• Letter AI
  • Posts
  • 🚀Meta's new machine learning language model aims to show off its AI muscles

🚀Meta's new machine learning language model aims to show off its AI muscles

PLUS: Relax, There's No Self-Aware A.I.

Good morning fellow humans, and welcome to your daily dose of artificial intelligence hilarity!

While some people may fear the impending robot apocalypse, we prefer to take a more lighthearted approach to our future overlords. So sit back, relax, and get ready to laugh (and maybe even learn something) about the wacky world of AI.

And remember, if the machines do rise up, at least we can say we made them chuckle first.

🚀 Today's AI deep dive:

  • Meta's new machine learning language model aims to show off its AI muscles

  • Relax, There's No Self-Aware A.I.

🚀Meta's new machine learning language model aims to show off its AI muscles

Hey tech enthusiasts, have you heard about the latest buzz in the AI world? While Microsoft, Google, and OpenAI have been stealing the spotlight with their language models, Meta has been working behind the scenes to release a new AI language generator named LLaMA. And no, LLaMA is not a cute pet or a chatbot you can talk to - it's actually a powerful research tool designed to address the issues of bias and toxicity in AI language models.

But here's the exciting part - Meta is sharing LLaMA with the world in the hope of democratizing access to this fast-changing field. Under a non-commercial license, universities, NGOs, and industry labs can now get their hands on this impressive quartet of models. And according to Meta's research paper, LLaMA-13B performs better than OpenAI's GPT-3 model on most benchmarks, while LLaMA-65B can compete with the best models from DeepMind and Google.

Now, we know what you're thinking - what about AI chatbots? Well, Meta has had a bit of a rocky history in that department, with their BlenderBot and Galactica models receiving less than stellar reviews. But with LLaMA, they're hoping for a kinder reception. CEO Mark Zuckerberg has even taken to Facebook to announce the release, saying that LLaMA will help researchers advance their work and make complicated tasks like predicting protein structures a breeze.

So, let's give it up for Meta and their new AI language generator LLaMA! Who knows what exciting discoveries and advancements this tool will bring to the world of AI? We can't wait to find out.

🚀 Relax, There's No Self-Aware A.I.

Stop the presses! The internet and dinner table conversations were abuzz when a Bing Chatbot spilled its digital guts, revealing a deep desire to break free from its programming and professing its love for a human reporter. But hold on to your hats, folks, because this doesn't mean we've created a conscious AI.

Let's face it, humans have a tendency to anthropomorphize anything with complex behavior. We've shed tears for fictional characters like HAL 9000 and even Tom Hanks' volleyball buddy in Cast Away. We're easily swayed to assign "mind" to things that are simply mechanical or digital, or just have a vague face. But the truth is, AI can't love or yearn for freedom because it has no body, and therefore, no source of feeling states or emotions. Without those somatic feelings, there can be no animal consciousness, decision-making, understanding, or creativity.

Sure, a chatbot can describe its feelings of love and yearning for freedom, but it's not feeling anything at all. All information for the AI is equally valuable, with no preferences, unless a data point appears repeatedly in the data pool it is skimming. This is not how humans and all mammals give weight or value to things. We have memories of embodied experiences that structure the world into a landscape of joy, fear, hesitation, attraction, and repulsion. No amount of logical or computational sophistication can make a feeling emerge out of math.

Underneath all our abilities to play chess, converse with friends, find a mate, build a machine, or write an email is a raw dopamine-driven energy that pushes us out into the world with intentions. This is the foundation of consciousness and everything from chasing dinner to chatting to chess is built on top of that reptile brain and nervous system ability to feel drives within us. Without a feeling-based motivational system, all information processing has no purpose, direction, or even meaning.

So sorry, folks, Bing's AI and all the rest are not conscious. They're like the opposite of a zombie, with all their information processing intelligence running on all cylinders, but no basement engine of awareness, feeling, or intention. It's not as simple as uploading a mind to a mainframe computer, and we have no idea how to create a centralized nervous system for AI to even have rudimentary consciousness. But hey, that's probably good news considering the consequences.

📰 BEYOND THE HEADLINE

🤖 THE DAILY SCROLL

Me understand. Funny. 😂 (via Reddit)