Good morning, it’s Tuesday! In this week's edition, we've got AMD flexing its AI muscles with a mega-acquisition, brain cells for rent (yes, you read that right), and AI helping entrepreneurs go from zero to hero. Plus, California's AI safety bill gets a makeover, and we explore why most companies are still playing it cool on AI.
—
Help us improve Forward Future: If you haven’t already, submit your answer to this week’s Forward Future poll. It’s at the bottom! 👇️
Get the best of multiple LLMs in one place! ChatHub enables you to compare and combine the responses from six different AI bots simultaneously, enhancing your use of LLMs and transforming your interactions with AI. Try ChatHub now.
👉️ Top Stories
AMD is going after Nvidia with a $5 billion acquisition
Advanced Micro Devices (AMD) is set to acquire cloud computing solutions provider ZT Systems for $4.9 billion, aiming to enhance its AI chip competitiveness against firms like Nvidia. This strategic move is expected to accelerate the deployment of AI infrastructure for cloud and enterprise customers and follows AMD's significant investments in expanding its AI software capabilities over the past year, with the deal potentially being profit-generating by the end of next year. Continue reading here.
Swiss Tech Firm Launches AI Made of Human Brain Cells Rental Service
Swiss company FinalSpark has introduced a cloud rental service for “biocomputers” composed of human brain cells, offering access for $500 per month. These organoid-based systems aim to provide a more energy-efficient alternative to traditional silicon computing, potentially being 100,000 times more efficient for AI model training. The technology uses dopamine and electrical signals for reinforcement learning, mimicking the way human neurons respond. However, the organoids have a limited lifespan of around 100 days, after which the AI effectively "dies," raising ethical and practical implications for future development. Continue reading here.
How A.I. Can Help Start Small Businesses
Entrepreneurs are increasingly using generative A.I. tools like ChatGPT and GitHub Copilot to accelerate business development, from coding and marketing to legal work and customer acquisition. In educational settings like Carnegie Mellon University, A.I. has notably sped up students’ startup progress. Startups like Skittenz and Orb.net have relied heavily on A.I. to overcome knowledge gaps and reduce initial costs, enabling them to reach milestones faster. Although still emerging, A.I.'s ability to amplify efficiency and reduce overhead suggests that it could reshape the startup landscape by driving quicker paths to profitability and scaling. Continue reading here.
California weakens bill to prevent AI disasters before final vote, taking advice from Anthropic
California’s SB 1047, a bill designed to hold AI developers accountable for potential catastrophic events, has been weakened following amendments influenced by AI firm Anthropic. Significant changes include removing the authority for the state attorney general to sue AI companies preemptively and replacing criminal liability for developers with more lenient public safety statements. The bill also eliminates the creation of a dedicated Frontier Model Division, replacing it with a larger advisory board under the existing Government Operations Agency. Despite these adjustments, the bill still faces strong opposition from critics concerned about stifling innovation, as it heads to a final vote in the Assembly. Continue reading here.
Most companies aren’t talking about AI
Despite the notable buzz surrounding AI, only 44% of S&P 500 companies mentioned "AI" or "artificial intelligence" in their recent earnings calls, indicating a gradual yet incomplete adoption among major corporations. While some industry leaders recognize AI's potential for efficiency and consumer service improvements, others foresee challenges, and a few have already integrated AI into their operations, like John Deere with autonomous tractors, without necessarily highlighting it in discussions. Continue reading here.
✍️ Editor Picks
Research
HybridRAG: Integrating Knowledge Graphs and Vector Retrieval Augmented Generation for Efficient Information Extraction
The paper presents an in-depth analysis of a novel approach called HybridRAG, which combines knowledge graphs and vector-based techniques to enhance large language models' (LLMs) ability to extract and interpret complex information from financial texts, such as earnings call transcripts. HybridRAG outperforms traditional RAG methods (VectorRAG and GraphRAG) in retrieval accuracy and answer generation when tested on a dataset of financial documents. The approach is instrumental for financial analysts, given the unstructured and complex nature of financial data, and the study indicates potential broader applications beyond finance. Continue reading.
Robotics
Apple Pushes Ahead With Tabletop Robot in Search of New Revenue
Apple is advancing the development of a high-end tabletop robot with a screen attached to a robotic arm, aiming to position it as a smart home command center and video conferencing tool. Led by executive Kevin Lynch, the device marks a shift into robotics and forms part of Apple's strategy to boost revenue as iPhone sales slow. Slated for potential release between 2026 and 2027, the project has encountered internal debate regarding its consumer appeal and technical demands. This move aligns with Apple's broader efforts to explore AI-driven products and new market opportunities.Continue reading.
Models
Introducing Hermes 3
The Hermes 3 AI model by Nous Research, built on Llama 3.1, introduces groundbreaking advancements in open-source AI with features like long-term context retention, enhanced conversation abilities, and complex roleplaying. Hermes 3 is fine-tuned on a dataset largely composed of synthetic responses and showcases superior reasoning and creativity compared to its predecessor. This decentralized project emphasizes experimentation in ways that prioritize adaptive instruction-following and innovative AI applications. Continue reading.
🚀 AI Launches: Ignition Sequence Start!
Mystic by Freepik - Freepik launches a beta test for its text to image generator
Runway Gen-3 Turbo - Runway has enhanced its Gen-3 artificial intelligence video platform with a "turbo" option, dramatically speeding up video creation from images and text prompts.
Pindrop Pulse - Pindrop says it can detect deepfake audio with 99% accuracy
Robotic Tutorials by Hugging Face - Learn how to train an LLM for a homemade robot
📽️ New Videos
🧰 Toolbox: New AI Use Cases
New Geekbench AI benchmark can test the performance of CPUs, GPUs, and NPUs - Primate Labs, known for Geekbench, is launching Geekbench AI version 1.0, an app designed to test the performance of Neural Processing Units (NPUs) by simulating a range of artificial intelligence workloads and precision levels across various platforms.
Midjourney’s New Web Editor - Offering a streamlined interface that unifies various previously separate image actions into one cohesive platform.
How to Prune and Distill Llama-3.1 8B to an NVIDIA Llama-3.1-Minitron 4B Model - NVIDIA researchers have developed a method to create efficient small language models (SLMs) by pruning and distilling a larger model, achieving higher accuracy with significantly less computational cost and fewer training tokens. This approach, validated with models like Llama-3.1-Minitron 4B, outperforms similarly sized models across various benchmarks and is set to join the NVIDIA HuggingFace collection.
🛰️ Houston, we have more headlines!
Donald Trump shares AI-generated images - The images show Taylor Swift fans, known as "Swifties," supporting his campaign. However, the images were later identified as deepfakes, with one pro-Trump account admitting the post was satire.
AI is Not our Future - Procreate is taking a strong stance against generative AI, emphasizing respect for artists’ skills and ownership of their work.
Automated Design of Agentic Systems - The new research seeks to automatically generate advanced agentic systems capable of incorporating new or recombined components, where a meta agent can theoretically program unlimited varieties of agents due to the Turing Completeness of programming languages.
LVMH CEO Bernard Arnault's family office goes shopping for AI startups - In 2022, Bernard Arnault, CEO of LVMH and the world's fourth-richest person, engaged in a series of artificial intelligence (AI) investments through his family office, Aglaé Ventures, partaking in over $300 million worth of funding rounds in AI companies.
Sam Altman’s Worldcoin Is Battling With Governments Over Your Eyes - Worldcoin, the brainchild of Sam Altman, aims to scan the irises of every person on Earth to create unique digital IDs and distribute cryptocurrency, but it faces widespread regulatory pushback.
Google’s AI Search Gives Sites Dire Choice: Share Data or Die - Google’s AI-powered search results are sparking a crisis for publishers, who must choose between allowing Google to use their content or risking disappearance from search listings altogether.
🏡 House Keeping
Add forwardfuture@mail.beehiiv.com to your contacts.
And add us to your Primary inbox OR Reply to this email with “spaceman”
Share Forward Future with your AI buddies
