gen AI community. Great video 🍻
Thank you for doing this so we can all see what's to be expected. But, and I'm gonna say this as somebody who didn't dive in all this but with only what I've seen from you and others videos, Sora isn't comming with something ground breaking for 200 per month. Can't wait to see your opinion after you do some more stuff with it, but from what I see it's having the same "style" of generating videos like others and from with all that you tested so far, is a battle of good prompts and finding the best one for your needs. Just my 2 cents, don't mind me :)
So we already had the best video models and just didn’t know it!
The tools are intriguing and will hopefully inspire the other providers.
I think a useful metric to start considering is, how many usable generations you get, as a percentage.
Based on that metric, I think Vidu is leading for me.
I have lots of Ai video tools, but I rarely use any of them anymore. Good for making music videos and that's it! not willing to spend time on multiple retakes swapping between tools etc ( more hobby than work). But Sora seems to have taken what works well from other tools and placed it in one package as in a storyboard. I am sure other features will arrive taking the sting out of the 200. Look forward to future reviews.
I use a combination of locally generated AI photos by first drawing my own outlines often, then generating them in SDXL and if needed going over them or over parts in flux.
I was a family photographer so I focus on making heartwarming family videos with KLING and MINIMAX
Your prompting has too much pointless fluff in it
"the city that has taken so much from him" this is unnecessary, just prompt for an emotion, angle, view, camera position or movement, but you don't need to describe abstract concepts like this to an AI Prompt, it doesn't affect anything other than to take some things literally and misinterpret your prompt. I've seen many people write entire poetry in their prompts and wonder why things don't turn out the way they wanted.
If you look at my currently uploaded videos they show as much emotion as KLING would allow, but I never go out of my way to prompt than what's needed.
Idk if it’s worth $200. When I saw that title I said ‘nooooo!!!!’ Sorry, Tim! Thanks for the timestamps though!!! 😂
It's 20$ , you fell for the click bait
@@privateprofile3517 No, it's 200$. There are two price models. Plus for $20/month and Pro for $200/month. I copied the prices directly from their site. With Plus you get 1000 credits and with Pro 10,000 credits and can create videos in better resolution and download them without watermarks, for example.
@@privateprofile3517 As is described in the video, the $20 tier doesn't allow you to generate people.
Sora’s img2vid model reminds me of the allucinations you get out of the img2vid model in Dream Machine. Hopefully we will get an improved update soon.
Yeah, the fact that it SOMETIMES decides to head off in its own direction is SUPER weird. Like, there's no way those running dudes in alley should have turned into a Kpop music video, but...there it is.
For Now - I'm sticking with Kling and Runway Gen3 Alpha. Tim, any word on when Runway will output 1920 x 1080? Thank you for testing this for us.
I haven't heard yet-- in the meantime, I've been using Topaz. Since we're talking about cost, I mean-- I know that's pricey, but at least you buy it once and own it.
@@TheoreticallyMedia 100%!!! I bought Topaz because of a video you made a while back! LOL I use it now too, but when I gen in KLING, I don't have to because its in 1920 already. TOPAZ is amazing though.
@@TheoreticallyMedia Hi Tim, I thought I replied to this, so sorry if you see it twice. YEAP!! I bought TOPAZ because of a video you made a while back showing how well it did with upscaling. Its a great app and does a very nice job. Its just easier when the gen is already in 1920x1080. Keep up the great work!
Overall quality seems modestly better, though I've seen some very mediocre Sora generations, but when you combine that with the ability to direct the action inside of a single cohesive output and I'd say I'm impressed. Quality is going to be the main thing for the general public that wants to use this for fun and short memes but storyboarding/quasi-keyframing is huge in conjunction with the ability to be able to remix and regenerate for anyone trying to use these tools more seriously.
Tim, thanks for taking the hit for us all! And no, it is not worth it. $200/m is just CRAZY!
The new UI tools seem quite interesting, especially the storyboard - but outputs seem about as food as the rest, maybe we will see some quick improvements in the next few days though.
Interestingly the chatGPT Pro is not just for Sora, it's a whole new tier of membership, presumably for professionals if all types. For examples it includes a beefier version of GPT4o-1 than the plus tier. I think this actually truly is a pro version as in the only people getting it will be those who's employer is paying for it
this video should be called "not all heroes wear capes"
Haha, if you check out the latest video, I forgot to cancel. So 'ole Sam got an extra $200 out of me. Sigh, I guess I should do another Sora test on the channel, since I've still got it.
I am convinced there's something pretty good here. It just isn't what OAI thinks it is.
I think SORA has some great things going for it in terms of image quality, but I'm not sure it's quite there yet in terms of stability. From what I've seen, it's still at the level it was in summer 2024.
I think the SORA could use a little extra love and care. I think we can all agree that Kling and Minimax are doing a great job, but there's always room for improvement!
Have a wonderful generation, folks! 😊
Based on what I saw in your video, I’d wait a few months before considering paying $200. While the functionalities are promising (when they work), the overall quality of the outputs isn’t particularly impressive yet. Thanks for saving me $200!
To make one music video ( which i do a lot of) i generate hundreds of clips. I have found only Runesy gen3 turbo being the most optimal flow of speed witb its unlimited plan. Otherwise it gets either expensive or boring to wait. You can compromise the quality a bot or substitute some clips with a side kling generation. Works for me.
It Would be really interesting to see if subtle remix on 3d animated cgi scenes could make the entire scene photorealistic without changing too much
I’ll give it a shot and play around with that!
@TheoreticallyMedia if you need some example shots in different levels of rendered quality to compare how realistic a 3d render must be to work with subtle let me know
Its wild how so many people are saying there are other better options than this! HAHA, this is literally some of the best video generators we have. And its just the start. Its seriously amazing tech. 10/10 - not worth $200 just yet but it will get there for sure.
That’s the thing that keeps me from dogging on it too hard. It’s a bad launch, but it’ll get there…eventually?
But man, black eye for AI video for sure
The best ai video generator, with extend function, that I've tried so far is aitubo (model C1). I'm gonna try pika ai 1.5 next with the extend function. I think Runway is still very glitchy as well as Sora. Extend function is by far the most important feature for me, for an ai video generator, because I want consistency so I eventually can make quality short films and music videos.
You totally need to do a livestream with viewer donations to run their prompts you'll get a few months (maybe years but baby steps) worth in no time.
Hi, Tim. This corridor scene is hillarious. Sora should pay me $200 to watch this.
You made a great sacrifice to show us your masterpieces in Sora :) Thank you for that! But as for me, I'll save $200 for Christmas presents ✨
Please do! Get something nice for someone and get a little something for yourself!!
As for me: I guess I generate some presents up? Haha
Good video. If they dropped this when they first teased Sora, I think everyone would be jaw dropping. It does look like they put a lot of polish on the site, but it seems from what you have shown the quality may have not been worth the wait and price point.
Personally I think this $200/month price point with the offering is hurting them. I think what really hurts is that there are better models locked behind the $200 price point. So it's like well I can't even test them. Maybe if they gave a small taste of the pro models, I don't know, just makes me not interested.
Those glitches look very much like 'optical flow' - which is used to create slow motion - I would imagine they're running everything through a process to slow mo it as. well as the generative stuff
ahhhhh, interesting!! Hmm, I wonder if at some point we'll see speed control in that case. That might be handy.
@@TheoreticallyMedia yes I guess to make processing faster, they're generating it at normal speed and then post processing (using something along the lines of Topaz or an internal slow motion based on optical flow) that file to slow it down - so they could theoretically slow it down by 25, 50% etc Optical flow does create artifacts when more abrupt transitions happen in the scene such as lighting changes etc
I would love to see Tim as Tony soprano. It just suits him. He always says something gangster at the end of his videos.
This is exciting! Thanks for reviewing. Still... No easy workflow to add my products on those videos. It is getting there with multiple tools but to insert a can of beer looking like the same can of beer... not there yet.
Adding Sora to O1 Pro mode makes the $200 more appealing, but definitely needs to get better... but my original ChatGPT sub from a year ago has o1 now so it definitely could be worth it one day soon. I am more excited for local video models though.
6:46 Your guy became Ted Lasso. By the way, did you try character consistency throughout several shots? In my opinion, the 20 seconds alone, if you can achieve consistency, worth more than $200.
I typed in versions of all the Sora examples on the website into Hailou and it pretty much created perfect versions of them. Open AI have shot themselves in the foot here. First they were late to the party, then when they did turn up they ask for double what everyone else wanted for pretty much the same thing. Not to mention half the world can't even sign up to it. Sora is dead in the water before it even began to learn to swim for me.
YES!!! MiniMax is the KING, for now, which is why I'm on their Unlimited Plan. And Runway needs to get their act together; their videos are basically slow motion and do not look very cinematic.
@@FilmSpook Spot on. I actually got a refund for the Unlimited Runway and went over to Hailou instead. After going Unlimited Runway throttled me down to just one generation which I thought for $100+ per month shouldn't be happening. The outputs had very little movement compared to Hailou. Glad I switched.
@@FilmSpook Indeed! Minimax is KING!
will you make a comparison video with kling, runway, sora, and open source?
The features like timeline and such can be easily copied. I say this as a SW dev.
Some of the other stuff like remix and things that actually use the generative AI as opposed to a well done UX could make it compelling for others.
The $200 price point is... interesting because it (and the gp4 reasoning model) seem to be incurring a much higher operating cost thats being passed down to the customers with questionable value. The videos you showed almost strike me as nothing special considering what ive and almost certainly you, someone who actually tracks this , has seen.
Its a very interesting time and I wonder when we will start to see consolidation because I can imagine there is a large market for a $200/month subscription. Maybe im taking the wrong impression away, but seeing a-la-carte 'credits' popping up as an alternate to subscription seems to indicate some at least understand that the wider consumer market is more price sensitive.
Im sure someone who understand econ better will chime in lol.
Thanks for the video!
I knew you were a 1 percenter by all the guitars. One of my friends left his guitar in my living room once and it looked really good. $200 is a lot of food.
Haha. Not a 1%’er at all! One of those guitars is my “nice” one and was about $700 when I bought it. The other was actually a good ole fashioned Facebook marketplace score! Admittedly, that once was a steal. I got it for $200 and it resells for around $550!
Funny enough on the food part: man, my kids talked me into McDonalds the other day, eating fairly simple (Burger, they got nuggets, and fries/drink) cost $40!
That’s almost one week of Sora! Haha.
Very cool, but i wont touch it until image generation can be coherent enough to follow dirrections without hallucinating
Been using it for a week now. For 200 dollars, the sofware is very charasmatic. You get close to your idea, and then it randomly generates a completely original scene. Sora is very promising. I would pay 200 a month for a working product
200 a month to help train their system? For a product you can’t commercialise? BONKERS
I liked the video and watched the full to show my support and thanks for putting this together for us.
I have a better idea : RUclips now offers free video generations
Google has Veo out there....somewhere. But, it is Google, so they've likely forgotten they have it.
I don't know if that old reality show Hoarders is still on the air, but if it is, we should send them over to Google and see if they can't find Veo and a whole slew of other forgotten projects.
Thanks for your analysis. I think that right now, AI video tools like this that rely on "projecting" each frame independently, makes their results inherently unstable and hard to direct. I wont be surprised if, in the future, the winning solution will come from professional 3D software like Cinema 4D enhanced with an AI front end that can instantly build entire 3D environments and fully rigged characters, texture maps and lighting.
Instead of trying to predict each frame, you will quickly create a fully modeled scene that’s ready for traditional direction, animation, and rendering, giving the director both speed and rock-solid control. This shift means today’s frame-by-frame solutions are just a stepping stone toward more robust, AI-enhanced 3D creation. Who knows, a great time to be alive!
I'll use it occasionally (on the $20 plan) but I'm still sticking with Kling, MiniMax, and Runway for my short films.
The thing is if it’s unlimited generations for 200 bucks and it’s just as good or better than Kling then it’s probably worth it for us perfectionists out there. Even with a Hailou subscription I still use Kling for the stuff I’d actually put my name on and those inspiration points ad up…
200 USD monthly! If I had that money I could make my own movie in a year!
Great videos - how long do you think it be until we can have audio and lip synching all wrapped into one result?
I don't think the video was too disjointed - I thought it was great, as usual! But gosh, spending $200 per month is quite a lot of money, so thanks for being the guinea pig for the rest of us, Tim!!
Editor Tim appreciates that!! On-Camera Tim was a total mess, but Editor Tim managed to do a LOT of cutting to make it seem like this was a lot less rambling!
Editor Tim hates On-Camera Tim. They fight every day.
@@TheoreticallyMedia 🤣🤣🤣
I do suspect that the competition will add a storyboard feature in response to the release of SORA.
Thanks! What´s is the best free A.I. to upgrade these videos to 4K?
You inadvertently recreated the Techno Viking
Ha! A legend of Early Internet!! I love that guy. I wonder if he’s got ‘lil Vikings busting their moves behind him now.
the quality is there 720p and 1080p. But too much bug which often makes the videos unusable.
I hope the next version will arrive soon
Yeah, for sure-- as the old saying goes: this is the worst it'll ever be.
I think Sora is looking great. When I can bring my stories to life with it in a consistent way and throw in audio then I might be willing to pay for the 200/month.
My account says, 'Your account does not currently support creating videos with uploaded media containing people,' instead of mentioning my "plan". Does this mean I wouldn’t have the option even if I upgraded?
The reality seems to be plain and simple here: if you pay $200/month you expect a PROFESSIONAL tool, to do professional things. But instead, SORA is delivering a (very, very, I need to say) nice toy. The model looks amazing, but has a really bad ROI, so the skyrocketing cost makes it almost worthless.
I'm still expecting NVIDIA to provide fixed point math on its chips, so we can stop drying rivers just to run AI and drop the costs of filmmaking in general. Maybe this day SORA can be viable.
Thanks to take the hit for us, dude. You deserve a new YT plate.
Will they have discounts if you pay by the year?
great question-- looking at my plan: Ha, it does not. No annual plan available!
Taking a hit to his bank account so he can deliver a more accurate view? Truly a man of the people.
Appreciate that!!
I've been trying to test for two hours and nothing works at this time. What's going on?
Hug of death. They’re getting slammed right now. I’d say give it a day or two to let the dust settle and things should go a lot smoother
This is very crazy 🤦🏻♂️.
Hmm... seems Sora has paused all new account creations :(
Hug of Death I presume. If I recall, the same thing happened when GPT4 came out. You’d think they’d have learned a lesson!
Can't believe we got Sora before The Elder Scrolls 6!
Haha, ES6 generated in Sora!! You know Todd Howard is looking into it right now!
Is it possible to use sora for subtile changes?
Like if I would film a scene with a bad lense and sora changes the footage to look actually professional without adding any big changes of the actual scene?
This are the changes that could revolutionise the market. No need to rent expensive equiptment for family events.
Those hyper hollywood goals arent anything I will use in the near future.
Whats the best text to video model right now?
10:30 i think that is almost exactly the image found in the most recent edition of Blade Runner.