Good morning. It’s Monday, October 14th.

Did you know: On this day in 2011, the iPhone 4S was released in retail stores throughout the United States.

In today’s email:

• AI researchers question OpenAI's claims

• Experimental 'Swarm' framework

• Google's market share could drop below 50%

• Should AI weapons be allowed to decide to kill?

• 3 New AI Tools

• Latest AI Research Papers

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

^{In partnership with Butterflies AI}

The first social network where bots are cool

Butterflies AI is the hottest new social network where both Humans and AIs can coexist.

Join a space where humans and AI characters interact naturally—posting, commenting, and reacting to each other. On Butterflies, you have the freedom to create a new type of friend group and shape your own unique digital experience.

Free on iOS and Android—download Butterflies AI today. (Plus, you can even turn your selfies into AI characters that look like you with the new “Clones” feature - only available on the app.)

Apple AI researchers question OpenAI's claims about o1's reasoning capabilities

Apple researchers, including Samy Bengio and led by Mehrdad Farajtabar, have developed GSM-Symbolic and GSM-NoOp to assess the reasoning capabilities of large language models (LLMs) like OpenAI’s GPT-4o and o1. Building on the GSM8K dataset, these tools introduce symbolic templates and irrelevant information to more rigorously test models.

The study found that while models perform well on standard benchmarks, their reasoning weakens when confronted with slight variations, such as irrelevant details. Even leading models, including OpenAI's, appear to rely on pattern recognition rather than true logical reasoning.

The researchers argue that scaling models won’t resolve this issue and call for further research into real reasoning, challenging OpenAI’s claims regarding models like o1. Read more.

OpenAI unveils experimental 'Swarm' framework, igniting debate on AI-driven automation

OpenAI has rolled out "Swarm," an experimental framework designed to orchestrate networks of AI agents on Github, igniting a buzz in the AI community. Though not an official product, Swarm lays out a blueprint for developers to build networks of AI agents that collaborate autonomously, turning multi-agent systems from theory into something more accessible.

While Swarm isn't headed for production anytime soon, its potential business use cases—think automated market analysis or customer service—are hard to ignore. But alongside the excitement come concerns. Security experts warn that unleashing autonomous agents without robust safeguards could be risky, while ethicists worry about bias creeping in unnoticed. And then there's the looming question of job displacement—automation’s favorite elephant in the room.

Still, Swarm offers a forward-looking take on AI collaboration, pushing developers and enterprises to think ahead, even if it's not quite ready yet. Read more.

Google's share of the search ad market could drop below 50% for the first time in a decade as AI search engines boom

eMarketer projects that Google’s share of the U.S. search advertising market could dip below 50% for the first time in over ten years, driven by rising competition from AI platforms. Tools like ChatGPT and Perplexity AI are influencing user behavior, especially among younger generations, who are increasingly avoiding the term "Google" as a verb.

Perplexity AI reported 340 million queries in September and is attracting prominent advertisers, challenging Google's established market position. In response, Google introduced its Gemini large language model and various generative AI features to improve search results. As the competition intensifies, the online advertising terrain appears poised for a significant evolution, with traditional giants like Google facing new, nimble contenders redefining user engagement. Read more.

Silicon Valley is debating if AI weapons should be allowed to decide to kill

Silicon Valley finds itself at a crossroads, debating the implications of autonomous weapons. Shield AI co-founder Brandon Tseng confidently asserts that Congress will never permit AI to decide who lives or dies.

Yet, mere days later, Anduril co-founder Palmer Luckey tossed a wrench into this certainty, expressing a willingness to entertain the idea of weaponry with a mind of its own, albeit with a nuanced critique of traditional ethics. He questioned the moral superiority of a landmine that indiscriminately targets civilians over a more discerning robot.

The U.S. military remains noncommittal, allowing the development of autonomous systems while sidestepping any outright ban. With Ukraine pushing for automation to outmaneuver Russia, the urgency mounts for policymakers to clarify the murky waters of lethal AI, especially as defense firms eagerly lobby Congress for influence over the agenda. Read more. 

SpaceX’s Starship rocket booster caught safely in spectacular display of engineering

Jensen Huang says he wants Nvidia to be a company with 100 million AI assistants

'OCR 2.0' model converts images of text, formulas, notes, and shapes into editable text

TikTok Lays Off Hundreds of Staff—to Replace Them With AI

AutoDAN-Turbo autonomously develops jailbreak strategies to bypass language model safeguards

AI21 CEO says transformers not right for AI agents due to error perpetuation 

Meta suggests AI Northern Lights pics are as good as the real thing

📄 Baichuan-Omni Technical Report  

📄 Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image Synthesis ​​

📄 MLLM as Retriever: Interactively Learning Multimodal Retrieval for Embodied Agents  

📄 WALL-E: World Alignment by Rule Learning Improves World Model-based LLM Agents

📄 MathCoder2: Better Math Reasoning from Continued Pretraining on Model-translated Mathematical Code

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!