Good morning. It’s Friday, November 22nd.

Did you know: 10 AI agents are currently writing a fully autonomous book?

In today’s email:

• Simulation Agents

• LLM Siri

• ChatGPT’s Live Camera Code

• New Altman Interview

• 4 New AI Tools

• Latest AI Research Papers

You read. We listen. Let us know what you think by replying to this email.

Hire Ava, the AI SDR & Get Meetings on Autopilot

Ava automates your entire outbound demand generation process, including:

• Intent-Driven Lead Discovery

• High Quality Emails with Waterfall Personalization

• Follow-Up Management

Free up your sales team to focus on high-value interactions and closing deals, while Ava handles the time-consuming tasks.

Book a demo to see how Ava can 10x your outbound.

‘Simulation Agents’ Can Now Create a Replica of Your Personality

Stanford and Google DeepMind researchers have engineered "simulation agents," AI models capable of replicating human values and decision-making with 85% accuracy. By distilling insights from two-hour qualitative interviews, these agents encapsulate nuanced personality traits, sidestepping the logistical and ethical hurdles of live human studies. Rigorous testing via personality metrics and behavioral tasks revealed their capacity to emulate structured human responses, positioning them as a scalable tool for social and behavioural research.

While promising, this capability comes with caveats. Unlike task-driven AI systems built to execute actions, simulation agents venture into the realm of human-like reasoning. Their utility in replicating human dynamics could redefine research methodologies, though limitations in behavioral depth and emerging ethical dilemmas warrant a measured approach to deployment. Read more.

Apple Reportedly Working on ‘LLM Siri’ to Compete with ChatGPT

Apple is reportedly advancing Siri’s capabilities through "LLM Siri," a generative AI enhancement designed to rival OpenAI's ChatGPT and Google's Gemini Live. This overhaul will integrate Apple's proprietary AI models to enable more fluid, conversational interactions while enhancing Siri's ability to summarize text and perform complex tasks. Expanded App Intents functionality will deepen its interoperability with third-party applications.

Set to debut by spring 2026, the system marks a shift in Siri's underlying architecture. Unlike cloud-reliant systems, Apple emphasizes device-local processing, preserving speed and security for tasks such as messaging and app navigation. Recent updates, including ChatGPT integration in the iOS 18.2 beta, reflect Apple’s cautious but deliberate AI adoption. Read more.

ChatGPT’s ‘Live Camera’ video features were just found in the code of the latest beta

OpenAI’s latest GPT-4o update sharpens its creative prowess, making it more fluid, user-centric, and naturally engaging. The model now produces responses that are not only clearer but more aligned with user intent, enhancing overall readability. File processing also sees a significant boost, offering deeper insights and more thorough responses.

Alongside this, OpenAI continues to push the boundaries of interactive AI. The latest beta (v1.2024.317) uncovers features like live camera functionality, real-time processing, and enhanced visual recognition. These advancements, building on May’s real-time interactions with objects, highlight OpenAI’s growing capabilities in visual AI. Read more.

WATCH: Sam Altman has an idea to get AI to ‘love humanity,’ use it to poll billions of people about their value systems

Sam Altman has articulated a vision for AI to “love humanity,” a concept he believes could be engineered into AI systems. During an interview, Altman proposed a method where AI engages in deep conversations with individuals worldwide to gauge their values, using this information to align AI behavior with humanity’s collective interests. According to Altman, this is entirely within reach, given AI’s current architecture.

Yet, as the field progresses toward Artificial General Intelligence (AGI), the stakes grow. Critics like Geoffrey Hinton and Elon Musk continue to warn of AI’s existential risks. Meanwhile, OpenAI’s internal alignment team has splintered, with departures signaling friction over how much safety is prioritized in the race to AGI. Read more.

1. AI Research and Development

• OpenScholar: The open-source A.I. that’s outperforming GPT-4o in scientific research

• Alibaba Just Released Marco-o1: Advancing Open-Ended Reasoning in AI

• Alibaba's Qwen2.5 Turbo reads ten novels in just about one minute

• A Chinese lab has released a 'reasoning' AI model to rival OpenAI's o1

• MIT researchers develop an efficient way to train more reliable AI agents

• New GPT-4o Model Appears Smaller Than Previous Updates

2. AI Applications and Hardware

• Adobe announces development of SLM that can run locally on a phone with no cloud connection

• Samsung unveils AI smart glasses with Google and Qualcomm

• Nanorobots move closer to clinical trials with new model that helps them navigate through the bloodstream

• Robotic model with real pigeon feathers can fly like a bird

• Rabbit launches ‘teach mode’ allowing people to create custom agents for its AI mobile device

3. AI in Business and Market Trends

• Nvidia's CEO defends his moat as AI labs change how they improve their AI models

• Three mystery whales have each spent $10 billion-plus on Nvidia's AI chips so far this year

• Google dropping new models left and right - Gemini Experimental 1121

• ​​Google Cloud launches AI Agent Space amid rising competition

• OpenAI Considers Taking on Google With Browser — The Information

4. AI Ethics, Regulation, and Controversy

• OpenAI accidentally deleted potential evidence in NY Times copyright lawsuit

• Elon Musk asked people to upload their medical data to 𝕏 so his AI company could learn to interpret MRIs and CT scans

• The CEO of Anthropic blasts VC Marc Andreessen's argument that AI shouldn't be regulated because it's 'just math'

• Elon Musk's Neuralink receives Canadian approval for brain chip trial

5. AI in Emerging Technologies

• AlphaQubit tackles one of quantum computing’s biggest challenges

• aiOla unveils open source AI audio transcription model that obscures sensitive info in real time

• YouTube Shorts' Dream Screen feature can now generate AI video backgrounds

📄 SymDPO: Boosting In-Context Learning of Large Multimodal Models with Symbol Demonstration Direct Preference Optimization

📄 RedPajama: an Open Dataset for Training Large Language Models ​​

📄 SageAttention2 Technical Report: Accurate 4 Bit Attention for Plug-and-play Inference Acceleration

📄 Ultra-Sparse Memory Network  

📄 VideoAutoArena: An Automated Arena for Evaluating Large Multimodal Models in Video Analysis through User Simulation

Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X!