Texti Newsletter #7 |
Happy New Year Everybody! Welcome to a new challenge and a new year! It's officially 2024 almost everywhere (China sorry you live in a different dimension). I am really excited about the following year! We have grandious plans for the following year and we look forward to it! From everybody at Texti, we are wishing you a sparkling New Year filled with bright moments! May the coming year wrap you in the warmth of love and the joy of laughter. Here's to new adventures, health, and happiness in 2024. Happy New Year! 🌟✨🎉 The highlights of the week are a little short (everybody's on vacation)
Video PoetAs predicted, this year is going to be all about text 2 video. Google is not ready to stand aside, and it has been in the works for a while. Google has posted their own research paper on Text2Video, unfortunately no hands-on yet, but there are plenty of examples on their web page. They even a pretty cool demo where present a little story about a raccoon and his travel across space via portals. The major differentiation between Video Poet and competition like Pika, is that it is meant to produce a continuous stream of video, with a narrative and an actual story line, rather than an arbitrary image. As Google describes it, it's a state of art model that uses under the hood a bunch of other individual components like MAGVIT V2 for video generation, SoundStream for audio matching and an autoregressive language model to orchestrate all of this together and provide a unique content format. This indeed looks like a game changer for the media creation and future tiktoks. As described it currently targets short content format. It even allows you to add stylization, in video modifications and generate video from static images 🤯. All we need to do is just work on our imagination to create original content ideas. I think we're getting closer to fully self-sustained AI generated content creators. At the beginning of last year somebody said that in the next 5-10 years, software engineers won't exist. In 2023 we were able to make a bot do anything just with a set of simple instructions, simply by using Auto-GPT. By 2025 we should be able to create a bot that will create short content form videos, targeted to every specific ethnicity, that will be fully automated, follow the news and current trends. I think this is already wild - the future is today, we live it and we just need accept that this is here and now, and enjoy the really unlimited possibilities we currently have! The New York Times Sues OpenAI!It ain't a new thing that OpenAI has been training it's data on the free data available on the internet. Ultimately all this information that AI is aware of, is just an approximation algorithm of words provided by humans, and written over and over again, in the last couple of decades we published on the interwebs. That's why Reddit API is now paid, that's why OpenAI has had conversations with many publications on paying for their models to train their dataset. Apparently it seems that the discussions haven't had any advancement since August, because The Times and OpenAI couldn't agree on a specific amount to pay (hint this seems the root cause of the scandal). The Times complains that OpenAI is making and raising billions while they produce the raw material that has to chew up by a language model. Note to say, OpenAI has been signing similar agreements with a multitude of media companies out there, to make it right for them and fair or at least somewhat. This is the first major lawsuit against OpenAI, and I'm extremely curious to see how this is going to end out. After-all OpenAI promised that they will protect and step-up for every company that has been using OpenAI's API and has had a law-suit filled against them. That's exciting and interesting, as this is going to be a really decisive moment in the history of AI. If this is going to be solved with money, the decisive moment will be postponed for a while, otherwise, we'll know how the further evolution of large language models (LLM) will follow through. Stepping out a little, and looking at the image from top, we can see a classic battle of the industrial revolution, a business that has been at it's peak about a hundred years ago, fights against the new emerging technology, that will replace it. The fact that The Times complains in clear text that OpenAI's raising money is a clear sign they don't understand the Open AI's investors view over the company and why the times doesn't get money poured in. The Times's value is clear as day, they hit their ceiling, Open AI's just starting. We have yet to see it thrive! Your Go-To Topic Tip for Daily Discovery!I love Texti.app and I love generating content with it, however I often encounter issues and it has limited abilities so it keeps answering with answers like:
This is really annoying and it makes the answer really long and pointless from time to time, especially because I don't really need this answer. I need to get straight to the answer. Well there's a really cool trick you can do. If you apply the following topic to a conversation AI model will make sure to answer the questions way better. If you're using Texti.app you can apply the topic to any conversation by clicking on the cog ⚙️ in the chat menu, and selecting General -> Everyday Conversation. Or alternatively just paste the following in your topic area:
That's it folks, see you next week ❤️️️️️️️ Happy Prompting! Remember to invite your friends to subscribe at https://texti.app/newsletter |
