Artificial General Intelligence (AGI) News and Discussions

Vakanai
Posts: 313
Joined: Thu Apr 28, 2022 10:23 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by Vakanai »

wjfox wrote: Wed Jan 04, 2023 6:51 pm A Princeton student built an app which can detect if ChatGPT wrote an essay to combat AI-based plagiarism

Pete Syme
Jan 4, 2023, 10:46 AM
  • A Princeton student built an app that aims to tell if essays were written by AIs like ChatGPT.
  • The app analyzes text to see how randomly it is written, allowing it to detect if it was written by AI.
  • The website hosting the app, built by Edward Tian, crashed due to high traffic.
A new app can detect whether your essay was written by ChatGPT, as researchers look to combat AI plagiarism.

Edward Tian, a computer science student at Princeton, said he spent the holiday period building GPTZero.

https://www.businessinsider.com/app-det ... ism-2023-1
I approve of this for one simple reason - it gives us a new benchmark goal, creating a new ai that passes as human to this app.
User avatar
wjfox
Site Admin
Posts: 8733
Joined: Sat May 15, 2021 6:09 pm
Location: London, UK
Contact:

Re: Proto-AGI/Transformative AI News and Discussions

Post by wjfox »

User avatar
wjfox
Site Admin
Posts: 8733
Joined: Sat May 15, 2021 6:09 pm
Location: London, UK
Contact:

Re: Proto-AGI/Transformative AI News and Discussions

Post by wjfox »

User avatar
wjfox
Site Admin
Posts: 8733
Joined: Sat May 15, 2021 6:09 pm
Location: London, UK
Contact:

Re: Proto-AGI/Transformative AI News and Discussions

Post by wjfox »

User avatar
Yuli Ban
Posts: 4631
Joined: Sun May 16, 2021 4:44 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by Yuli Ban »

And remember my friend, future events such as these will affect you in the future
User avatar
Yuli Ban
Posts: 4631
Joined: Sun May 16, 2021 4:44 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by Yuli Ban »

Spooky

Getting similar vibes to:


Possibly, they realize now that AGI is nowhere near as far away as once believed, which would distress shareholders who had been reassured for years now that AGI was "decades or even generations away"

Society is very clearly not ready for AGI. As much as I want it now (and I certainly would rather accelerated timelines), unleashing AGI right now would be akin to unleashing the atomic bomb in the 18th century, so treading carefully is necessary.
And remember my friend, future events such as these will affect you in the future
User avatar
Ozzie guy
Posts: 486
Joined: Sun May 16, 2021 4:40 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by Ozzie guy »

Yuli Ban wrote: Fri Jan 13, 2023 7:10 am Spooky

Getting similar vibes to:


Possibly, they realize now that AGI is nowhere near as far away as once believed, which would distress shareholders who had been reassured for years now that AGI was "decades or even generations away"

Society is very clearly not ready for AGI. As much as I want it now (and I certainly would rather accelerated timelines), unleashing AGI right now would be akin to unleashing the atomic bomb in the 18th century, so treading carefully is necessary.
Some key differences

Unlike the government OpenAI can't Epstein people for releasing secrets which I am sure workers of the nuclear program would be fearful of.
OpenAI has less resources and capacity in general to prevent leaks.
The kinds of people who work on AGI strongly overlap with the kinds of people who want AGI released to "save their sister from cancer, stop themselves dying of old age etc"
Pushback from the public and government institutions is nothing compared to push back if nukes were leaked.
We have social media and other tech that makes it easy to spread leaks and hard to cover them up.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by funkervogt »

ChatGPT can (and will) be considerably improved.

Stephen Wolfram shows how it could generate more accurate answers to many types of questions if it uses Wolfram Alpha.
But why does ChatGPT get this particular thing wrong in the first place? If it had seen the specific distance between Chicago and Tokyo somewhere in its training (e.g. from the web), it could of course get it right. But this is a case where the kind of generalization a neural net can readily do—say from many examples of distances between cities—won’t be enough; there’s an actual computational algorithm that’s needed.

The way Wolfram|Alpha handles things is quite different. It takes natural language and then—assuming it’s possible—it converts this into precise computational language (i.e. Wolfram Language), in this case:
https://writings.stephenwolfram.com/202 ... o-chatgpt/

There's also this:
Many people mistakenly think that large language models that generate one word at a time is the end game.
Connecting LLMs with tools, e.g.: search engines, python interpreters, etc., is super exciting, and leverages tools' power, robustness & more!


A few takeaways from this:

1) There's still a lot of low-hanging fruit to be picked to improve ChatGPT's performance.
2) People who write news columns and post transcripts on the internet dedicated to exposing ChatGPT's mistakes and limitations are idiots. It will rapidly improve, as in, it will be noticeably better every six months and won't make most of the old mistakes anymore. It will get harder and harder to trip it up.
3) A chatbot that can't pass the Turing Test could still be an incredibly powerful resource with many effects on the economy, society, and daily life.
agi
Posts: 21
Joined: Tue Apr 05, 2022 3:03 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by agi »

deleted
Last edited by agi on Tue Apr 04, 2023 12:28 pm, edited 1 time in total.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: Proto-AGI/Transformative AI News and Discussions

Post by funkervogt »

agi wrote: Fri Jan 20, 2023 2:13 pm AGI by 2024, the hard part is now done

https://medium.com/@paul.k.pallaghy/cha ... 79d31a7277
But look at how widely he defines an "AGI."
Post Reply