Bakz T. Future
Vova – @VovaZk
https://twitter.com/vovazk
http://gptify.com/
Are we Wasting the Potential of GPT-3 and AI Technologies?
https://bakztfuture.substack.com/p/are-we-wasting-the-potential-of-gpt
Original clubhouse event link:
https://clublink.to/event/PbDz8DGG
—
► Remember to Like, Comment, and Subscribe!
—
Connect with me:
Substack Newsletter – https://bakztfuture.substack.com
Twitter – https://www.twitter.com/bakztfuture
Instagram – https://www.instagram.com/bakztfuture
Github – https://www.github.com/bakztfuture
Feel free to send me an email or just say hello:
bakztfuture@gmail.com
The idea that nobody’s doing good stuff with GPT3 is ridiculous when the developers can’t get access.
if this is the case why don't you build what you want to exist? it's hard, really really hard to think through something new and it takes longer than a year
Instead of altruism understand Open Source.
Where money is made by providing a service,. Where liabilities can be disclosed. Where it can attract more "general" people. Where it can be transparent and provide true privacy… etc.
Many of your "altruistic" wishes are resolved by it.
But don't over estimate GPT-3, as it is stated in the video, is just a NLP tool, it's nowhere to be a philosopher !
Isn't my idea, post 2 videos ago, about helping elderly person's loosing their memory (Alzheimer), not a worthy thing to tackle, I don't mean to cure it but to mitigate it's effect, no need for a supervising doctor needed.
I'm playing around with some BERT code, where you can ask the "bot" questions that are include in a text, having a kind of local (because of privacy issue) summarisation…
I can imagine myself having a little robot parrot on my shoulder, listening to my conversation (local ASR/TTS), helping me out when I loose the thread (NLP) and some YOLO image recognition for searching my misplaced glasses…
Well, it's not exactly super-intelligent to begin with. So to have a sub-human AI with super-human data, solve problems that human intelligence has worked on for thousands of years and a whole religion based on it (Buddhism)?
I guess that the best usage of GPT-3 is to find verbal problems with repetitiveness and low difficulty and without it being any high-stakes problems, of course.
It's great to think about, and maybe we can find some altruistic applications for it, but I doubt GPT-3 will be the first AI to really have a great positive impact on society.
Well, anyways, I enjoyed the chat! Altruism is a huge complex subject. So it certainly makes you think.
thanks for sharing man. cool stuff
Openai has to prove to the masses that their model is capable and able to carrying out the expected NLP tasks especially in business or their premise to AGI or whatever it is will be a catalyst for a new AI winter, so far it needs fine tuning and complicated engineering prompts, I hope in the future this will become more and more easier
Google just made a 1.3Trillion parameter model called "Switch Transformer".
Use GPT3 to generate altruistic GPT3 applications. Prompt it by brainstorming a few.