Barely two years after OpenAI introduced ChatGPT (and sparked a global fascination with artificial intelligence), pre-Christmas 2024 has brought a veritable blizzard of AI releases.
The rivalry between OpenAI and Google has taken centre stage, with both firms unveiling significant advancements in the scope and functionality of their respective platforms.
According to the reviewers, Google emerged as the clear winner, thanks to its groundbreaking video generation tool, Veo 2, which outshone OpenAI’s competing offering, the long-awaited Sora platform.
But there was more to the search giant's offering than just video.
It has peppered us with a Gemini upgrade, Whisk, an innovative image generation tool, and Willow, a quantum computing chip that promises to revolutionise complex calculations.
As Nate Jones, a social media follower of all things AI, colourfully observed: "The hunt for AGI (artificial general intelligence) is like the battle for the Iron Throne — a high-stakes game where only the most powerful survive.....
“Why the Game of Thrones reference? Google’s comeback reminds me of Daenerys Targaryen rising from exile to claim the Iron Throne."
A Packed month
The pre-Christmas period saw both OpenAI and Google vying for attention with a series of high-profile launches.
OpenAI kicked off its “12 Days of OpenAI” campaign on December 5, introducing a string of tools and updates designed to solidify its position as a leader in conversational AI and creative tools.
Google followed suit with a December 11 launch of multiple innovations, spearheaded by Veo 2, part of its rapidly advancing AI portfolio.
OpenAI’s goes for versatility
OpenAI’s headline release was the o1 Model, a major upgrade that enhanced reasoning capabilities and promised more human-like conversational skills.
This marked a significant step forward in the AI’s ability to handle complex queries with speed and precision. Alongside the o1 Model came the premium ChatGPT Pro Subscription, priced at $200 per month, aimed at professional users seeking advanced features such as the o1 Pro mode for more demanding tasks.
In the creative domain, OpenAI transitioned its Sora AI video generator from research preview to general availability.
Designed for marketers and creators, Sora promised accessible text-to-video generation, enabling users to craft short videos from written prompts or uploaded files.
“OpenAI’s Sora is a step forward in making AI creative tools more accessible,” noted TechCrunch, adding, “Its ease of use and intuitive design make it appealing, but there’s room for improvement in terms of output realism.”
Google pushes the button(s)
Google, owned by Alphabet (NASDAQ:GOOGL) Inc (NASDAQ:GOOG), countered OpenAI’s campaign with a flurry of its own releases, led by Veo 2, a next-generation video generator praised for its stunning realism.
Capable of producing 4K resolution videos, Veo 2 surpassed OpenAI’s Sora in nearly every technical metric.
The Verge called Veo 2 “a game-changer in video generation, with outputs that are so realistic they blur the line between artificial and real-world footage.” The tool’s ability to incorporate physics-based motion and lifelike textures left reviewers and users alike in awe.
Social media was filled with side-by-side comparisons of Veo 2 and Sora outputs.
One user remarked: “The difference is night and day. Veo 2 feels like watching real footage, while Sora’s videos, though creative, lack grounding in reality.”
Accompanying Veo 2 was Gemini 2.0, an advanced multimodal AI capable of processing text, images, and audio. Its experimental version, Gemini 2.0 Flash, offered cutting-edge speed and versatility.
Google also introduced Whisk, an AI-powered image generator that allowed users to create visuals using reference images instead of text prompts.
Further reinforcing its technical prowess, the Silicon Valley behemoth launched Willow, a quantum computing chip that can perform calculations in minutes that would take classical computers billions of years.
“Willow is a leap forward, not just for Google but for quantum computing as a whole,” wrote Wired.
Veo 2 vs. Sora: A decisive victory for Google
The head-to-head comparison between Veo 2 and Sora became the most talked-about aspect of the December AI launches. Reviewers overwhelmingly favoured Google’s offering.
“Sora is imaginative and user-friendly, but it can’t compete with Veo 2’s realism,” noted TechRadar. “Veo 2 doesn’t just generate video—it redefines what’s possible in the space, with a level of detail and physics accuracy unmatched by any other tool.”
While Sora excels in creative applications such as storytelling and remixing, its limitations became evident when users sought realistic outputs.
As one analyst pointed out: “Sora’s videos often feel overly stylised and artificial. They’re fun and creative, but Veo 2 delivers realism that’s in a class of its own.”
Users echoed these sentiments. A widely shared Twitter post read: “Veo 2 isn’t just an AI video generator—it’s like having a cinematographer in your computer. The realism is jaw-dropping.”
Both companies are pursuing distinct strategies. OpenAI’s focus remains on versatility and user accessibility, while Google appears to be prioritising technical excellence and realism in its outputs.
“There’s no doubt that OpenAI and Google are driving the next wave of AI innovation,” said a report from Forbes. “Their December launches show just how competitive this space has become—and how quickly the technology is advancing.”
2025: Year of the agent?
Going into the new year commentators will be throwing around the word 'agentic'. Put simply, this means they are able to perform complex tasks autonomously.
OpenAI's o1 model, for instance, employs a "chain of thought" approach, improving reasoning and decision-making processes. Similarly, Google's Gemini 2.0 introduces AI agents capable of multi-step problem-solving and autonomous task execution.
Already we are seeing AI that automates the sales process from prospecting to booking calls with the business development teams.
Savvy social media outfits, using tools such as Make and Zapier, are able to create and distribute content at the press of a button.
Next (LON:NXT) year will likely see this trend move at light speed as companies look to cut costs and extract the human element from mundane administrative tasks.
Analysts predict a significant increase in enterprise deployment of AI agents. Indeed, Deloitte forecasts that 25% of enterprises using generative AI will implement AI agents by 2025, with this figure rising to 50% by 2027.
And what about AGI?
Again the o1 model marks a shift towards AI systems with improved reasoning abilities.
However, while it represents a notable advancement, it does not yet achieve the comprehensive understanding and adaptability characteristic of artificial general intelligence (AGI).
Experts continue to debate the timeline for achieving AGI. Some, like futurist Ray Kurzweil, predict its arrival by 2029, while others anticipate a longer trajectory.
The consensus suggests that, despite rapid advancements seen in December, AGI is not imminent and requires further breakthroughs in AI research and development.