Customizing the R1 with Generative UI

Though this seems like a fun way to customize the R1, Rabbit says responses on the AI-generated interfaces are “slower than the default UI on R1 and can take over 30 seconds to appear.” It suggests disabling this feature if you want the fastest responses possible.

Enabling Generative UI

You can turn on generative UI by logging into your rabbithole account, selecting settings, and then clicking “profile.” Once you hit “enable generative UI,” you can enter a prompt, such as: “You are a talented front-end UI designer and your favorite season is autumn. You will generate a very high quality UI in this style and it will look amazing.” Then, hit “customize.”

Limitations of Generative UI

Following the R1’s major update, some reviewers have revisited the device. CNET says the experience has “vastly improved,” but that you’re still much better off with a smartphone. Android Authority was less impressed even after the latest patch.

Conclusion

In conclusion, while generative UI may be a fun feature to play with, it may not be the best option for those looking for fast and efficient responses. The slower response times and limitations of the feature may make it less appealing to some users. However, for those who are interested in exploring the possibilities of AI-generated interfaces, it may be worth giving it a try.

FAQs

Q: What is generative UI?
A: Generative UI is an AI-generated interface feature that allows users to customize their R1 device with unique and personalized designs.

Q: How do I enable generative UI?
A: To enable generative UI, log into your rabbithole account, select settings, and then click “profile.” Hit “enable generative UI” and enter a prompt to customize your interface.

Q: Are there any limitations to generative UI?
A: Yes, responses on the AI-generated interfaces can take over 30 seconds to appear, making it slower than the default UI on R1. Additionally, some reviewers have reported that the feature is not as impressive as expected, even after the latest patch.

Q: Is generative UI worth trying?
A: If you’re interested in exploring the possibilities of AI-generated interfaces and don’t mind slower response times, then yes, generative UI may be worth trying. However, if you prioritize fast and efficient responses, you may want to consider other options.

Elon Musk Updates Lawsuit Against OpenAI, Accuses Microsoft of Antitrust Violations

Elon Musk has updated his fraud, breach of contract, and racketeering lawsuit against OpenAI to make antitrust claims against Microsoft, accusing the two companies of attempting to “monopolize the generative AI market.” The amended complaint filed on Thursday names Microsoft as a new defendant, as well as LinkedIn co-founder Reid Hoffman and Microsoft VP Dee Templeton, a former OpenAI board member.

New Defendants Added to the Complaint

Microsoft has invested $14 billion in OpenAI since 2019 and holds exclusive rights to commercially license the AI startup’s technology and a 49 percent stake in its for-profit subsidiary. Musk’s lawyers say that former OpenAI CEO Sam Altman “engaged in rampant self-dealing” to create a “de facto merger” between the two companies to promote anticompetitive practices.

Antitrust Allegations

The lawsuit argues that OpenAI and Microsoft are “actively trying to eliminate competitors” by exchanging “competitively sensitive information,” and making their investors refrain from funding rival companies like xAI, Musk’s AI company. Both companies are competing for funding in the growing AI market, with OpenAI securing $6.6 billion in October to build more powerful AI models. xAI raised $6 billion in its own funding round in March to accelerate its development of “future technologies.”

New Plaintiff Added to the Complaint

xAI has also been added to the complaint as a new plaintiff, alongside former OpenAI board member Shivon Zilis — an executive at Musk’s Neuralink company and mother of three of his 12 children.

Background on the Lawsuit

Musk co-founded OpenAI in 2015 with Altman and Greg Brockman, who are also defendants in the complaint, before leaving the company in 2018. A previous lawsuit that Musk launched in March (and dropped without explanation in June) accused OpenAI of abandoning its founding agreement to develop AI that will benefit humanity, and claimed its partnership with Microsoft made OpenAI a “closed-source de facto subsidiary” focused on maximizing profits.

Conclusion

The amended complaint aims to stop OpenAI and Microsoft from allegedly engaging in anticompetitive practices that harm the development of AI technology. The lawsuit will likely have significant implications for the future of AI and the tech industry.

FAQs

Q: What is the main accusation in the lawsuit?

A: The lawsuit accuses OpenAI and Microsoft of attempting to monopolize the generative AI market and engaging in anticompetitive practices.

Q: Who are the new defendants in the lawsuit?

A: The new defendants are Microsoft, LinkedIn co-founder Reid Hoffman, and Microsoft VP Dee Templeton, a former OpenAI board member.

To teach an AI agent a new task, like how to open a kitchen cabinet, researchers often use reinforcement learning — a trial-and-error process where the agent is rewarded for taking actions that get it closer to the goal.

In many instances, a human expert must carefully design a reward function, which is an incentive mechanism that gives the agent motivation to explore. The human expert must iteratively update that reward function as the agent explores and tries different actions. This can be time-consuming, inefficient, and difficult to scale up, especially when the task is complex and involves many steps.

Researchers from MIT, Harvard University, and the University of Washington have developed a new reinforcement learning approach that doesn’t rely on an expertly designed reward function. Instead, it leverages crowdsourced feedback, gathered from many nonexpert users, to guide the agent as it learns to reach its goal.

While some other methods also attempt to utilize nonexpert feedback, this new approach enables the AI agent to learn more quickly, despite the fact that data crowdsourced from users are often full of errors. These noisy data might cause other methods to fail.

In addition, this new approach allows feedback to be gathered asynchronously, so nonexpert users around the world can contribute to teaching the agent.

“One of the most time-consuming and challenging parts in designing a robotic agent today is engineering the reward function. Today reward functions are designed by expert researchers — a paradigm that is not scalable if we want to teach our robots many different tasks. Our work proposes a way to scale robot learning by crowdsourcing the design of reward function and by making it possible for nonexperts to provide useful feedback,” says Pulkit Agrawal, an assistant professor in the MIT Department of Electrical Engineering and Computer Science (EECS) who leads the Improbable AI Lab in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL).

In the future, this method could help a robot learn to perform specific tasks in a user’s home quickly, without the owner needing to show the robot physical examples of each task. The robot could explore on its own, with crowdsourced nonexpert feedback guiding its exploration.

“In our method, the reward function guides the agent to what it should explore, instead of telling it exactly what it should do to complete the task. So, even if the human supervision is somewhat inaccurate and noisy, the agent is still able to explore, which helps it learn much better,” explains lead author Marcel Torne ’23, a research assistant in the Improbable AI Lab.

Torne is joined on the paper by his MIT advisor, Agrawal; senior author Abhishek Gupta, assistant professor at the University of Washington; as well as others at the University of Washington and MIT. The research will be presented at the Conference on Neural Information Processing Systems next month.

Noisy feedback

One way to gather user feedback for reinforcement learning is to show a user two photos of states achieved by the agent, and then ask that user which state is closer to a goal. For instance, perhaps a robot’s goal is to open a kitchen cabinet. One image might show that the robot opened the cabinet, while the second might show that it opened the microwave. A user would pick the photo of the “better” state.

Some previous approaches try to use this crowdsourced, binary feedback to optimize a reward function that the agent would use to learn the task. However, because nonexperts are likely to make mistakes, the reward function can become very noisy, so the agent might get stuck and never reach its goal.

“Basically, the agent would take the reward function too seriously. It would try to match the reward function perfectly. So, instead of directly optimizing over the reward function, we just use it to tell the robot which areas it should be exploring,” Torne says.

He and his collaborators decoupled the process into two separate parts, each directed by its own algorithm. They call their new reinforcement learning method HuGE (Human Guided Exploration).

On one side, a goal selector algorithm is continuously updated with crowdsourced human feedback. The feedback is not used as a reward function, but rather to guide the agent’s exploration. In a sense, the nonexpert users drop breadcrumbs that incrementally lead the agent toward its goal.

On the other side, the agent explores on its own, in a self-supervised manner guided by the goal selector. It collects images or videos of actions that it tries, which are then sent to humans and used to update the goal selector.

This narrows down the area for the agent to explore, leading it to more promising areas that are closer to its goal. But if there is no feedback, or if feedback takes a while to arrive, the agent will keep learning on its own, albeit in a slower manner. This enables feedback to be gathered infrequently and asynchronously.

“The exploration loop can keep going autonomously, because it is just going to explore and learn new things. And then when you get some better signal, it is going to explore in more concrete ways. You can just keep them turning at their own pace,” adds Torne.

And because the feedback is just gently guiding the agent’s behavior, it will eventually learn to complete the task even if users provide incorrect answers.

Faster learning

The researchers tested this method on a number of simulated and real-world tasks. In simulation, they used HuGE to effectively learn tasks with long sequences of actions, such as stacking blocks in a particular order or navigating a large maze.

In real-world tests, they utilized HuGE to train robotic arms to draw the letter “U” and pick and place objects. For these tests, they crowdsourced data from 109 nonexpert users in 13 different countries spanning three continents.

In real-world and simulated experiments, HuGE helped agents learn to achieve the goal faster than other methods.

The researchers also found that data crowdsourced from nonexperts yielded better performance than synthetic data, which were produced and labeled by the researchers. For nonexpert users, labeling 30 images or videos took fewer than two minutes.

“This makes it very promising in terms of being able to scale up this method,” Torne adds.

In a related paper, which the researchers presented at the recent Conference on Robot Learning, they enhanced HuGE so an AI agent can learn to perform the task, and then autonomously reset the environment to continue learning. For instance, if the agent learns to open a cabinet, the method also guides the agent to close the cabinet.

“Now we can have it learn completely autonomously without needing human resets,” he says.

The researchers also emphasize that, in this and other learning approaches, it is critical to ensure that AI agents are aligned with human values.

In the future, they want to continue refining HuGE so the agent can learn from other forms of communication, such as natural language and physical interactions with the robot. They are also interested in applying this method to teach multiple agents at once.

This research is funded, in part, by the MIT-IBM Watson AI Lab.

EU Releases Draft General-Purpose AI Code of Practice for Regulatory Guidance

The development of the "First Draft General-Purpose AI Code of Practice" marks a significant effort by the EU to create comprehensive regulatory guidance for general-purpose AI models. This draft has been a collaborative effort, involving input from diverse sectors, including industry, academia, and civil society.

Key Objectives

The draft aims to clarify compliance methods for providers of general-purpose AI models, facilitate understanding across the AI value chain, ensure compliance with Union law on copyrights, and continuously assess and mitigate systemic risks associated with AI models.

Recognizing and Mitigating Systemic Risks

The draft emphasizes the need for robust safety and security frameworks (SSFs) and proposes a hierarchy of measures, sub-measures, and key performance indicators (KPIs) to ensure appropriate risk identification, analysis, and mitigation throughout a model’s lifecycle.

Taxonomy of Systemic Risks

The draft outlines various threats, including cyber offenses, biological risks, loss of control over autonomous AI models, and large-scale disinformation. The document acknowledges that this taxonomy will need updates to remain relevant as AI technology continues to evolve.

Proactive Stance on AI Regulatory Guidance

The EU AI Act mandates that the final version of this Code be ready by 1 May 2025. This initiative underscores the EU’s proactive stance towards AI regulation, emphasizing the need for AI safety, transparency, and accountability.

Conclusion

While still in draft form, the EU’s Code of Practice for general-purpose AI models could set a benchmark for responsible AI development and deployment globally. By addressing key issues such as transparency, risk management, and copyright compliance, the Code aims to create a regulatory environment that fosters innovation, upholds fundamental rights, and ensures a high level of consumer protection.

Frequently Asked Questions

Q: What is the purpose of the First Draft General-Purpose AI Code of Practice?
A: The purpose of the draft is to create comprehensive regulatory guidance for general-purpose AI models, ensuring their safe and responsible development and deployment.

Q: What are the key objectives outlined in the draft?
A: The draft aims to clarify compliance methods for providers of general-purpose AI models, facilitate understanding across the AI value chain, ensure compliance with Union law on copyrights, and continuously assess and mitigate systemic risks associated with AI models.

Q: What is the EU AI Act, and how does it relate to the Code of Practice?
A: The EU AI Act mandates that the final version of this Code be ready by 1 May 2025, emphasizing the EU’s proactive stance towards AI regulation, emphasizing the need for AI safety, transparency, and accountability.