Did You See This Week's AI News Yet?

Matt Wolfe covers this week's AI news: ChatGPT's growth is slowing, Meta launched Threads, Midjourney added a pan feature, and Steam is cracking down on AI-generated images. Google changed its privacy policy, OpenAI announced Super Alignment, and someone made an AI-generated Twitch stream of Trump and Biden.
Questions this video answers
July 7, 2023
Over 95 million posts
About 24 hours
20% of their compute power
July 3, 2023
Chapters
Now, this is something I don't get to say very often, but this week has legitimately been a pretty slow week in AI, which is perfect for me because I'm on a trip with my family to Colorado right now, and I'm not my normal studio, as you can tell by the pictures of bears that are behind me. But there are a few bits and bobs that I think would be cool to cover to keep you in the loop with what's going on with AI and what you might have missed this week. So let's check it out. Speaking of a slowdown, ChatGPT for the first time in a long time, has actually slowed the pace of growth. You can actually see that it peaked in May and in June they actually had less users than they normally have. This is the first cooling off month that Chat GPT has had since its inception. Now, this isn't necessarily AI news, but it is pretty much what the entire tech world is talking about this week, and that is Threads, which is Meta's response to Twitter. It's basically the same thing as Twitter just put out by Meta. And so far in the early phases of it, a lot of people are claiming that it's been a lot more positive than Twitter. A lot of the negativity that you get with Twitter, people just haven't been seeing as much of on threads. So that's a kind of cool thing to know about Threads. Personally, I'm a little bit worried about adding another social media platform into my life. I'm having a hard time keeping up with the ones that I'm already on, but I might get in and play around with threads. I mean, it's already got over 95 million posts and as of this recording, it's been open for about 24 hours. We did get a cool new toy this week in the form of of Mid Journey's new pan feature. This is actually similar to the zoom out feature, except now you can, instead of zooming out on the picture, move the picture to the left or to the right or to up or down. So for example, if I use a prompt like a penguin wearing a crown while lounging on a beach, we get something that looks like this. And if we upscale number one here, we now have new options here. Not only do we have a zoom out and custom zoom, we now have left, right, up, down, these are the pan buttons. So if I want to see what's going on to the right, I can click the right arrow here and then add an additional prompt. So let's say sitting next to a lobster and submit that we can see it. Panned our image to the right and put a lobster sitting next to it. Let's upscale number three. Let's pan further to the right. And now the plot thickens as we learn that either this is a giant lobster or it's a half lobster, half human. So I'm excited to see what kind of creative uses people have for this new panning feature. My other favorite AI platform, Leonardo AI, also got a nice little upgrade this week. They announced their Leonardo Canvas 2.0. Taking a peek at their demo here, you can see that they added a crab over here. They highlighted this area, changed the area, added a person in, added another person, added a sun to the sky, similar to what we see in something like a Photoshop generative fill. We can actually now use Leonardo to add additional things into the image and mask out certain areas and put exactly what we'd like to see into that image. They turn these rocks into little rocks with crystals popping out of them. Added a little banner up here. So some really cool new things that we can now do with Leonardo AI as well. There's also been news going around this week that over on Steam, they've been rejecting games that use AI. In fact, somebody in our Future Tools Discord actually said that their game was recently rejected, saying that it couldn't be posted because it used AI images, and they were currently disallowing that. However, Valve did respond and and say that the goal isn't to discourage the use of AI on Steam, but right now they're being very, very cautious because of how the AI models are trained, because a lot of the legalities and the ethics of AI images are still sort of in this gray area where it was trained on other people's artwork. But it is still unique images. The legalities of that is all still being fleshed out, and Steam doesn't want to be caught in the crossfire. So they're not disallowing the use of AI, they're just disallowing the use of AI generated images with sort of gray area origin, let's call it, for right now. Also this week, Google updated their privacy policy, and they did it in a kind of sneaky way. The new privacy policy explicitly says that the company reserves the right to scrape just about everything you post online to build its AI tools. So anything you're posting on the Internet, they're saying, look, we're probably going to use this in our AI models. And if you're using Google, you're essentially giving US consent to do that. The exact words are Google uses information to improve our services and to develop new products, features and technologies that benefit our users and the public. For example, we use publicly available information to help train Google's AI models and build products and features like Google Translate, Bard and Cloud AI capabilities. So Google is essentially saying, if you're posting it online, we're probably scraping it and we're probably going to use it to improve our various AI models. So. So hopefully you're okay with that. Let's talk about OpenAI and ChatGPT for a minute, because there's a few things that have come out this week around OpenAI and ChatGPT. First is that they introduced Super Alignment, they're starting a new team, and they're dedicating 20% of their compute power to ensure that the AI is in alignment with the goals of humans, essentially. Right now, AIs essentially rely on reinforcement learning from human feedback, meaning that it's it relies on humans to sort of supervise it and make sure that it's doing everything it's supposed to do and it's all working properly. However, when the AI eventually becomes smarter than humans, humans are no longer going to be able to supervise it because the system will be smarter than the humans. And when the systems become smarter than the humans, we need to make sure that the goals of the AI systems are aligned with the goals of human beings. Because if an AI system believes that it needs to save the earth and that the only way to save the earth is to eliminate humanity, well then the goals of humanity and the goals of AI aren't aligned. And so that is an extreme case. But that is essentially what they're saying here is they need to build a team to make sure that they are putting all the safeguards in place to make sure that this is prevented. Probably a smart move. Also this week, OpenAI announced that the new GPT4 API is going to be available to everyone. That means that the companies that are building tools on top of GPT can now also use the GPT4 models. And also if you're a fan of using the OpenAI Playground, you should now have GPT4 as an available option inside of the playground. Also, if you've tried to log into Chat GPT this week and you have a Chat GPT plus account, you might have noticed that the browsing option is gone. There's now a message down in the bottom right that says we've temporarily disabled the Browse with Bing Beta feature. And to read the Help center article, if you read the Help center article, we have learned that ChatGPT Browse beta can occasionally display content in ways we don't want. For example, if a user specifically asks for a URL's full text, it might inadvertently fulfill this request. As of July 3, 2023, we've disabled the Browse with Bing Beta feature out of an abundance of caution. While we fix this in order to do right by our content owners, we're working to bring the beta back as quickly as possible, and we appreciate your understanding. So we have no clue how long we're going to lose Internet access inside of Chat GPT. But if you do need to browse with Chat GPT, there's still always Bing. Now, while OpenAI taketh away, they also giveth that was really dumb. They also announced this week that code interpreter will be available to all ChatGPT+ users over the next week. It lets ChatGPT run code optionally with access to files you've uploaded. You can ask Chat GPT to analyze data, create charts, edit files, perform math, et cetera. Plus, users can now opt in via Settings. So if we go back over to our ChatGPT here and we come up to our settings, click on Beta features, it should show up in here. We're going to when we have access. I clearly don't have access yet, but it's coming. Finally, to wrap up this week, I want to point you to this awesome Twitch stream. There is an ongoing live Twitch stream of Donald Trump debating Joe Biden, but it's all just nonsense. It's just AI generated nonsense and it's absolutely hilarious.
Weapon Crafting Thundercrash Kills in Bubbles Are you high on your own gas fumes or just that desperate to relate to these gamers? But let me tell you something Rooney Wolf, this drummer here is right on target when they say ad breaks during streams are a joke. Just like your entire presidential campaign was one big comedy act.
Hey there Scrap Catastrophe. I see you love hugging your titty bear. Well, who am I to judge? We all have our strange pleasures in life. But here's the deal folks. Let's focus on something more important than cuddling with stuffed animals. So put down that titty bear and get ready to tackle some real issues facing our country. It's dark Brandon time. Get out of my way. Let's get down to business because it's time to twist some nipples and build back better.
So yeah, that's pretty awesome. The first time I discovered this, my buddy who goes by Lord Travis Wright on Twitter point realizing what am I doing with my life? But it was hilarious. I just could not look away. That's something fun. Nonsense, but it's all hilarious. Like I said, it's been a slower week this week, but still some interesting news that I thought you should be aware of. If you want to stay up to date with all of the AI news on pretty much a daily basis, check out FutureTools IO, specifically the AI news page. I update this page almost every single day. Also, if you go to the homepage, I add new tools to this page almost every single day as well. Just all the cool AI tools that I come across. And then you can also join our free newsletter by clicking on this button. If you click on this button, I'll give you the TLDR of everything that happened in the AI world this week. If you enjoyed this video and you want to make sure you get a news video like it every single Friday in your YouTube feed, maybe subscribe to this channel and like this video and I'll make sure more of them show up in your YouTube timeline or whatever you call it. Also, give this video a little thumbs up if you liked it and that'll make me feel all warm and fuzzy. All right, checking out from my vacation in Estes park. Colorad. I'm gonna go check out for a few days, go on some hikes and get out into nature. So thank you so much for tuning into this channel. I really appreciate you. I'll see you guys in the next video. Bye Bye.
More from Matt

AI News: Dots, GPT-6.1 Sol, Sonnet 5.5, Gemini 4, and everything you need to know









