Forward Future Episode 25

Dall-E 3 Free

People figured out how to get Dall E 3 for free, right now, and I want to show you. Last week, Dall E 3 was released. It’s an incredible AI generative art model that rivals Mid Journey. It’s rolling out right now to ChatGPT plus users, which is $20/mo. In fact, I just got access to it, let me know in the comments if you want me to do a review of it. But, a few people figured out that you can get Dall E 3 completely free right now. All you need is a Microsoft account. I thought you might need to download the Edge browser, but turns out that’s only for Bing Chat. So with your Microsoft account, navigate over to bing.com/create, and…that’s it! You now have the incredible Dall E 3 image generator. You initially get 100 credits, which is enough for 100 generations of 4 images each, and after that might just have to wait a bit longer for your images to be created. Check out some of these images from Dall E 3, they are truly incredible. And, if I were MidJourney, which is $10/mo AND can only be used through Discord, I’d be really nervous right now.

Viral Image…Real or AI?

There’s a set of images going viral right now and people can’t figure out if they are AI-generated or not. That should really speak to the current state of artificial intelligence and generative art specifically. Take a look at these images they're wild they show a very large man, eating pizza with what looks to be an alligator in a swamp. I'm not sure if that's an alligator or a crocodile. Another image shows A piece of pizza hanging off the man's neck with the crocodile eating it, and his face seems extremely excited, in the next one, he's doing a karate kick move to the face of a crocodile with other crocodiles around him. These images are so fantastical. It's hard for me to believe that they might be real. It's unverified right now but the quality of the images is really good to the point where this Reddit post has over 37,000 upvotes at the time of this video and no one can really seem to tell if it's real or not. What do you think? Let me know in the comments if you think this is real or not.

MSG Sphere

Next, the MSG Sphere in Las Vegas had its first concert. The first artist to perform at the Sphere was U2, and the Sphere itself is one of the most impressive feats of engineering that I've ever seen. I live in California, and I'm only a 5 Hour drive away from Las Vegas I've seen this thing in person, and it's as incredible as it looks in all of the videos. Let me tell you some stats about it. This sphere is by the same company that made Madison Square Garden in New York, and this one is called the Madison Square Garden sphere. It cost $2.3 billion to create stands 366 feet tall and 516 feet wide. It has 160,000 high-quality speakers and 260 million video pixels. It looks stunning from both the outside and the inside. We also found out this week that the cost to advertise on the sphere just for one day is $450,000 and an entire week $650,000, so you really got a great price discount for that whole week, and the price includes the production of a 90-second spot. I'm a huge fan of Las Vegas, and I can't wait to go see a concert here it looks like an incredible experience that mixes both the live aspect of a concert with the incredible visuals in almost like a mixed reality experience.

Three AI Wearables Launch

Next, this week seems to be the week of AI wearable hardware. Three new innovative hardware devices were launched this week, and I'm gonna talk about all three of them but with these launches come some controversy which I'll talk about in a minute. The first device is the humane AI pin, which was created by X Apple executives and uses projectors, cameras, and AI tech, all packed into a small form factor that you're supposed to wear the company humane, which makes the device unveiled the AI pin at the Paris fashion show with supermodel Naomi Campbell as the first person outside of the company to wear the device in public, according to end gadget, the company describes the device as a screenless, standalone device and software platform Bill from the ground up for AI. It's powered by an advanced QUALCOMM snapdragon platform and equipped with a mini projector that takes the place of a smartphone screen along with the camera and speaker. So basically, if you look at this video you can see that somebody's holding their hand out, and the AI devices projecting a user in a face onto their hand. The device also performs functions like AI-powered optical recognition, but it's supposedly privacy first, and there is no always listening mode. Personally, I don't think I'd use this device. I don't see it replacing my phone.

Next is the tab device by Avi Schiffman. If that name doesn't ring a bell, he's a young technologist who was behind the Covid dashboard and an Airbnb-like platform for Ukrainian refugees. The tab device, again is a wearable AI hardware device and is touted as an AI companion, it also integrates data from users’ daily lives. It monitors conversations actively and provides instant access to the world’s knowledge through AI models. According to Avi by having Tab listen to key conversations throughout the day whether it's concerns that the user mentions or ideas, they brainstorm you don't have to worry about forgetting anything you chat with the device is personal AI later as it retains a full context of your daily life and work with it to brainstorm new ideas ask it what should I do today or just generally have someone to talk to.

The last device is by a company called Rewind AI rewind AI is a very impressive product that basically records a lot of what you do on the computer every day, and it allows you to query against it and it provides suggestions all through AI but now they've designed and launched a wearable device that essentially records everything you do all day very similar to the previous products mentioned. According to the website they've already had over 3000 pre-orders for the device. Now let's talk about what you're all probably thinking right now do I really want a device recording? Everything I do all day every day. Personally, I understand the value of never forgetting anything and having an AI assistant remind me of things before I even know, I need them, now with that out-of-the-way, I would not use these. I think there's actually a benefit to forgetting some things. Our brains are not meant to remember everything and even if we had a device to help us remember everything I don't think we're supposed to. I believe we were actually meant to forget things that are just not as important as other things in our lives and of course, the considerable controversy is the privacy factor. All three of these products claim to be privacy first but how can they be when it's recording everything around you at all times? I can imagine someone walking into a public space with one of these devices on and everyone around them would be pretty upset about it. There's an argument to be made that we're already in this world because everybody has a device in their pocket that can record everything in video, but it still takes someone taking out their phone and actually actively recording, and it's pretty apparent when somebody's doing that. With these devices it's going to be obvious when somebody's recording you I suspect either out of federal level or a state level at least in the United States. There's going to be a lot of legal pushback. If not banning these types of devices are you going to get one let me know.

LLaMA 2 Long

Meta released LLaMA 2 Long which is an AI model based on the Lu model by meta-released a few weeks ago, but allows for longer context windows, which is always a welcomed feature. According to Venture Beat Meta’s new elongated AI model outperforms some of the leading competition in generating responses to long user prompts, including open eyes GPT 3.5 turbo with its 16,000 character context window and Claude two with its 100,000 character context window. Longer context windows mean longer prompts and responses which is suitable for a range of AI use cases, including coding. I'll drop a link in the description below to download the LLaMA 2 Long version.

Mr. Beast Deepfake

Next, the biggest YouTuber in the world, Mr. Beast called out a deep fake of himself, promoting a scammy two-dollar iPhone. Take a look at the video now as you can see the visuals look pretty good, and the voice is decent, but it's likely enough to convince many people that this is real. MrBeast took to Twitter, now known as X, to let people know to avoid this scam. MrBeast said lots of people are getting this deep fake scam ad of me on our social media platforms, ready to handle the rise of AI deep fakes. This is a severe problem. And I agree entirely. I know there are a lot of researchers that are trying to invent ways to detect AI-generated content, but there's always going to be bad actors who are able to create this and subvert detection. This isn't the first deep out there, Tom Hanks and Gail King have both been targeted in AI deep fakes. Because these celebrities have so many videos of them, it makes it especially easy to create deep fakes based on their likeness. So, as has been the case since the beginning of the Internet, don't believe everything you see online and always verify before giving your information, especially your credit cards.

Apple AI Hiring

It looks like Apple is bucking the trend of mass layoffs within the tech industry. Apple CEO Tim, Cook said the company will be hiring researchers and engineers in the UK. Although Apple hasn't directly announced any AI features or large language models, there have been a number of rumors of them building an internal large language model likely to power future versions of Siri called Ajax. According to the BBC article, Mr. Cook said AI was behind. Several prominent features on Apple devices, such as software that detects if a person has fallen or been in a crash, as well as more commonly used tools, such as predictive typing. But these features are great, but only scratch the surface of what's possible with AI especially when you have incredibly powerful chips in most people’s pockets that can power many devices. I just reviewed a very small large language model that could likely fit on any phone, and it performed exceptionally well for most use cases. It's only a matter of time until they release a large language model-powered version of Siri, and I can't wait for it, especially if it's able to actually execute commands using something akin to Code Interpreter.

Google Pixel 8

It looks like Apple isn't the only tech giant making AI news this week. Google released their new Pixel 8 and Pixel 8 Pro phones that are packed with AI features. According to their blog post, these new versions of the pixel phones are built with AI at the center for a more helpful and personal experience powered by the tensor G3 chip. For example, you can use Magic Editor within Google Photos, which uses generative AI to help you bring your photos in line with the essence of the moment you were trying to capture. And they're also including an improved AI call screener, which helps you receive 50% fewer spam calls on average. I'm an iPhone user, and I get spam calls all day, every day so I would love a feature like that. They're also including a beef-up version powered by a lot of the AI research that they've been working on. It might be time for me to switch back to an Android phone.

RT-X

Google Deep Mind has launched a new model called RT–X. This model is supposedly extremely good at writing instructions for robots. This is something that I made a video about a while ago, but it's finally coming out and being published. according to the deep mind blog post today we are launching a new set of resources for general purpose robotics learning across different robot types or embodiments together with partners from 33 academic labs. We have pooled data from two different robot types to create the open X-embodiment data set. We also release RT one X, a robotic transformer, model, derived from RT one and train Ed on our data set that shows skills trans across many robot embodiments. The basic way to think about this is like code interpreter, which is able to write code that can be executed in an environment, but rather than writing Python code, it's writing code that can make robots do things. This is another step in the direction of having an AI robot personal assistant in every house. Ever ever seen the movie iRobot? If not, I'll save you the time it ended really well and there were no problems at all.

Sam Altman Foot In Mouth

Next, Sam Altman can't seem to keep his foot out of his mouth. Elizabeth Weil notes in a new profile of the OpenAI CEO that he has mentioned that AI will likely replace what he calls "a median human.” According to the article in futurism.com, Altman hopes that artificial general intelligence will have roughly the same intelligence as a median human that you could hire as a coworker. Although he might be bright in predicting what the future looks like with artificial intelligence and the workforce, he needs to think about how his words are coming across to the general public, who aren't as excited about this AI revolution. Last week he joked about open AI achieving AGI internally, which he quickly clarified was a joke and then this week he's talking about as if they are entirely replaceable as the CEO of the leading AI company. He should be much more thoughtful about how he conveys his predictions of the future regarding AI.

StableLM 3B

In what seems like a trend that I'm all for, more small models are being released. This week StabilityAI released a 3 billion parameter version of their stable LM model in their Twitter post stability. AI says bringing sustainable high-performance language models to smart devices. I am a massive fan of the smaller, highly performant models that can be loaded onto almost any device with no Internet required. It absolutely blows my mind that we can have the entirety of human knowledge baked into just a few gigabytes of storage. Let me know if you want me to fully test the stable LM3B model.

AutoGen

In probably my favorite story of this week, Microsoft released AutoGen, which is a framework to build multi-agent capabilities into large language model applications. You can think of this as ChatGPT plus Code Interpreter, plus plug-ins, but fully customizable and flexible. Essentially, you can give a group of AI agents an assignment, and they will go do it. They have a bunch of tools at their disposal, and they can execute code as well. I've already made a video tutorial showing you how to use it, and I plan on making more videos about it because I'm completely enamored with it. I've also been building things that will help me automate stuff that I do every day, and I plan to make a video showing off the different things that I'm building. I'll drop a link in the description below for where you can check out AutoGen as well as the video that I created so you can know how to use it. If I were to make a prediction about which AI technology would be the most valuable in the future, this would undoubtedly be a top contender.

Awesome Robot

Next, we have another story about a robot. This one's design is insane. This robot, which stands on two legs or two wheels, can carry hefty things and is highly agile. Built by a German company,

this is an autonomous mobile robotic system. Take a look at this video showing off what this robot is capable of. I believe robots are going to be increasingly used in our society, especially as AI continues to be injected into these robots. They just become that much more valuable. There are concerns about having intelligent robots everywhere, but I tend to be pretty optimistic about the future of robotics.

Canva AI

Canva, the web-based image editor, has released its generative art product. I am a huge fan of Canva, and in fact, Canva is what I use for all of my thumbnails. I can't draw a stick figure, so creating those thumbnails I thought would be impossible for me, and I thought I would need to hire somebody. And then I found Canva, which allows me to easily create those thumbnails and is infinitely simpler than Photoshop. This is not a sponsorship by Canva. I'm just a huge, huge fan. Canvas's new product called Magic Media lets you do text-to-image easily and not only images, but it also includes video, which is unique compared to Mid Journey and Dall E 3. Right now, I'm only aware of Gen 2 being able to generate AI video. I haven't played around with this new feature yet, but I plan on it.

AI Video of the Week

This week’s video of the week is maybe my favorite one so far. I found this one through Bilawal Sidhu’s X post, and it's created by Reuben Fro on X. Reuben says this is created by using burning Gaz Ian splat in Unity 3-D. The results are gorgeous. Check out this video now. And here's another example by the same creator, using gaussian splashing. I don't know much about these techniques, but they're gorgeous. I'm going to have to do more research into them.

Arc Browser AI

For our last story, we're talking about the arc browser. I haven't personally used the Arc browser, but many of my friends use it as their daily driver. Josh Miller, who is the CEO of the browser company that makes the Arc browser, is going all in on AI. In a story by The Verge, he talks about being happy to have missed the crypto wave but has identified the AI wave as something that he's willing to bet the company on. He wants to embed AI into every aspect of your web browser. Here are some of the features that are going to be coming for the Arc browser powered by AI first ask ChatGPT, which will allow you to ask the AI questions right from the Arc command line. This seems like the most noticeable feature and probably the least innovative one. Next, Arc includes something called tidy tab titles, so when you pin a tab in the browser, Arc automatically renames the tab to something that is cleaner and more concise. Similarly, tidy downloads will automatically rename files that you download. For example, sometimes you get these long and cryptic file names when you download something from the Internet, but Arc now will rename it to something that will describe the download. Also, they're going to include five-second previews, which allow you to hover over any link, hold down the shift key, and then arc will fetch a summary and preview of that webpage. Last called Ask Page, so when you use command F, which typically allows you to find a word on a page, it'll not only look for that keyword as usual, but if it can't find it, it'll also use AI to get an answer to your query specific to the page that you're looking at which seems incredible. I use command F every day, all day. According to the story, all of these features Parrot says serve the same goal they make your Internet life easier and help you do stuff without screaming look at all this AI. So maybe it's time for me to try out the Arc browser.