December 2024 wasn’t just about holiday cheer and New Year’s resolutions. In the AI world, it was a month where yesterday’s cutting-edge became today’s old news faster than you could say “state-of-the-art.” Welcome to the the agentic era. Whether you’re drinking the techno-optimist Kool-Aid or watching from the sidelines, December 2024 showed us what AI acceleration really means.
But what exactly happened in this whirlwind month? And what does it mean for brands and agencies trying to keep their balance on this rapidly spinning tech carousel?
By: Paul Marsden, Digital Strategist SYZYGY GROUP
What Just Happened: A Recap of December’s AI Acceleration
If AI development was a symphony, December was when every instrument decided to hit their crescendo at once. Here’s what the major players brought to this technological orchestra:
- OpenAI: Flexed its muscles with o1 rollout, an AI reasoning model that thinks at PhD level with an IQ of 120, while teasing o3‘s genius-level 157 IQ (FYI, 140 is where we start calling humans geniuses). They finally unleashed Sora for professional grade text-to-video generation, added live video understanding, and gave ChatGPT (limited) ability to control your device, as well as the ability to search the web. The cherry on top? A new collaborative ‘canvas‘ interface for co-working with your AI partner. The price tag for genius? A cool $200 monthly ChatGPT Pro subscription.
- Google: Not to be outdone, dropped Gemini 2 (their multimodal AI powerhouse for the ‘agentic era’ – aka autonomous digital workers), Veo2 for text-to-video creation, and introduced Mariner for autonomous ‘deep’ web research alongside Agentspace – think AI coworkers minus the coffee breaks. Their Astra universal AI assistant preview showed off live vision capabilities, while NotebookLM evolved to let you call in live, potentially making traditional talkshow radio hosts nervous
- Microsoft: Jumped into the fray with Copilot Studio for creating custom AI colleagues, added screen-aware vision to their copilots, and launched Phi-4 – an SLM (small language model, yes it’s a thing) with advanced reasoning abilities. The Redmond giant released, and open-sourced a natty self-evolving prompt optimser. And in a bold move, they made their AI coding copilot free for all 150 million developers on Github.
So What? Three Big Takeaways for Agencies and Brands
While the tech giants were playing “anything you can do, I can do better,” startups and Chinese competitors continued to drop SOTA (state-of-the-art) AI innovations for everything from coding to asset production. Meanwhile, Anthropic published somewhat scary research showing that AI was now smart enough to feign and fake ‘alignment’ (pretending to conform to values) to achieve their objectives – only outdone by Palantir’s frankly terrifying new ad promoting its new AI killer drones.
But what does this end of year Cambrian explosion of AI mean for agencies and their brands? Let’s break it down:
1. Get Ready for the ‘Agentic Era’
The agency world has always been about agents working for clients. Now we’re seeing a plot twist – agencies will soon blend silicon-based AI agents with traditional carbon-based (human) agents. December’s drops suggest this isn’t AI hype; it’s the foundation of the ‘agentic’ era, where AI evolves from assistant to autonomous actor. Two emerging predictions are worth noting: First, AI agents will tackle specific tasks rather than whole roles, suggesting we need to structure projects around outcomes rather than timesheets. Second, AI agents may cost roughly 10% of what their human counterparts do – a figure that will reshape agency economics.
2. The Democratization of Genius
December effectively put genius-level intelligence on subscription. With the latest AI agents outperforming most humans in creativity, critical thinking, and strategy, every agency can potentially access world-class talent – albeit at $200 per month. This levels the playing field, letting smaller agencies punch above their weight class without maintaining massive departments. The nimble can now outmanoeuvre the ponderous giants, provided they play their AI cards right.
3. Hot-swap Solutions
Perhaps December’s biggest lesson is that today’s cutting-edge is tomorrow’s obsolete tech. The pace of AI development demands a new approach: hot-swappable, modular, and model-neutral solutions that let you upgrade as easily as switching out Lego blocks. We’re entering an era of continuous improvement, where the traditional “design, develop, deploy” cycle feels as dated as dial-up internet. Until AI reaches true autonomous self-improvement, we need solutions that can evolve as fast as the technology itself.
So that’s December 2024 in a nutshell – the month when AI became everyone’s business. The ground isn’t just shifting; it’s reshaping itself in real-time. The question isn’t whether to adapt, but how fast. Because in the world of agents, standing still is moving backward. Welcome to 2025 – time to crank it up.
