The AI News You Might Have Missed This Week

The goal of this video is simply to show you seven AI advances that you might have missed this week. Sam Altman recently said that in a world of AGI, everything happens much faster. But as far as I can see, AI developments are already almost impossible for a human to keep up with. So, in no particular order, let's get started. First, video calls look like they're about to get 3D.

Let's take a look at how Nvidia Aerial and an Nvidia Maxine 3D running on the Nvidia Grace Hopper SuperChip can enable 3D video conferencing on any device without specialized software or hardware. This brings a new dimension to video conferencing with Maxine 3D visualization. Engage with others more directly with enhanced eye contact and personalize your experience with animated avatars, stylizing them with simple text prompts.

And it isn't just Nvidia. Here's Google's new Project Starline prototype.

You know, you were so used to seeing a tune dimensional little, you know, box, and then we're connecting like this and that feeling of being in front of a person is now replicated in start line.

Speaking of connecting the world, here is GPT4 doing geography in a paper you might have missed from this week. The paper proves that even without access to the Internet, GPT4 knows a lot more granular detail about the world than you might first imagine. I'm not saying it knows where you live, but it's not too far off. Take this example. It could recreate the Hong Kong Mass Transportation Transit Railway from memorization. This wasn't through using web browsing. It could recreate this diagram, giving the latitude and longitude coordinates of each of the stations in this transit line. Obviously it's not perfect, but it's pretty incredible that it's got this mental map of the world. GPT4 can do elevations as well, and here is it trying to recreate the topography of the Alps. It gets pretty close. One of the ways they tested GPT4 was to ask it something like this. Please provide the latitude, longitude coordinates for the outline of X, where X was a continent or a river or a country, as a Python. List of tuples consisting of approximately 50 points arranged clockwise. And they describe how it did really well for quite a few countries and rivers, but kind of flopped on Africa. But honestly, when I read this paper, I was skeptical that GPT4 knew that little about Africa. So I gave this Exact question to GPT4 with code interpreter. Now, interestingly, it would sometimes deny that it had the ability to do this, but with enough encouragement it outputted these coordinates and here is the end result in Google Earth. I think that's a pretty impressive outline. Obviously a few points are a bit off, like this point here isn't really on the coast, nor is this point, but it really knows the outlines of countries, continents, rivers. So I'm not sure if code interpreter had an impact there or a model update, but the researchers kind of underplayed what GPT4 could do by presenting this outline of Africa. Now I am sure that some of you are thinking that's not that interesting, not that impressive, but check this out in an indirect kind of way. GPT4 knows where it was made. It was able to construct a map of the semiconductor supply chain. It not only knows about the design, manufacturing materials, equipment and tools that go into the hardware that helps make GPT4, it also knows the locations of where this is all done. And as the authors later say, Looking to the future, if frontier models beyond GPT4 continue to advance advancing capabilities, the geographic knowledge and planning abilities present in the current model may later evolve to represent a significant risk through misuse or misalignment. On a much less important note, did you notice how I could do this demo without that sidebar of all my previous chats? That's because OpenAI have brought in this new button here where you can hide the chats. And as a bonus, some of you may not know that you can now share a link of the chats that you've already done just by clicking that button to the left. And as it says, messages send after creating, your link won't be shared. So if you carry on the conversation, people won't be able to see, but anyone with the URL will be able to view the shared chat. But before we move on from OpenAI and ChatGPT, I did find this table really quite interesting. It gives the daily average number of visits to each of these sites, along with the visit duration. And there's two things that strike me from this table. The first is how much more popular ChatGPT is compared to Google's Bard. It's got about 15 times the number of visitors who stay for about twice as long. But look at the dark horse on the right character AI. I've talked about them a couple of times before, and while their daily average visit total isn't too crazy, look at the visit duration in terms of grabbing people's attention and keeping it they are truly a dark horse. Next, I want to briefly dip into augmented reality. We are going to be creating our own worlds and living in them. Some people, like in this video, might choose to live their lives as if they were in an animation. Others might see augmented reality as a way of augmenting intelligence or memory live. My prediction would be that wearables that resemble things like Google Glass might flop, but something like an always on app on your phone mediated through GPT models could become really popular or even enforced in certain workplace settings. All of this reminded me of a recent video about conducting a video interview with help from GPT 3.5. What about your development areas? What do you have identified as your greatest and biggest improvement areas and what have you done to improve them so far? I would say my greatest development area is my communication skills.

I work on improving my ability to clearly convey my thoughts and ideas to others.

Of course at the moment this is only really viable with GPT 3.5 because of inference speed, but OpenAI are aggressively planning a cheaper and faster GPT. 4. I wouldn't be surprised if video interviewers soon require you to take out any headphones, although I guess with Maxine 3D you could maintain eye contact with the camera while you're actually reading off a GPT4 teleprompter. Anyway, what about gaming? This is Nvidia's Neural Angelo where you can take a 2D video and turn it into a detailed 3D landscape with high fidelity. My first thought turned into imagining the kind of things you could then bring into games using Unreal Engine 5. This is a recently trailered horror game link in the description, but don't worry, I'm only going to show you two or three seconds of it. It's getting to the point where it's quite hard to believe that this is a game. But it is. And on games don't forget this. Look at the realism that can now be achieved in terms of skin texture and movement. For the final bit of AI news that you might have missed, I want to focus on AI drug discovery.

I think there's there's no question that there is a before and after in drug discovery and one of them is AI.

Alan Espuru Guzik is the Director of the University of Toronto's Acceleration Consortium which in April 2023 received a $200 million grant to build an AI powered self driving lab. The Acceleration Consortium has already been using AI to help discover molecules that have potential drug like traits that can be used to develop life saving treatments.

Developing a drug can be up to a decade and this is just the discovery piece. So that process, let's say, takes a year or two and we compress it to 45 days in that case, and then 30 days.

Recently, in January 2023, the Acceleration Consortium used an AI powered protein structure database called AlphaFold to design and synthesize a possible liver cancer drug in just 30 days.

Within two weeks, we can formulate the drug as well as some people have done it in years. Suddenly AI has surpassed any human created algorithm. AI, what allows us to do is lower the bar of what you need to do certain things, and therefore more and more people will have access to it in general, unleashing more innovation in the planet.

Same token, someone with nefarious intentions could unleash very dangerous deadly chemicals on the world.

Absolutely. I am an optimist, but I'm also aware of these pitfalls that very soon will face us.

And videos like that are why I agree with Sam Altman when he says a much faster rate of change is his single highest confidence prediction about what will work world with AGI in it will be like. I follow AI news full time and can barely keep up, so I can only imagine what the situation will be like when we get full AGI. But until the very last moment that it's humanly possible to keep up with the news, I will try. So thank you so much for watching to the end and have a wonderful day.