Google's Veo 2 Pricing, NEO Gamma, and Grok 3’s AI Benchmarks Under Fire

Summary

AI Deep Dive: Google's Veo 2 Pricing, NEO Gamma, and Grok 3’s AI Benchmarks Under Fire

Hosted by Daily Deep Dives
Release Date: February 24, 2025

Welcome to this episode of AI Deep Dive, where the hosts delve into the latest and most impactful developments in artificial intelligence. This episode covers three major topics: Google's cutting-edge video-generating AI, the controversy surrounding Elon Musk's Grok 3, and the future of home robotics with NEO Gamma. Below is a comprehensive summary of the key discussions, insights, and conclusions drawn from the episode.

---

1. Google's VO2: Revolutionizing Video Generation

The episode opens with an in-depth look at Google's VO2, a groundbreaking AI capable of generating entire videos. Hosts A and B express their astonishment at the rapid advancements in AI technology.

Cost Implications:
B highlights the staggering cost associated with VO2, stating, “the cost is, get ready for this, $32,000 per hour of video” ([00:56]). To put this into perspective, A compares it to the production cost of Avengers Endgame, which was approximately $32,000 per second ([01:12]). This comparison underscores the significant investment required to utilize VO2 for high-quality video production.

Target Audience:
The hosts discuss the primary users of VO2, noting that it is likely aimed at large-scale productions such as Hollywood studios and major advertising agencies, rather than individual content creators or YouTubers. A remarks, “it's definitely not for, you know, just your average YouTuber” ([01:20]).

Pricing Models and Accessibility:
A and B explore the differing pricing strategies within the AI industry. B mentions OpenAI's Sora model, which offers access for a flat monthly fee combined with ChatGPT Pro ([01:43]). This contrast raises questions about whether subscription-based models will democratize access to advanced AI technologies or if high costs will keep tools like VO2 exclusive to elite users. A speculates, “could that even put pressure on Google to like lower their price for VO2?” ([01:49]).

2. Grok 3 and the Benchmark Controversy

The discussion shifts to Elon Musk's Grok 3, touted as the "world's smartest AI." However, the hosts reveal that Grok 3's claims are under scrutiny due to questionable benchmarking practices.

Benchmarking Practices:
B elaborates on the controversy, explaining that Grok 3 uses a method called "consensus at 64," which allows the AI to attempt multiple solutions to a problem and then select the most frequent answer ([02:11]). A criticizes the omission of this data when comparing Grok 3 to OpenAI’s models, describing it as “kind of shady” ([02:46]).

Comparative Performance:
While initial graphs make Grok 3 appear superior, deeper analysis using consensus at 64 reveals that some of OpenAI’s models outperform Grok 3 on specific tasks ([02:50]). This raises concerns about transparency and the integrity of AI performance claims.

Efficiency Considerations:
A points out that without knowing the computing power each model utilizes, it's difficult to make fair comparisons. “It's like comparing two cars' gas mileage without saying one's a tiny compact and the other is a giant SUV” ([03:19]). This analogy emphasizes the need for comprehensive data when evaluating AI capabilities.

Takeaway:
The hosts conclude that consumers and stakeholders should approach AI performance claims with skepticism and demand greater transparency from developers to ensure accurate and honest assessments.

3. ByteDance vs. Deep Seek: The Open Source AI Battle

The episode transitions to the competitive landscape of AI development, focusing on ByteDance, the company behind TikTok, and its rivalry with Deep Seek.

Deep Seek’s Open Source Approach:
A and B discuss Deep Seek’s innovative strategy of adopting an open-source model, making their AI technology freely available for modification and collaboration ([03:37]). B highlights that Deep Seek’s chatbot is already surpassing ByteDance’s Dubai in daily users, despite being relatively new to the market ([04:12]).

Impact on AI Development:
A explains that open source fosters a collaborative environment, enabling rapid advancements and widespread adoption. “It's a whole different way of thinking about AI development” ([03:55]). This approach contrasts sharply with companies that keep their AI technologies proprietary, potentially shaking up the industry dynamics.

Future Implications:
The hosts ponder whether Deep Seek’s success will pressure other AI companies to adopt more open and collaborative models, potentially accelerating innovation and accessibility across the board.

4. NEO Gamma Humanoid: The Next Generation of Home Robots

One of the standout segments of the episode is the exploration of NEO Gamma, a humanoid robot developed by One X Technologies, heralding the future of home robotics.

Design and Functionality:
B describes the Neo Gamma’s design as “friendly,” with features like soft covers for safety and emotive ear rings that move to display expressions ([04:44]). This user-centric design aims to make robots less intimidating and more acceptable in domestic settings.

Advanced Language Model:
A delves into the robot’s proprietary language model, which allows for genuine conversational interactions rather than just pre-programmed responses ([05:04]). This advancement enables users to engage in meaningful dialogues, ask for assistance with chores, receive news updates, or simply chat about their day.

Real-World Testing:
The CEO of One X Technologies is quoted saying, “developing robots in homes, not just in labs, is super important” ([05:25]). The hosts agree, emphasizing the importance of testing robots in real-life environments to ensure they can handle the complexities and unpredictability of everyday homes.

Ethical and Social Considerations:
A and B discuss the broader implications of integrating humanoid robots into households. Concerns about privacy, safety, and the potential impact on human relationships are raised. A notes, “what if they start like predicting our needs before we even know what they are?” ([06:24]), highlighting the delicate balance between technological advancement and ethical responsibility.

Future Prospects:
The discussion touches on the dual-edged nature of such advancements—while they offer convenience and enhanced interactions, they also raise significant questions about autonomy, dependency, and the ethical use of AI in personal spaces.

5. Ethical Considerations and Future Implications

Throughout the episode, hosts A and B stress the importance of ethical considerations in AI development. They argue that as AI becomes more integrated into various facets of life, establishing safeguards to prevent misuse becomes paramount.

Transparency and Accountability:
The controversy surrounding Grok 3 underscores the need for transparency in AI benchmarking and performance claims. Without honest reporting, stakeholders cannot make informed decisions or trust in the technology’s efficacy.

Collaborative Development:
Deep Seek's open-source model suggests a pathway toward more inclusive and democratic AI development, potentially mitigating some ethical concerns by fostering community oversight and collective responsibility.

Human-Robot Interaction:
With the advent of humanoid robots like Neo Gamma, the ethical landscape becomes more complex. Issues of privacy, data security, and the psychological impacts of human-robot relationships must be addressed to ensure these technologies benefit society without causing harm.

Final Thoughts:
In their closing remarks, the hosts encourage listeners to contemplate the kind of future they want to build with AI. A poses thought-provoking questions about shaping AI’s evolution, emphasizing the collective responsibility to embed values and safeguards into these powerful technologies.

A: “If you had the power to shape the future of AI, what kind of world would you create? What values would you focus on? What problems would you solve?” ([07:10])

B: “It's something we all need to be thinking about. Well, that's all the time we have for this deep dive into AI.” ([07:25])

The episode concludes with a call to action for listeners to stay informed, question developments, and actively participate in the conversation surrounding AI’s role in shaping our world.

---

Key Takeaways:

1. Google's VO2 represents a significant leap in video generation technology but comes with high costs, limiting its accessibility to large-scale producers.
2. Grok 3's benchmark claims are under scrutiny, highlighting the need for transparency and comprehensive evaluation metrics in AI performance assessments.
3. Deep Seek's open-source approach challenges traditional proprietary models, potentially leading to more collaborative and rapid advancements in AI.
4. NEO Gamma exemplifies the future of home robotics, blending advanced conversational capabilities with user-friendly design, while also raising important ethical questions.
5. The episode underscores the crucial balance between technological innovation and ethical responsibility, urging stakeholders to consider the broader implications of AI integration into daily life.

Stay tuned to AI Deep Dive for more insightful analyses and updates on the ever-evolving landscape of artificial intelligence.

Transcript

[A]

Foreign.

[B]

Welcome back, everybody, for another deep dive. This time we're going to be jumping headfirst into all the latest AI news that's got everyone talking.

[A]

Feels like every time you turn around, there's some new AI thing blowing up online.

[B]

I know, right?

[A]

Like, seriously, it's insane.

[B]

It is. And today we've got some great articles lined up for us. We'll be talking about Google's wild new video generating AI, VO2.

[A]

Ooh, yes.

[B]

The latest controversy around Elon Musk and his claims about Grok 3.

[A]

Always a fun time.

[B]

Always a fun time. Plus a peek into what might be the future of home robots with this thing called the Neo Gamma Humanoid.

[A]

Yeah. And what's interesting is that a lot of the articles this week are about stuff that's starting to move from just research into like real world applications. You know, like there's actual costs now and implications that people are having to think about. It's getting real out there.

[B]

Yeah, like those articles a while back about AI writing movie scripts.

[A]

Right.

[B]

And now we've got Google's VO2 that can generate entire videos.

[A]

It's crazy.

[B]

And get this, the article compares the cost, which is, get ready for this, $32,000 per hour of video.

[A]

Oh, wow.

[B]

To the cost of Avengers Endgame, which was about $32,000 per second of film.

[A]

Oh, my gosh.

[B]

Yeah, it's a crazy comparison.

[A]

That is a striking comparison. I mean, you think about it. And yeah, right now this VO2 is probably really aimed at like, productions with huge budgets.

[B]

Right?

[A]

Like big Hollywood studios and advertising agencies.

[B]

Right.

[A]

It's definitely not for, you know, just your average YouTuber.

[B]

Right. Especially when OpenNI is out here offering their Sora model for a flat monthly fee with ChatGPT Pro.

[A]

You know, it is interesting to see how these different pricing models are going to play out, you know?

[B]

Right.

[A]

Like, is subscription based access going to be the thing that like, really opens this tech up to everyone?

[B]

Yeah.

[A]

I mean, could that even put pressure on Google to like lower their price for VO2?

[B]

Right.

[A]

Or is this always going to be like this super fancy high end tool? I guess we'll have to see.

[B]

Yeah, well, and talking about competition and who's claiming to be the best, let's move on to this article about elon Musk and Xai.

[A]

Okay.

[B]

And their model, Grok 3, they're saying it's the world's smartest AI, but there's a bit of, let's call it disagreement over how they got to that conclusion.

[A]

Ah, yes, the benchmark wars always A heated topic in the AI world, always.

[B]

So this article talks about consensus at 64, which, if I'm understanding this right, lets the AI take multiple stabs at a problem and then picks the most frequent answer.

[A]

Right. Like giving it a bunch of tries to get it right. Which, you know, can definitely make the scores look better. But it seems like XAI kind of left that data out when they were comparing Grok 3 to OpenAI's models.

[B]

Yeah, and they even included a graph that makes Grok 3 look way better at first glance.

[A]

Right.

[B]

But then when you look at the consensus at 64 results, some of OpenAI's models are actually doing better on some techs.

[A]

Yeah. And it really makes you think, why wouldn't you include that information?

[B]

Right.

[A]

It just seems kind of shady.

[B]

It does.

[A]

Come on, be transparent, you know? Plus, you know, we don't even know how much computing power each model is using to get those results. It's like maybe Grok3 is actually more efficient even if it doesn't win every time.

[B]

It's like comparing two cars gas mileage without saying one's a tiny compact and the other is a giant suv.

[A]

Exactly. You got to look at the whole picture.

[B]

Exactly. You're not getting the whole story. It's a good reminder to always, you know, dig a little deeper, especially in the AI world, where everyone's always making these big claims.

[A]

Yeah, don't just believe the hype.

[B]

Don't believe the hype. Okay. Well, speaking of competition and trying to be number one, our next article takes us to ByteDance.

[A]

The TikTok folks.

[B]

Yeah, the TikTok folks. And it sounds like they're scrambling a bit in their AI department because of this competitor called Deep Seek.

[A]

Oh, yeah, Deep Seek. They've been shaking things up.

[B]

Yeah. So Deep Seek is making waves with their. And I love this open source approach.

[A]

That's a big deal. Yeah, it's a whole different way of thinking about AI development.

[B]

Okay, so for those who, like me, might not totally get what open source means.

[A]

Oh, sure. So basically, it means they're making their technology freely available for other people to use and modify.

[B]

Okay.

[A]

It's the opposite of how a lot of companies do things where they keep their code super secret.

[B]

Okay.

[A]

But with open source, everyone can collaborate and build on each other's work.

[B]

And it sounds like Deep seap's strategy is working because the article says their chatbot is already beating bytedance's Dubai about in daily users.

[A]

Yeah. Even though it hasn't been around as long.

[B]

It's like David and Goliath in the AI world.

[A]

It really is. It'll be interesting to see how this all shakes out.

[B]

Yeah.

[A]

Could really change how AI is developed, you know?

[B]

Yeah. So far we've been talking about AI in this kind of abstract way.

[A]

Right.

[B]

But now I want to talk about something a little more, I don't know, tangible.

[A]

Okay. I like where this is going.

[B]

The Neo Gamma humanoid robot.

[A]

Okay.

[B]

This thing is being developed by One X Technologies. And I have to say the pictures in this article are one wild.

[A]

I know, right?

[B]

They've really put a lot of work into making this thing look friendly.

[A]

Like they want you to feel comfortable having this thing in your home.

[B]

Yeah, it's got these soft covers for safety and get this, emotive ear rings.

[A]

Wait, what?

[B]

Yeah, like little rings that move to show expressions.

[A]

Oh, that's cute.

[B]

And they even built their own language model so you can like talk to it.

[A]

That's pretty wild. It really is. Like something straight out of a sci fi movie.

[B]

I know. So there's this quote from the CEO of 1x Technologies where he says something like, developing robots in homes, not just in labs, is super important.

[A]

Makes sense, right?

[B]

Yeah.

[A]

You gotta see how it works in the real world, not just in some controlled environment.

[B]

Yeah. Like a robot might be great at navigating a lab, but then totally freak out in a messy house.

[A]

Exactly. Real life is chaotic, but it does.

[B]

Make you think, like, what does it mean to have these things in our homes?

[A]

Right, right.

[B]

Like, what about our privacy or even safety?

[A]

Yeah. And like, how does it change our relationships with each other if we've got robots doing all this stuff for.

[B]

Okay, this is where it gets really interesting. So they mentioned this new language model that One X Technologies deal. What's so special about it?

[A]

Well, it sounds like they're trying to make the robot understand us better.

[B]

Okay.

[A]

Like, most home robots just have a few pre programmed responses. Right?

[B]

Right.

[A]

But this Neo Gamma is supposed to be able to have an actual conversation.

[B]

Whoa, whoa, whoa. Like, hold on, an actual conversation?

[A]

Yeah. Imagine being able to ask it to help with chores or read you the news or just like chat about your day.

[B]

That is both cool and kind of creepy. Like, what happens when these things understand us better than we understand our ourselves?

[A]

Right.

[B]

What if they start like predicting our needs before we even know what they are? That's some next level stuff.

[A]

It is, and it's a valid concern. Like we've already seen with social media how algorithms could be used to like, mess with our emotions and influence our choices.

[B]

Okay. So it's not just about how advanced the technology is, but also about, like the ethical side of things.

[A]

Exactly. We have to make sure these robots are used for good, not for bad.

[B]

And how do we even do that?

[A]

That's the million dollar question, isn't it? And it's one we gotta be thinking about now, before it's too late.

[B]

It seems like with every exciting possibility in AI, there's also a whole bunch of challenges that come with it.

[A]

That's the thing about exploring the unknown, right?

[B]

Right.

[A]

It's exciting and terrifying all at the same time.

[B]

I like that. Exploring the unknown. So if you could leave our listeners with one final thought, what would it be?

[A]

I'd ask them this. If you had the power to shape the future of AI, what kind of world would you create? What values would you focus on? What problems would you solve?

[B]

That's a great question.

[A]

And don't forget about the safeguards. What would you do to prevent things from going wrong?

[B]

It's something we all need to be thinking about. Well, that's all the time we have for this deep dive into AI.

[A]

Thanks for joining us.

[B]

Hopefully you found it informative, thought provoking, maybe even a little bit mind bending.

[A]

And remember, the future of AI is not set in stone. It's something we're all building together.

[B]

So let's make it a future where AI helps us create a better world for everyone. Until next time, keep learning, keep questioning and keep imagining the possibilities.