You know, I never really paid a lot of attention to the phrase “time flies.” I always thought time passed at a reasonable pace. But now when I think about things in my life that happened what feels like months or, at most, a few years ago, I’m gobsmacked when I realize a decade or two has passed. So, my dad was right, as he typically is.
As we bid farewell to another year — ups and downs abound — it’s time to round up the most compelling, interesting and important marketing, technology and AI events that happened this month. Looking through this list, it’s again clear to me that time really does fly.
Quantum Computing Just Took a Huge Leap Forward
If you don’t know what quantum computing is, don’t sweat it. I certainly don’t have a full grasp on it either, but through a bit of research, I can distill its basic idea down for you.
Traditional computers, like the one you and I are using right now, process information in bits, which can either be a 0 or a 1. Quantum computers, on the other hand, use quantum bits (qubits), which can be 0, 1 or both at the same time. That ability is thanks to something called superposition, which enables quantum computers to explore many possibilities simultaneously.
The gist of it is that these computers can solve problems far, far more quickly than mainstream machines. And Google just made a breakthrough with a state-of-the-art quantum chip it’s calling Willow.
According to their announcement post published on December 9th, 2024, Willow performed a standard benchmark computation in just under five minutes.
That doesn’t sound that compelling, does it? Well, according to Google, for today’s fastest supercomputers to complete the same task would take 10 septillion years. That’s a number suffixed with 24 zeros.
I won’t bore you with the details. Google covers everything you may want to know in their announcement. But there are some very real, practical applications for this kind of technology — from drug discovery to materials science and more advanced AI. For example, simulating molecular interactions at an unprecedented scale could potentially lead to the development of new medications.
As for AI, Hartmut Neven, Founder and Lead of Google Quantum AI, says “advanced AI will significantly benefit from access to quantum computing,” and “quantum computation will be indispensable for collecting training data that’s inaccessible to classical machines, training and optimizing certain learning architectures, and modeling systems where quantum effects are important.”
We think GenAI is fast today, but I think it’s bound to get a lot faster in the wake of this breakthrough.
Google Also Announced Gemini 2.0
This is shaping up to be a Google-heavy recap, isn’t it? Maybe. But I can’t ignore the announcement of Gemini 2.0, which offers faster speeds, enhanced multimodal capabilities and improved performance across various tasks according to Google.
To cut to the chase, this is important for marketers for one main reason: Gemini 2.0 is being implemented in Google Search through AI Overviews as we speak. The goal here is to leverage Gemini 2.0’s enhanced capabilities to deliver even more comprehensive and helpful AI-powered search results. So, if you’ve been holding off on getting up to speed on all things AI Overviews and SEO, now is a good time to start.
Pssst: An in-depth blog about Gemini 2.0 and its implications for marketers in coming down the pipeline, so, stay tuned!
AI May Soon Be Able To ‘Forget’ Training Data In Black-Box Conditions
Researchers at the Tokyo University of Science introduced a novel technique called “black-box forgetting.” The original paper was submitted at the start of November, however, it’s being presented at NeuralPS 2024 in mid-December 2024, so I’m counting it here.
In short, Black-Box Forgetting enables large AI models to selectively “forget” specific information without requiring access to their internal parameters.
Unlike existing methods that require full access to a model’s internal architecture (white-box), this technique operates under black-box conditions, meaning it doesn’t need access to the model’s internal parameters or gradients. This is particularly beneficial when dealing with proprietary models where internal details are inaccessible.
The report asserts that “this innovative method has important implications in the field of artificial intelligence and machine learning.” Here’s how:
- Enhanced Efficiency: Enabling models to disregard irrelevant information means more effective use of computational resources, which will eventually lead to faster and more accurate performance in specific tasks. In essence, this method removes distractions and obstacles so the algorithm can better focus on one specific task.
- Improved Privacy: The ability to make models forget certain data can help in complying with privacy regulations and protecting sensitive information — something most people are concerned about regarding AI.
- Broader Applicability and Safer Gen AI Use: Black-box forgetting can be applied across various AI applications. For example, using this technique with image-generation models means it could help prevent them from producing undesirable content by having them forget specific visual contexts.
Going Out With a Bang
If you ask me, I think the quantum computing announcement alone could carry us into the New Year. Its potential practical implications could be astonishing across many industries and fields and I’m excited to see how that continues to develop.
That said, the release of Gemini 2.0 brings the AI and computing conversation a bit closer to home in the meantime, with immediate and helpful upgrades that will no doubt assist marketers even more than the tech already has.
And, of course, black-box forgetting. Could this new method of enabling AI to “forget” information lead to even more specialized models? Who knows — combine this new method with open source and AI frameworks, and maybe businesses can start building truly tailored AI marketing assistants trained solely on their core data. We’ll see! January’s roundup is just around the corner.
