The New Bard and AI Images, Videos, and Translations

Just when I was going to do a video on the revolution that has happened in translation dubbing, along comes a new AI image technique that blew my mind. Got ready to film that and then the new Bard comes out, which I have now tested in a dozen ways. I'm going to try to cover it all and show you how you can benefit. And I'll also cover how anyone with expertise in 26 diverse fields could contribute to red teaming what might become GPT5. But so let's start with Bard where less than 24 hours ago we learned of Bard extensions. It's a bit like ChatGPT plugins, but it works specifically for YouTube, Gmail, Google Docs, Google Drive, etc. But if we cut out some of the hype, what would we actually use this for? Well, here's one use case that I will be using it for that I couldn't have done before. The image you can see is a photo I took while exploring some Roman remains and where previously Bard could analyze the image and tell me a bit about Mithras, the God shown, it couldn't have done this. Recommend a YouTube video to learn more about this figure. Notice I didn't specify Mithras. It deduced just from the image that this was Mithras and then suggested a video for me. If this had been available while I was out exploring, I would have probably watched that video while looking through the ruins. I must say I really like the seamlessness of not having to separately analyze the image and then go onto YouTube and search instead. It all happens in one window in one app. Likewise, imagine seeing a travel image you like and you want to know about flying there. This was again a photo I took while traveling recently and I didn't separately ask it what location is this and how would I fly there from London? Instead, it was all one request. Flight duration and typical cost from London to location shown. They actually gave me additional information about the Fisherman's Bastion and then found specific flights using Google flights with the airlines and the typical costs. And I can say that these flight durations and timings are accurate. Let me show you one more cool thing before we get to some of the anti hype. I asked Bard to search my Google Drive for the recent Smart GPT doc and said write a Shakespearean sonnet to summarize it. And you can almost imagine this happening for the minutes of a meeting you've recently attended or anything that might be a little dry that you want to spice up a little bit now. Yes, this sonnet doesn't accurately follow the Shakespearean template, but it's really not bad. Look at this from humble prompts Thou dost create such light thy creators, AI explained. And Josh, with foresight Keen have harnessed thy power to solve the unseen with chain of thought, research, and reflection Keen Thou dost answer questions with precision keen. However, we are back to the problem of hallucinations. If you can't fully rely on what Bard is saying, its use drops massively. Now I have successfully got Bard to read a lot of my Gmail messages. It can do that. And of course some of the messages I get are comments on my YouTube channel. But this time when I asked, it simply made up feedback that I hadn't gotten. Let's be honest, you guys know YouTube. Do any YouTube comments sound like any of these three? It seems very rare to me to get a YouTube comment that is so neutral and polite and calm. Of course, I quickly checked and no such a comment didn't exist. This was Bard hallucinating. In fact, all three were. And I just want to quickly show you it can sometimes find genuine comments like this one. But the problem is what is the utility of searching for something like that if you can't rely on it? Also, sometimes you have to remind it to do things that you know it can do. For example, here it successfully went through my Gmail and found my replit bill. But when I asked how much more is my replit monthly bill than my Elevenlabs bill, it said it can't find anything. But then when I prompted again and again, it first found the receipt from Elevenlabs and then compared it. But it wasn't smooth and I would say at this point, not terribly useful. Likewise, this image recognition of Darth Vader was great even without naming the character, it got it. But the figures for gross revenue were that button where you can check the validity of an answer, and I have tried that out a few times. It's this Google button here that says double check response. After all, bard, powered by Palm 2, is supposed to be the only language model to actively admit how confident it is in its response. And while this did work to flag up that the gross revenue figure might be inaccurate, saying you should consider researching further, it doesn't normally flag up much, to be honest. Here there was just a single green line for one answer to this mathematical problem, which it got wrong. This, by the way, is a classic mathematical question that I have tested GPT4 on many times, and it also normally gets it wrong. But remember that thing about Bard being super honest when it doesn't know an answer. I would have been really impressed if this was full of orange and red saying I'm really not sure. And I even directly prompted Bard asking this how confident are you in this response? And it said I I am very confident in my response. I then pointed out where it was wrong and it did admit the error. Anyway, I don't want to be too you see a castle like this in the distance and gave me YouTube videos for it. So if I was walking along I could watch that getting hyped about the castle I'm about to see. It goes without saying that of course I have been following the Alpha Missense news, but I am still working my way through the science paper and it does seem like that it's more of a step forward than a sea change as the article technology is not only going to revolutionize movie dubbing, it's also maybe as soon as next year gonna make us question all historical footage we see. Let's start with Oppenheimer in Spanish.

And now the famous network scene in Germany. Handelst bitter sted aleph on erenstulen auf du solstjetst zofart aufstein zum fenstagen as irfnen undinen kopferau strecken und shrine ich bin so wutend und verder das nicht lenger hin nemen.

And now how about Andre Karpathy in Hindi. Form factor design kigai hai. Or let's rewrite history and make Kennedy's famous speech in French. And two more quick ones before we move on. These are the last two languages it can do. Or let's rewrite history and make Kennedy's famous speech in French. And two more quick ones before we move on. These are the last two languages it can do. Here's Liam Neeson in. Say Solo di parli Grazie Babene non se my stato forida Boston I should note that this only works for translation. You can't just get someone to say anything, not without a consent form. That's why I don't think it will be as early as next year, when a politician in a major election is able to claim that they lost due to deepfake imagery. I think it'll be very close, and I could see it happening in 2025, but not 2020 24. The guardrails are still quite tight and the technology isn't quite good enough. I'm of course predicting this on the wonderful Metaculus website. Not only are they sponsoring the video, but it's actually free to sign up using the link in the description. As I mentioned in my last video, you can see the aggregated forecast of hundreds of AI questions and I think my audience is going to be better informed than average. So do sign up and start making your own forecasts. I genuinely want our community to be among the best inform on the planet in AI progress. Speaking of which, I think many of you might be genuinely interested in some of the guests and topics I've got lined up with one of the most famous professors who argues that GPT models can't reason versus the recent news that GPT 3.5 instruct can play chess. And you may even hear the result of my game against GPT 3.5. But now it's time for this the inventive AI image generation you saw at the start of the video. I'm going to show you not only how you can create it yourself, but also how you can turn it into a video. Hopefully you can see image and maybe you like a bit of movement like this. Here are some of the other generations if you prefer things to be a little bolder. And can you spot the letters AGI in this somewhat quirky video? I just love how two of the characters just yeet their way out of the video randomly. There's two quick tools I want to show you. The first one is Fusion Art where you get 25 free credits. What I've been doing is to create bold text in Adobe Express. I download that output and then click upload black and white image. I've changed it to 16 by nine and then you can basically get creative with the prompt. You can see one of the other generations I came up with here. Then what I did is I took those Images to Runway Gen 2. You can sign up for free and then you get 45 seconds of generation. You can see here what some other people have been doing. This is from Stealthy the Time Traveller. This by the way, is the free and unlimited hugging face space that you can use alternatively, and let me just give you some tips before I move on, the higher up you make the illusion strength, the more you can see the original illusion. In this case that might be the letters AI explained, or in other cases it might be the background shapes or spirals. But one big benefit of reducing the illusion strength is that you get more interesting lifelike images. If you want different outputs, you can also change the prompt. Of course, course. Or if you want to keep the prompt the same and get different outputs, just click on Advanced options and just randomly change the seed and you'll get a different output if you ever find the resolution a little low. There are free websites where you can upscale images linked in the description, and this kind of stuff is only going to get crazier when it moves into the world of 3D, as you can see here. But if you find testing for dangerous capabilities more compelling than generating AI images, how about the OpenAI Red Teaming Network announced less than 24 hours ago? Essentially, if you are a domain expert in any of the 26 areas I'm about to show you, you can join up and get paid. And that's even if you can commit to as little as five to 10 hours in a year. Honestly, if I didn't have a bit of a conflict of interest. But if you find testing for dangerous capabilities more compelling than generating AI images, how about the OpenAI with the channel, I would definitely have signed up to this. Getting paid to get early access to what could be GPT5 and Red Team. It seems like sweet deal to me. If you are a subject matter in any of these areas, then do check out the link in the description. At the very least it should be highly interesting work. That's it for today, a lot of diverse topics covered and a lot more to come. Do let me know in the comments if you found any of this helpful and thank you so much for watching to the end. As always, have a wonderful day.