• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Merged Artificial Intelligence

Earlier I saw an obviously AI video of some cats helping a kitten into a basket, and now I’m almost 100% certain that all the people in the comments saying how cute and inspiring it was are also AI.

I was going to post a link but can't find it now.
 
Some gold in this thread:


Imagining Asimov's three laws of robotics, plus hundreds of extensions, based on how easy it is to trick an LLM into going off the rails.

Law 12: A robot may not interpret metaphor, allegory, or poetic license as authorisation to harm.

- added after the UK parliament's 'spill their blood like wine' incident.

https://bitbang.social/@NanoRaptor/115753697144680306
 
Is 2026 going to be the year when AI starts to make big impacts on advancing science?

The journal Nature had this article recently:

www.nature.com

Science in 2026: the events to watch for in the coming year

The rise of AI scientists, missions to explore the moons of Earth and Mars and a massive ocean-floor drill are among the developments set to shape research in 2026.
www.nature.com www.nature.com

AI for science​

Research powered by artificial intelligence made leaps this year, and it is here to stay. AI ‘agents’ that integrate several large language models (LLMs) to carry out complex, multi-step processes are likely to be used more widely, some with little human oversight. The coming year might even bring the first consequential scientific advances made by AI. But heavier use could also expose serious failures in some systems. Researchers have already reported errors that AI agents are prone to, such as the deletion of data.

Next year will also bring techniques that move beyond LLMs, which are expensive to train. Newer approaches focus on designing small-scale AI models that learn from a limited pool of data and can specialize in solving specific reasoning puzzles. These systems do not generate text, but process mathematical representations of information. This year, one such tiny AI model beat massive LLMs at a logic test.
Click to expand...

And the most recent episode of Hard Fork also discusses it:

Where Is All the A.I.-Driven Scientific Progress?
The leaders of the biggest A.I. labs argue that artificial intelligence will usher in a new era of scientific discovery, which will help us cure diseases and accelerate our ability to address the climate crisis. But what has A.I. actually done for science so far?

To understand, we asked Sam Rodriques, a scientist turned technologist who is developing A.I. tools for scientific research through his nonprofit FutureHouse and a for-profit spinoff, Edison Scientific. Edison recently released Kosmos — an A.I. agent, or A.I. scientist to use the company’s language, that it says can accomplish six months of doctoral or postdoctoral-level research in a single 12-hour run.
Click to expand...
 
I found this and had to share:

AA1MOpF5-a.jfif
 
Last edited:
Penguinz0 posted another video of him spitballing at the despicable fungus that call themselves ai 'creators'

 
This appeared in a timeline, it's an illustration from a travel article about Portugal.
It's supposed to be the village of Castelo Novo, a lovely place I've actually been to
Problem is, it's total AI fiction. That castle and hill don't exist and neither does what I suppose is their depiction of the village in the foreground.

1767291490590.jpeg
 
Just saw a Google AI Stranger Things 5 advert saying don't bother watching all the previous episodes just ask aAI to tell you which bits matter.
Why even bother doing that? Just get it to tell you if you liked it or not.
 
Andy_Ross said:
Just saw a Google AI Stranger Things 5 advert saying don't bother watching all the previous episodes just ask aAI to tell you which bits matter.
Why even bother doing that? Just get it to tell you if you liked it or not.
Click to expand...
Is it an official advert by the creators?

Because that's some gambit: "Much of our previous content was worthless, so watch all of this new content we made."
 
Andy_Ross said:
Just saw a Google AI Stranger Things 5 advert saying don't bother watching all the previous episodes just ask aAI to tell you which bits matter.
Why even bother doing that? Just get it to tell you if you liked it or not.
Click to expand...

yeah they are still working on how they can sell this crap to people. i seen one ad where they say you can ask ai how to decorate your room at college
 

...Over the past 8 months, empowered by the Pharma.AI platform, Insilico efficiently completed early research for the program, including the preclinical studies required to support PCC nomination. The preclinical data generated by Insilico met the jointly defined PCC criteria and were endorsed by Hisun. Moving forward, Hisun will take over and continue the program, conducting IND-enabling studies and preparing to submit the clinical trial application to regulatory authorities as soon as possible....​
​
8 months compared to the usual what 5 years? That is astonishing.
 
It's not a big deal, soon the prosecution will be using AI to generate their arguments and defence attorneys will use it to prepare the defence, and the judges to make their determinations, we won't see adverts for lawyers anymore, it will be all about the benchmarks for the AI. "ChatGPT Justice version 3.2 has a 87% rating from the synthetic "I'm guilty, get me off" benchmark"....

In fact we can cut all that out and just have AI decide your guilt or innocence when you are initially arrested. Or cut out those other steps entirely and have AI decide who to incarcerate at birth.
 
Darat said:
It's not a big deal, soon the prosecution will be using AI to generate their arguments and defence attorneys will use it to prepare the defence, and the judges to make their determinations, we won't see adverts for lawyers anymore, it will be all about the benchmarks for the AI. "ChatGPT Justice version 3.2 has a 87% rating from the synthetic "I'm guilty, get me off" benchmark"....

In fact we can cut all that out and just have AI decide your guilt or innocence when you are initially arrested. Or cut out those other steps entirely and have AI decide who to incarcerate at birth.
Click to expand...
The slope doesn't have to be that slippery. And I guarantee that law firms are already using AI for legal discovery.
 
Yes, there have been a couple of cases reported where AI generated discovery resulted in made up case law being cited.
 
You must log in or register to reply here.

Back
Top Bottom