Theta EdgeCloud has recently announced a new chapter in their partnership with the Vegas Golden Knights. This time around they are hosting an EdgeCloud based chatbot set to appear on the Sin City hockey team’s homepage. It has not yet been released and reportedly still in beta testing according to the chief of tech, aka the Sultan of Swing: Jieyi Long.
Theta Labs CEO Mitch Liu’s comments in the press release make mention that the LLM bot is Gen AI – broadly speaking – but also utilizes “agentic AI” elements layered atop of it.
Ok, what is Agentic AI?
Agentic AI refers to AI systems that possess a degree of autonomy and can act on their own to achieve specific goals.
Hmm, sounds ominous. What goals might those be?
Unlike traditional AI models that simply respond to prompts or execute predefined tasks, agentic AI can make decisions, plan actions, and even learn from its experiences – all in pursuit of objectives set by its human creators.
Well, this shouldn’t be all that surprising. Gen AI all have some level of prescribed outcomes. The chatbot appearing on the EdgeCloud AI Showcase is specifically: Meta-Llama-3-8B-Instruct. This flavor of Meta Llama specializes answering complex questions, instructively on a wide range of topics. Its improved accuracy, better understanding of instructions, and increased versatility make it a valuable tool for a wide range of applications.
In a word – a generalist – without much prescribed outcomes other than to inform the user. Other broad based versions of LLama include:
- Meta Llama 3.8B-Finetuned: This version of Meta Llama 3 is fine-tuned on a specific dataset or task, such as conversational dialogue or text classification. This fine-tuning process allows the model to learn more specific patterns and relationships in the data, making it more accurate and effective for that particular task.
- Meta Llama 3.8B-Large: This version of Meta Llama 3 is a larger version of the base model, with more parameters and a larger training dataset. This model is useful for tasks that require more complex and nuanced understanding of language, such as natural language processing and machine translation.
- Meta Llama 3.8B-Transformer: This version of Meta Llama 3 uses a transformer architecture, which is a type of neural network that is particularly well-suited for natural language processing tasks. This model is useful for tasks that require more complex and nuanced understanding of language, such as machine translation and text summarization.
- Meta Llama 3.8B-BERT: This version of Meta Llama 3 uses a BERT (Bidirectional Encoder Representations from Transformers) architecture, which is a type of neural network that is particularly well-suited for natural language processing tasks. This model is useful for tasks that require more complex and nuanced understanding of language, such as question answering and sentiment analysis.
Of course, the Golden Knights want to sell more tickets and more merchandise. That’s the obvious reality of capitalism – increase revenue, increase profits.
But on the softer side of salesmanship, we can speculate that the chatbot will also seek to deepen relationships with it’s users. To befriend the user – and ascertain their wants and needs. Finding the correct angles for wit, comedy, knowledge of hockey history, specific favorite players, achievements, reminiscing about moments in time for the franchise and fan alike.
These will all be baked into the LLM’s programming. We can envisage the bot having a certain level of decorum in their soft-selling. Again, more focused on building a meaningful relationship with the specific user. And less obtusely, a ticket and merchandise selling machine.
As we all move along into the uncertain future of generative AI, these biases will become more and more entrenched in what will likely become a huge population of chatbots and AI interfaces.
Digital Natives are often talked about as someone who has grown up with digital technology and is comfortable using it in their daily lives. Milennials specifically sit at the cross-roads of many digital technologies but can also remember the time of analog. Boomers and Gen X certainly know of the old times as well.
But what about these newer generations of Gen Z and beyond. Growing up with iPad’s in their cribs. Having smartphones since age eight. Being locked-in to social media at such young ages. So many nuanced layers of comprehension about algorithms and what they are implicitly conveying about what has value. An addiction to likes, reposts, interactions.
Now comes a new generation of Generative AI, with it’s own pre-ordained biases to “achieve certain goals.” Whether that’s a chatbot or a text-to-image generator, image-to-video generator. What will the values of these varieties of AI convey?
Admittedly, it’s not all entirely nefarious. Hockey nerds wanna talk puck. And they’re going to get that with this bot. Nerds deserve an outlet for their nerd-dom. The LLM will provide.
The perhaps darker future which needs more interrogation is what this does to the human communication. Is general conversation as we know it headed to a more function oriented place – where the user is switching from bot to bot in seeking a particular support or sounding block for their current line of questions?
Where does this place the cognitive processes of people really thinking through things on their own. I’m loathe to conclude this article with a note of: “moderation is key to utilizing generative AI.”
But, that’s really up to each of you.
Go big. Go small. Ignore it if you want. But much like the dominance of search and social media over the last 20 years, Gen AI is about to become ubiquitous in a big way. We ought to begin to understand the biases in these programs while they are still in their infancy.
It’s probably all good for Theta and their future Edgeclouds. It may not be bullish for a future thoughtful generation of humanity going forward. Or maybe it’ll take our evolution even further and enlighten us in way we can’t yet imagine. But do choose carefully which bot you’re putting your secrets and your trust.