The Path Forward for AI

There’s been a lot of interesting AI news lately, and I wanted to share some thoughts on where things might be heading. Open source models are quickly catching up to—and in some cases even outperforming—closed source alternatives. Admittedly, these are just projections and might not capture the complete picture, but here’s what seems to be converging towards:

1. Increased Compute for Local Use
   As everyday devices pack more computing power, it’s now feasible to run complex AI models locally rather than relying solely on cloud services. This setup reduces latency and can lower costs, making powerful AI accessible right from your computer or phone.
2. Better Performance for Robotics and Real-Time Applications
   In areas like robotics, every millisecond counts. When quick responses are essential, running AI locally—right at the source—avoids delays from data traveling over the internet. This approach can be a game-changer, ensuring that robotic systems can adapt to changes in real-time.
3. Enhanced Data Privacy and Security
   Local AI models keep data closer to home. Processing sensitive information on your device or within your company’s network minimizes risks associated with data transmission. This setup offers greater control over your data, which is especially important in today’s privacy-focused environment.

While there’s still a lot to work on in AI, this shift toward open source and local processing looks promising. It paves the way for more immediate, secure, and adaptable AI applications that can benefit a wide range of users and industries.

Meanwhile, I’m having a great time running large models on my M4 Max MacBook Pro. Admittedly, it’s still a bit slow—chugging along at 4-5 tokens per second for Deepseek R1—so it’s not quite ready for production use yet.

---

AI News for the Week (from Gemini)

Key Technological Advancements:

• Robotics: Agibot developing humanoid robots for tasks like steaming clothes and making sandwiches, aiming to assist aging populations.
• Scientific Discovery: AI being used to virtually “unscroll” fragile ancient scrolls buried by Mount Vesuvius.
• Image Generation: MIT researchers developed a tool that generates high-quality images faster and with lower energy consumption.
• Food Industry: AI revolutionizing R&D, accelerating product development, and improving supply chain management through predictive analytics and LLMs.
• Quantum Computing: MIT researchers made progress in enabling communication between multiple quantum processors.

Ethical and Societal Implications of AI:

• Viral spread of sexually explicit deepfake videos of Taylor Swift on X sparked outrage and calls for regulation.
• Research highlights public concerns about transparency, misinformation, bias, and the impact on journalism due to AI.
• EU continues scrutiny of OpenAI’s ChatGPT and is working on an AI Liability Directive.
• Major tech companies in the US are seeking federal AI regulation.
• Professionals are divided on the ethical acceptability of AI providing professional advice.

Funding and Investment Trends in the AI Sector:

• Tencent plans to increase AI spending.
• Nvidia CEO unveiled new AI chips.
• Significant funding rounds for AI startups: Anthropic ($3.5B), Turing ($111M), Hippocratic AI ($141M), Figure (potential $1.5B), Lambda ($480M), Together.ai ($305M), and others.
• Nvidia and xAI joined a $30 billion AI Infrastructure Fund.
• OpenAI announced a $12 billion investment in CoreWeave.

Strategic Partnerships and Collaborations:

• Nvidia partnered with General Motors (GM) in humanoid robotics.
• OpenAI and Meta Platforms are in discussions with India’s Reliance Industries for potential partnerships.
• The AI Infrastructure Partnership collaborated with energy sector leaders GE Vernova and NextEra Energy.
• Collaborations between agricultural companies, academic institutions, and startups are crucial for AI in the food industry.

New AI Products and Services Launched:

• Bloomberg Law introduced new AI-powered legal intelligence solutions.
• Cloudflare launched new cybersecurity products, including Cloudflare for AI.
• The New York Times is using internal AI tools to assist journalists with tasks like summarization and question generation, while adhering to ethical guidelines.