New AI Robot Autonomy OpenAI SORA 2 Gen AI Video Claude Sonnet 4.5 AI News

But as we can see here, it looks more to be like it’s just dropping it in the basket. Watch the hips here. You can see the way that the robot moves with its 31 degrees of freedom. It uses dual arm tracking and it can actually extend up to 200 centimeters or about 6.5 feet vertically. And 180 centimeters or just under 6 feet horizontally with its reach. And this is the cutting edge of embodied AI. And you can see the robot walking back and bending back down to grab more of these balls here. It bends with pretty good speed compared to other robots, certainly much faster in walking.

Overall, very realistic gait. Dropping the balls again in the basket, less of a toss, more of a placement, pick and place that we’re seeing here. But again, you can see the actual dexterity of the thumbs and the forefingers in the robot’s hands. Which makes it much more useful than many of these examples where the robot may have grippers, but maybe they’re just two fingers. Here, this seems to be much closer. You can see even the way that the index finger grabs the ball to secure it in the palm. But tell me what you think in the comments below.

Would you allow this robot to grab your balls if you’re playing tennis? Would it be faster than a human or is it just easier to just do it the old-fashioned way? Next, let’s take a look at OpenAI’s Sora 2 and see if you can notice the inconsistencies no one’s talking about. One year ago, Sora 1 redefined what was possible with moving images. Today, we’re announcing the Sora app, powered by the all-new Sora 2. It’s the most powerful imagination engine ever built. And it’s packed with new features. I’ll pass it to Bill for more details.

Did you catch that? Probably not on the first run. Here, the horses are white at first. They have no red socks, long white horns, long white tails. Both riders have blue capes and they’re holding nothing. And it’s packed with new features. But suddenly, these horses are brown. They have red socks. These riders are holding something. One cape is blue, one cape is red, short tails and short horns. Bill, for more details. No big deal. Now, at this dance, the elephant has long tusks. Now, every video comes with sound. And then short tusks. Next, the snowmobile has blue lights, no blue lights, blue lights again.

Sora 2 is also the state of the art for motion, physics IQ and body mechanics. Marking a giant leap forward in realism. And then no blue lights. Now, look at this figure skater, black shirt with nothing on it. Now, as she spins, there is a design if you pause it. No big deal. Let’s take a look at this next one. Dragon has no ears. Now it has ears. Small details. Let’s see if you can notice this next one. Good water dynamics. Can you see it? On the path to AGI, the gains aren’t just about productivity.

It’s about creating new possibilities. It’s also about creativity and joy. One, two, three, four. I didn’t catch that at first. If we go back and we see Sam Altman, he’s looking at these figurines. It’s about productivity. And at first they are without color. It’s about creating new possibilities. Then they’re blue. Then now here they’re black and this character has five fingers on one hand. Where it has four fingers or maybe four and a half on the other. One, two, three, four. And some of these characters have five fingers, some four. Now, with the duck race, some of these ducks are large ducks and some are small ducks.

And the background of the track actually changes if you look back at that. So tell me if you think VO3 is more or less consistent in the comments below. Now onto the release of Claude Sonnet 4.5, which now features all new upgrades and coding, complex agent construction and computer interaction, as well as improved reasoning and mathematical proficiency. Plus, Claude Sonnet 4.5 can now use tools to solve problems. But what’s most interesting are its upgrades. To start, in Claude Code, new checkpoints allow users to save and revert progress. The terminal interface has now been revamped with a new native VS Code extension that’s now available.

Also, the Claude API features context editing and memory tools so that AI agents can manage longer tasks and increase complexity. And for Claude app users, code execution and file creation, including spreadsheets, slides and documents, are now all integrated directly into conversations. But even more important is the release of Claude Agent SDK, which provides the underlying infrastructure used to build Claude Code to make custom AI agents. Plus, Anthropic says Claude Sonnet 4.5 is the most aligned frontier model released to date, showing improvements in reducing undesirable behaviors such as brown nosing, deception and power seeking.

And for its agentic and computer use capabilities, progress has also been made in defending against prompt injection attacks, with Claude Sonnet 4.5 currently being available via the Claude API at the same pricing as Claude Sonnet 4. And Sonnet 4.5 achieves state-of-the-art performance in benchmarks, plus it can focus for over 30 hours on complex multi-step tasks. And in terms of computer interaction on OS World, a benchmark for real-world computer tasks, Sonnet 4.5 leads at 61.4%, which is an increase from Sonnet 4’s 42.4% four months ago. In fact, the Claude for Chrome extension uses these capabilities as demonstrated by Claude navigating websites, filling out spreadsheets and completing tasks in the browser.

And there’s a temporary research preview called Imagine with Claude, which demonstrates Sonnet 4.5 generating all kinds of software dynamically in real time, and it’s available to Mac subscribers right now for five days. And finally, let’s take a look at two projects that could be big in the future. The first of these is Kite AI. They just raised $33 million, $18 million of which came from PayPal Ventures, plus they have other high-profile investors like SBI Investment, GSR Ventures, and Samsung Next. And this is a layer one being built on Avalanche, and it’s already being used live by PayPal and Shopify, and it allows AI agents to search, order, and interact with other merchants, plus settle payments.

And it’s already being used in production right now. The second project is called Sentient.agi, and it’s raised over $85 million by Peter Thiel’s Founders Fund and Pantera Capital, as well as other big names, and it’s using Polygon as its base layer, which is an L2 or sidechain of Ethereum, and they’ve already implemented some incredible features here as well. And make sure to share any AI projects that you know of in the comments below so that I can possibly check them out in a later video. Like and subscribe, share this video, and thanks for watching.

Bye. Thank you. [tr:trw].