AI & Robotics News and Discussions

[funkervogt]

[funkervogt]

[agi]

deleted

[wjfox]

Daily Mirror publisher Reach reveals AI is already writing articles

1 hour ago

Daily Mirror publisher Reach published its first AI-written articles last week, CEO Jim Mullen revealed today.

During Reach’s 2022 earnings call, Mullen said the publisher was already using artificial intelligence to write articles, after the business announced it was considering the possibility last year. The first AI-written articles have appeared on local news site In Your Area last week.

“There are three AI-written articles about Newport for In Your Area,” Mullen said. “One of them was ‘Seven Things to do in Newport’: that was written by a bot.”

Mullen said, though, that Reach’s current AI strategy still also involved human journalists.

“It was based off of content from our journalists, and the decision to run it was made by the editor: that is responsible use of AI,” he said.

https://www.standard.co.uk/business/dai ... 65292.html

[firestar464]

https://www.nature.com/articles/d41586-023-00641-w

Quality > quantity

[spryfusion]

[caltrek]

Soft Robot Developed That Easily Goes From Land To Water
by Aaron Aupperlee
March 16, 2023

Introduction:

(Futurity) Most animals can quickly transition from walking to jumping to crawling to swimming if needed without reconfiguring or making major adjustments. But most robots can’t.

“We were inspired by nature to develop a robot that can perform different tasks and adapt to its environment without adding actuators or complexity,” says study co-first author Dinesh K. Patel, a postdoctoral fellow in the Morphing Matter Lab in the School of Computer Science’sHuman-Computer Interaction Institute at Carnegie Mellon University.

“Our bistable actuator is simple, stable, and durable, and lays the foundation for future work on dynamic, reconfigurable soft robotics.”

The bistable actuator is made of 3D-printed soft rubber containing shape-memory alloy springs that react to electrical currents by contracting, which causes the actuator to bend. The team used this bistable motion to change the actuator or robot’s shape. Once the robot changes shape, it is stable until another electrical charge morphs it back to its previous configuration.

“Matching how animals transition from walking to swimming to crawling to jumping is a grand challenge for bio-inspired and soft robotics,” says Carmel Majidi, a professor in the mechanical engineering department.

Read more here: https://www.futurity.org/soft-robot-lo ... 2891132-2

[firestar464]

[spryfusion]

[wjfox]

[funkervogt]

In this interview, Sam Altman says that, while inelegant, LLM's like the GPT could become capable of AGI.

[Yuli Ban]

Bucking popular consensus, I'm not 100% convinced LLMs are the direct path towards AGI, only pseudo-AGI. I think they're the central nervous system of it, but there's still something else that we need to get to the promised land. We might be hearing more about what that something is in the next few months.

[jamestiago]

Did you say that with some work in mind?

[funkervogt]

Bucking popular consensus, I'm not 100% convinced LLMs are the direct path towards AGI, only pseudo-AGI.

I agree. Still, a "fake AGI" like GPT-5 or -6 will be enormously powerful. It will take years for us to find ways to employ it to its full potential across many thousands of tasks.

[wjfox]

[lechwall]

Bucking popular consensus, I'm not 100% convinced LLMs are the direct path towards AGI, only pseudo-AGI. I think they're the central nervous system of it, but there's still something else that we need to get to the promised land. We might be hearing more about what that something is in the next few months.

I agree I've been saying this for a long tine
Did you read the paper Microsoft released about an early version of GPT 4
https://arxiv.org/abs/2303.12712
If so well worth a read the main problem on the road to AGI appears to be developing planning and giving the AI a long term memory which its not clear LLM's can do with available resources

[weatheriscool]

Geordie Rose – AI Humanoid Robot is the Biggest Business in the World
March 26, 2023 by Brian Wang

https://www.nextbigfuture.com/2023/03/g ... world.html

Geordi Rose is the CEO of Sanctuary AI, which has about $100 million in funding. They have deployed humanoid robots into actual commercial retail stores under a test program. During a week-long pilot test, the store, owned by retail chain Canada Tire Corporation (CTC), saw its mechanical intern handle 110 different retail-related activities in the front and back of the store. These included picking and packing merchandise, sales floor replenishment, cleaning, tagging, labeling, store display compliance, and folding – tasks that previously had been demonstrated only in a Sanctuary AI lab set up to mirror the store.

[superfishy]

The RLHF methodology of OpenAI doesn't seem like a very elegant path to AGI. Somewhat akin to how fusion power will ultimately come from making steam to spin a turbine lol. Either way I'm excited and more than a little scared about the future of AI and how fast its seeming to come upon us.

I have a degree in history and wrote a paper on the microhistory of average citizens during the industrial revolution and how oblivious they were in the moment to the vast changes approaching society as a result. We are in one of those revolutions now and its quite exciting, though I'd be surprised if society 20-30 years from now resembles anything to the current paradigm. Technological change is far outpacing the ability for governments and economies to adapt and for the first time I'm questioning if we will be able to handle it. I'm generally not a pessimistic person but I'd say the chance of a societal collapse in the next couple of decades isn't as unlikely as I'd like it to be. As a result of this, I've officially stopped contributing to my 401k . Societal collapse or no, I'm convinced by the time I'm of retirement age things will be a lot different..

[weatheriscool]

The RLHF methodology of OpenAI doesn't seem like a very elegant path to AGI. Somewhat akin to how fusion power will ultimately come from making steam to spin a turbine lol. Either way I'm excited and more than a little scared about the future of AI and how fast its seeming to come upon us.

I have a degree in history and wrote a paper on the microhistory of average citizens during the industrial revolution and how oblivious they were in the moment to the vast changes approaching society as a result. We are in one of those revolutions now and its quite exciting, though I'd be surprised if society 20-30 years from now resembles anything to the current paradigm. Technological change is far outpacing the ability for governments and economies to adapt and for the first time I'm questioning if we will be able to handle it. I'm generally not a pessimistic person but I'd say the chance of a societal collapse in the next couple of decades isn't as unlikely as I'd like it to be. As a result of this, I've officially stopped contributing to my 401k . Societal collapse or no, I'm convinced by the time I'm of retirement age things will be a lot different..

I honestly believe that any collapse will come from idiocy and the hatred of knowledge. It will come from the rights utter hatred of knowledge and advancement and the acceptance of such. People that are attacking chat-gpt are probably going to be part of the reasons that these people get elected that are spreading nothing but hatred throughout our society.

[ººº]

It's not just the right.