The next few years in #AI are going to be wild. Check out this guitaring panda made by the new text-to-video model from China, KLING, that's rivaling OpenAI’s unreleased Sora model. KLING can generate up to 2-minute long videos with 1080p quality with just a text prompt 😲
James Kinsley’s Post
More Relevant Posts
-
OpenAI introduces a limited-scale preview of its latest tool, Voice Engine, capable of cloning any speaker's voice using just a 15-second audio clip. The technology promises natural-sounding speech with emotive and realistic tones. Under development since 2022, Voice Engine is an extension of OpenAI's text-to-speech API, already integrated into preset voices and the Read Aloud function. Samples demonstrate remarkably authentic renditions resembling genuine human voices. Visit www.smartnews.pk for more! #OpenAI #VoiceEngine #VoiceCloning #TextToSpeech #AI #smartnewspk
To view or add a comment, sign in
-
-
Having the source code (the parameters) won’t help us avoid SkyNet taking over, because we don’t actually understand our models. Recall from the movies that pulling the plug wasn’t an option. So, the options left are: (1) Predict when SkyNet is going to attempt a take-over, and intervene beforehand (2) Intervene by modifying SkyNet, when it makes an attempt, so that it’s still operational, but harm is avoided (3) Build a “loyal SkyNet” to fight it, after it has taken over. But, we don’t understand how our trained models work. So, we have no analytical tools to do prediction for (1), we do not know how to edit the parameters correctly to achieve (2), and we therefore have no way of preventing our “loyal SkyNet” (3) from also turning on us. Open source is not enough. We also need to understand our models. And we have to stop being content with alchemy. We need an exact theory of neural networks.
Since many AI doom scenarios sound like science fiction, let me ask this: Could the SkyNet take-over in Terminator have happened if SkyNet had been open source?
To view or add a comment, sign in
-
India's first #artificialintelligence regulation proposal has come to the fore, but people might be still wondering: Do we really need to regulate AI? Do we have enough GPUs? Can we afford them? How can we handle #deepfakes? How can we examine copyright mess around AI? Folks, this is why AIACT.IN version 3 matters. This is an explainer video right from the author of the bill, Abhivardhan - explaining why this bill matters, and what should you know. You can access this explainer of AIACT.IN Version for free at https://lnkd.in/gDsUWvSP And yes, please feel free to send in your feedback on AIACT.IN Version 3 at vligta@indicpacific.com.
To view or add a comment, sign in
-
OpenAI has launched a new AI model called SORA that can convert text into a video format. This model can generate a one-minute video. which raises the question: how will it impact the movie industry? Check out SORA here: https://openai.com/sora. #OpenAI #AIModel #SORA #TextToVideo #MovieIndustry
To view or add a comment, sign in
-
-
OpenAI's new introduction Sora, an AI model, can understand and simulate the physical world in motion. In simple words, Sora can create on-minute realistic videos in response to prompts. Here is a prompt (OpenAI's) and video: Prompt: Historical footage of California during the gold rush. (credit: OpenAI)
To view or add a comment, sign in
-
As an early adopter to Chat GPT and AI, this is one of the best explanations of the technology and practical uses. And it’s from a NASCAR driver. Skip to 10 minute mark if you just want the Chat GPT info https://lnkd.in/e-QcPk9J
To view or add a comment, sign in
-
🇨🇳OPENAI’S SORA HAS SOME SERIOUS COMPETITION China’s OpenAI competitor, Kuaishou Technology, just released their latest text-to-video AI tool, Kling. Kling has the ability to create two-minute videos in crystal-clear 1080p resolution at 30 frames per second, while faithfully replicating real-world physics. This is a simple and awesome example of a Mad Max commercial created in less than an hour. GenAI market is not thing for one or two players…. #genai #ai
To view or add a comment, sign in
-
Creating Video From Text Sora is an AI model that can create realistic and imaginative scenes from text instructions. All videos on this page were generated directly by Sora without modification. #openai #soraai #soraopenai #midjourney
To view or add a comment, sign in
-
I Can Help You with AI, Data Projects 👉atliq.ai | Helping People Become Data/AI Professionals 👉 codebasics.io | Youtuber - 1M+ Subscribers | Ex. Bloomberg, NVIDIA
The LLM market is heating up further 🔥 Anthropic (an Open AI competitor and a company backed by Amazon) released a new model that seems to outperform GPT-4o. We have used a prior version of Claude models at AtliQ Technologies for client projects and are satisfied with the results. Now with this new release, the cost for LLM API calls is going to go down further which is good news for companies building solutions using LLMs. What's your take on the ongoing competition in the LLM market? 👇🏼 #genai #LLM
To view or add a comment, sign in
-
-
Forever learning @ Dalitics. Ethan Mollick's "Co-Intelligence" book is an engaging exploration of AI as thinking companion who can inspire your creativity. Tools such as Claude and Perplexity can do that, in my experience. Now, it's time to try a new tool - Luma https://lnkd.in/edNHzQzN
I had Luma animate my book cover and I really like the effect of the uncanniness fading away to something kind of normal by the end. For now, AI video works best when you can embrace the weirdness. We will see what OpenAI's Sora and other fully multimodal models bring. It appears that Luma does not have a world model (it is not actually simulating reality) but if Sora does, as some of speculated, then the difference should be pretty huge.
To view or add a comment, sign in