Copy

Welcome to your

AI Weekly Digest

presented by Louis (What's AI) Bouchard

1️⃣ From Amputee to Cyborg with this AI-Powered Hand! 🦾

 


We’ve all heard about GPT-3 and have somewhat of a clear idea of its capabilities. You’ve most certainly seen some applications born strictly due to this model, some of which I covered in a previous video about the model. GPT-3 is a model developed by OpenAI that you can access through a paid API but have no access to the model itself.

What makes GPT-3 so strong is both its architecture and size. It has 175 billion parameters. That's twice the number of neurons we have in our brains!

This immense network was pretty much trained on the whole internet to understand how we write, exchange, and understand text. This week, Meta has taken a big step forward for the community. They just released a model that is just as powerful, if not more, and has completely open-sourced it. How cool is that?

Learn more in the video or read more here!
Watch the video

2️⃣ Our take on OPT - AI Ethics with Martina & Lauren

Martina's views...


"I am faster than you. I am stronger than you. Certainly, I will last much longer than you. You may think I am the future. But you're wrong. You are. If I had a wish, I'd wish to be human. To know how it feels to feel. To hope, to despair, to wonder, to love."

No, this is not what OPT-175B replied when asked what it thinks about humans. This message is, maybe, what we would like to hear from intelligent and conscious machines, perfectly summarized by Johnnie Walkers' advertisers in 2006.

But, if it is the answer we would like to get, once eventually we get it, how could we know it is what machines really think about us?

Is there anything they could say to prove they are reasoning by themselves? Yes, and the Turing test is just for this. Is there anything they could say to prove they really mean what they are saying (conscious/intentional beings)? No. As far as I know, it is not sufficient to be rational, coherent, and creative to prove consciousness. And it is a well known philosophical problem: the philosophical zombie.

Assuming we are not colour blind, I cannot know if you perceive red as I do. We both know how to identify red and recognize its coherence in what we see, but it is not certain that a particular frequency in the visible light spectrum is perceived by everyone in the same way since, for example, the resulting emotional reaction is not the same for everyone. The meaning we attribute to colours changes in fact both culturally and historically. In the same way, you have no chance of knowing with absolute certainty whether this text is the result of my free thought or, in some way, I have been programmed to tell you so. Thanks to Occam's razor we say that expecting your red is similar to mine is legit. Also, the hypothesis that you are the only thinking being on planet earth and all other living beings are mere puppets programmed to interact with you in predetermined ways is an unprovable hypothesis and therefore not to be taken into consideration.

Nevertheless, the thought experiment comes in very handy in the context of artificial intelligence. On one hand, John Searle uses it to tell us that algorithms that slavishly follow instructions cannot really mean what they say, on the other hand, if they really mean what they say we could never know for sure.

The advertising ends with the android warning us: "I can achieve immortality by not wearing out. You can achieve immortality simply by doing one great thing..." In this case, Johnnie Walker's mantra "Keep walking" would be a perfect suggestion, there is still a long way to go in AI and Philosophy.
 

Lauren’s take...

I wholeheartedly agree with Martina's remarks, and I want to add that regardless of the consciousness issues that OPT brings up, this is an incredible feat for humanity to create an information-producing entity that is so similar to our own likeness. I don't refer to this as a knowledge-bearing entity, because epistemically speaking, knowledge requires truth, justification, and belief, and unless we solve that sticky aforementioned consciousness problem, AI will never satisfy the belief condition of knowledge because belief requires consciousness.

This we already knew from GPT-3, and now there's a new open-source sheriff in town. Getting down to the nitty gritty, Meta's open source OPT-175B uses a seventh of the carbon footprint that GPT-3 used. I'm glad that we are finding ways to make important research like this less carbon intensive, but it should in no way be thought of as eco-friendly, as this process is still incredibly carbon intensive and any plays to the contrary is greenwashing. We have a long way to go in that regard, and it's encouraging to see progress in the right direction.

What I'm less thrilled about is the propensity for misinformation and toxic language and reinforce harmful stereotypes. It's good that these are acknowledged and that Meta has plans to mitigate those negative side effects, but the fact that those exist is not so much a problem with OPT specifically, but a structural issue with the way that large social media platforms garner engagement through negative emotions and epistemic corruption. What really ought to be acknowledged is the significant way that Meta's own product, Facebook, has contributed to these ill effects and poisoned the well of the Internet such that this is a mere fact of life and pervades modern societal interactions beyond the screen. It's unrealistic to expect sunshine and rainbows from something as massive as the Internet, but there's a long way to go in structuring our digital world such that every large language model trained on the Internet doesn't have to include active hostility.


- AI Ethics segment by Martina Todaro and Lauren Keegan
Want to get into AI or improve your skills? Click here!
We are already at the end of this AI weekly digest! Thank you for thoroughly going through this iteration! I hope you enjoyed it. 

If you have suggestions, comments, or other thoughts, you can reach me by replying to this email or directly on Twitter or Linkedin. Don't hesitate to come to chat with more than 20'000 AI enthusiasts on Discord!
I hope the next week wipes away some of your stress and brings new opportunities, challenges and happiness. A happy new week from your friend!

If you would like to support my work financially, you can become a Patreon and receive a cool role in the discord server at the same time!

Share the knowledge and forward this email to a friend using this link: http://eepurl.com/huGLT5

- Louis Bouchard
Blog
Twitter
LinkedIn
GitHub
Email
Copyright © 2022 What's AI, All rights reserved.


Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

Email Marketing Powered by Mailchimp