Artificial Intelligence thread

[Wrought]

Paper studying the development of smart cities w.r.t. business regulations in 82 cities.

As globalization advances, establishing a fair, efficient, and convenient business environment is vital to enhancing international competitiveness. Smart city development, as a new urban innovation model, drives digital transformation. It also improves market conditions, and speeds up urban reforms. Therefore, it is particularly important to examine how less developed regions can utilize smart city policy (SCP) to improve their business environment. This study is grounded in institutional theory, information economics, and the theory of competitive advantage.

Using panel data from 82 cities in western China spanning 2006 to 2021, it treats SCP as a quasi-natural experiment and applies a multi-period difference-in-differences model. The research examines the link between smart city development and regional business environments, while also exploring the underlying mechanisms. The findings reveal that SCP have significantly improved the business environment in western China, which remains robust even after a series of robustness tests. The effects of SCP are particularly pronounced in megacities and in cities with higher levels of human capital, stronger financial support, and more advanced digital infrastructure. Moreover, SCP fosters a market-oriented, internationalized, and law-governed business environment, while Fintech further amplifies this positive impact. Additionally, SCP exhibits a significant spillover effect. These findings offer valuable insights for policymakers seeking to enhance the business environment in less developed regions.

Please, Log in or Register to view URLs content!

 

[tphuang]

https://twitter.com/i/web/status/1988366113395351798

JD here claims its JoyAI platform was handling all the customer service and streaming programming with AI and that achieved huge improvement in satisfaction and cost savings. I'm really curious on how well it really worked, because companies tend to be overly positive when they report on things like this.

 

[Michael90]

https://twitter.com/i/web/status/1988366113395351798

JD here claims its JoyAI platform was handling all the customer service and streaming programming with AI and that achieved huge improvement in satisfaction and cost savings. I'm really curious on how well it really worked, because companies tend to be overly positive when they report on things like this.

Is JD only present in China? Or are they more global like Alobaba and Pingdodo's Temu?
Never heard of them abroad actually but they seem to be huge in China and they function more like Amazon..

 

[mossen]

Great interview with some of the guys behind InclusionAI's Ling model.

 

[Wrought]

AI-driven misinformation is on the rise, and the authorities have taken notice of it. Definitely something which warrants a decisive response.

GUANGZHOU -- China is moving to tighten the regulation of malicious images and videos created using generative artificial intelligence in response to an increase in fake content posted on social media, including those purporting to show kidnapping victims or earthquake damage. Revisions to a law on internet safety take effect in January. This law initially was intended to improve the security of telecommunications infrastructure, but the latest additions include risk management and safety monitoring of AI. The revisions were approved by the Standing Committee of the National People's Congress on Oct. 28.

Following a magnitude-6.8 earthquake that struck Tibet on Jan. 7, an image of a newborn covered in debris spread via social media. The image went viral as people sympathized with the baby, but it was later discovered that the image was created with generative AI and had nothing to do with the earthquake. The person who posted the image was detained by authorities for spreading misinformation.

Public security authorities also are intensifying their crackdown. In late October, these authorities publicized some of the illegal fake content they found, signaling that they will pursue those who take similar actions. The examples included an incident in which a 28-year-old Zhejiang province man claimed his daughter was kidnapped by using a photo of a girl that he found online. In another case, a 57-year-old Shanxi province woman posted text and images purportedly showing damage from an earthquake, but such damage did not exist.

Please, Log in or Register to view URLs content!

 

[tphuang]

https://twitter.com/i/web/status/1988958033196818886

more effort here from Alibaba to boost its stock prices, lol

I mean of course they are working on app improvement. They have the best open source model family. They should be doing so.

 

[GulfLander]

Anthropic making antiCN allegations?

https://twitter.com/i/web/status/1989033793190277618

 

[magmunta]

Anthropic making antiCN allegations?

https://twitter.com/i/web/status/1989033793190277618

The allegations could be true because that would definitely benefit china. What could also be true is to represent china and Chinese AI systems as existential threat so that the western citizens don't have access to the Chinese AI, which would, in turn, give western AI systems, such as Anthropic, greater market share and income. Both these could be true. By the way, is Anthropic a cyber security company?! How did they find first the Cyber operation? It sounds suspicious. I love the USA and would rather see western companies dominate AI, but this sounds like an attempt by Anthropic to get rid of Chinese competition.

 

[anzha]

This was an interesting article:

'Tiny' AI model beats massive LLMs at logic test

Please, Log in or Register to view URLs content!

Associated preprints from arxiv:

Less is More: Recursive Reasoning with Tiny Networks

Please, Log in or Register to view URLs content!

Hierarchical Reasoning Model

Please, Log in or Register to view URLs content!

Not claiming to be a replacement for LLMs, but interesting all the same.

 

[Overbom]

Anthropic making antiCN allegations?

https://twitter.com/i/web/status/1989033793190277618

Just another hit job. After the news were spread all around the world, they slightly changed a minor detail on their blog post about this supposedly big hack attempt


Original:

"The AI made thousands of requests per second — an attack speed that would have been, for human hackers, simply impossible to match."

Updated:

At the peak of its attack, the AI made thousands of requests, often multiple per second—an attack speed that would have been, for human hackers, simply impossible to match.

No biggie. Totally honest PR


Look they even made a small remark about the change. Well, nobody cares now as long as the news were spread lol

• Corrected an error about the speed of the attack: not "thousands of requests per second" but "thousands of requests, often multiple per second"