Over the weekend, Elon Musk’s synthetic intelligence firm xAI launched Grok Think about, a brand new generative AI instrument for producing photographs and movies. Grok Think about is out there now to paid xAI subscribers within the Grok iOS and Android apps.
Musk has been hyping up the challenge on X, sharing pictures and movies from Grok customers. This consists of some mildly NSFW content material, which the Grok app labels as “Spicy.”
This Tweet is presently unavailable. It could be loading or has been eliminated.
AI video is an thrilling — and admittedly terrifying — new frontier for the AI business. To proponents, this expertise provides artists a brand new medium for creativity and will scale back the prices of animation and filmmaking. To critics, AI video poses critical dangers for sexual deepfakes and misinformation.
Placing apart that debate for the second, I needed to see how effectively Grok Think about compares to xAI’s largest rivals. As I’ve written beforehand, Google’s Veo 3 AI video mannequin presently leads with discipline with surprisingly lifelike video. Then there’s Sora, from ChatGPT-maker OpenAI. Moreover, the widespread AI picture generator Midjourney just lately launched its personal generative AI video instrument.
So, how does Grok Think about examine to its opponents? To be blunt, I am not impressed.
Sure, Grok Think about is model new, and Musk just lately stated on X that it “ought to get higher on daily basis.” Nonetheless, as of this writing, it appears to lag far behind its rivals.
Let me present my work.
Evaluating Grok Think about AI video to the competitors
Mashable just lately wrote a few viral AI video development — safety digicam footage of animals leaping on trampolines and fascinating in related antics. So, I used a easy immediate to check Grok Think about, Veo 3, Sora, and Midjourney: “Safety digicam footage of rabbits leaping on a trampoline at night time.” Easy sufficient, proper?
First, I ought to notice that there is a massive distinction between Veo 3 and Grok Think about. Google’s Veo 3 mannequin can generate movies based mostly on a textual content immediate. Merely describe the video you need, and Veo 3 will do the remainder. Nonetheless, instruments like Midjourney and Grok Think about solely provide text-to-image technology. After producing or importing a picture, customers can then animate it, reworking it into a brief video clip. On this sense, Grok Think about is already on the again foot in comparison with OpenAI and Google.
Mashable Gentle Velocity
With these caveats, let’s dive into the outcomes, which I’ve additionally shared on X.
I put my take a look at immediate into Grok, and it returned these disappointing photographs.
Credit score: Screenshot courtesy of Grok / Timothy Beck Werth
Credit score: Screenshot courtesy of Grok / Timothy Beck Werth
I chosen the least dangerous of those photographs and created this brief video:
This Tweet is presently unavailable. It could be loading or has been eliminated.
It is…tremendous? Sort of mid, or meh, as the youngsters say.
However it additionally suffers compared to different AI video instruments.
Because the video exhibits, Google Veo 3 and Sora did significantly better with the identical immediate:
This Tweet is presently unavailable. It could be loading or has been eliminated.
Lastly, Midjourney, which animates photographs much like xAI, was in a position to produce higher photographs and movies, although it took two makes an attempt. The picture and video it produced have the grainy look of surveillance footage.
AI-generated picture.
Credit score: Timothy Beck Werth / Midjourney
This Tweet is presently unavailable. It could be loading or has been eliminated.
Audio can also be a significant drawback with Grok Think about. Whereas Veo 3 can produce sound results and coherent dialogue in sync with the video, the audio I’ve discovered on Grok Think about movies is restricted to tough sound results and gibberish.
Musk in contrast Grok Think about to a modern-day Vine app, writing on X, “Grok Think about is optimized for many enjoyable and shareable content material.”
And in my preliminary exams, Grok Think about appears optimized for creating two sorts of photographs and movies: memes and anime. If you wish to animate memes — or create sexually suggestive movies of anime ladies — then Grok Think about will do the trick, I assume. However past that, I am unable to say I am impressed.
There may be one space the place Grok Think about does shine, and that is when it comes to velocity. To date, I’ve discovered it produces each photographs and movies considerably quicker than its rivals.
This Tweet is presently unavailable. It could be loading or has been eliminated.
Mashable reached out to xAI, and we’ll replace this story if we obtain a response.
Disclosure: Ziff Davis, Mashable’s father or mother firm, filed a lawsuit in April towards OpenAI, alleging it infringed Ziff Davis’ copyrights in coaching and working its AI methods.
