How Trump’s admin could change US AI Policy

Today’s Highlights:

📰 News: How Trump’s admin could change US AI Policy + Anthropic models can now be used by US defense and intelligence agencies

💰 Funding: Perplexity is set to raise $500M

⚡️ Top News Stories:

1. Trump’s victory could lead to a significant change in the US’s AI policy:

• Trump’s Expected Repeal of Biden’s AI Policy: With his election victory, Trump has pledged to dismantle Biden’s AI Executive Order (EO), which required reporting on AI security and real-world safety impacts, including model vulnerabilities and bias checks.

• Biden’s AI Executive Order: Passed in October 2023, Biden’s EO created the U.S. AI Safety Institute (AISI) and set guidelines for companies to report model testing results to ensure AI safety, which some Republicans criticized as overreach.

• Potential Repeal of CHIPS and Science Act: House Speaker Mike Johnson, a Trump supporter, suggested that Republicans may repeal or “streamline” the CHIPS Act, a Biden initiative to boost U.S. semiconductor manufacturing crucial for AI tech.

• Trump’s Likely Light-Touch Regulation: Trump’s AI stance emphasizes free speech and limited regulation. Analysts expect his administration to focus on safety risks relevant to national defense but avoid broad regulatory measures, potentially dismantling the AISI.

• Increased Role of State Legislation: Trump’s approach may prompt states, particularly Democratic ones like California, to enact their own AI regulations. This year alone saw nearly 700 AI-related bills introduced at the state level.

• Geopolitical and Economic Implications: Analysts predict Trump will tighten AI-related export controls on China, possibly impose tariffs on tech imports, and restrict H-1B visas—all of which could impact AI R&D funding and U.S. leadership in AI.

• Influence of Elon Musk: Musk, a Trump supporter with major business interests in AI, could gain regulatory influence under the new administration, potentially benefiting ventures like xAI, Tesla, and SpaceX.

2. At the Conference for Robot Learning, NVIDIA introduced new advanced tools for robot learning.

• Project GR00T is a set of specialized workflows to enhance humanoid robot capabilities in areas such as motion generation and environmental perception.

• The open-source Isaac Lab framework enables developers to train robots across various forms using scalable, physics-based simulations.

• The new Cosmos tokenizer enables high-speed visual data processing, enhancing robotic vision tasks by processing data 12x faster than existing methods.

• NVIDIA partnered with Hugging Face to integrate LeRobot with its robotics tools, aiming to accelerate open-source AI robotics research and innovation.

3. Microsoft Research has introduced Magentic-One, an open-source, generalist multi-agent system designed to tackle complex tasks by integrating multiple specialized agents.

• Magentic-One employs an "Orchestrator Agent" to oversee and coordinate various specialized agents that can browse the web, manage local files, and execute Python code, among other functions.

• The system’s modular structure allows easy addition or removal of agents, enabling adaptability for various complex workflows without disrupting the primary system.

• It includes AutoGenBench, an extensible standalone benchmarking tool, to assess Magentic-One's capabilities on agentic benchmarks.

4. OpenAI’s new "Predicted Outputs" feature for GPT-4o and GPT-4o-mini reduces latency by reusing anticipated text segments, allowing developers to input a "prediction string" that speeds up response times, particularly useful for repetitive tasks or minor document edits.

5. OpenAI acquired the high-value domain chat.com from HubSpot founder Dharmesh Shah, who originally purchased it for $15.5M in March ‘23, with the domain now redirecting to the ChatGPT site and reportedly exchanged for OpenAI stock.

6. Anthropic has partnered with Palantir and AWS to provide U.S. defense and intelligence agencies access to its Claude AI models on Palantir’s IL6 platform, certified for handling “secret” data, with the goal of enhancing intelligence analysis, improving data security, and supporting tasks like covert campaign identification and decision-making in defense.

7. AI startup Mistral has launched a customizable content moderation API, powered by its Ministral 8B model, to classify text into categories like hate speech, violence, and PII, with support for multiple languages and adaptable safety standards to address model-generated harms and enforce strong safety measures.

8. Apple has introduced new developer tools for Siri screen awareness, allowing Siri and Apple Intelligence to interact directly with visible content in apps like browsers, documents, and photos, and enabling users to engage with onscreen elements through new APIs without relying on workarounds like screenshots.

9. Apple begins testing smart glasses features with an internal study, codename ‘Atlas’, to explore user-friendly, everyday wearable technology.

10. Google has launched its AI video creation tool, Google Vids, enabling users to create professional-quality videos from Google Drive files and text prompts using the Gemini AI model, with applications for customer support, company updates, training materials, marketing recaps, and project management.

11. T-Mobile is investing $100M in an OpenAI-powered customer service solution, IntentCX, designed to transform customer interactions with advanced, AI-driven decision-making capabilities.

12. Saudi Arabia is set to launch "Project Transcendence," a state-backed AI initiative with up to $100Bn in funding, aiming to establish a competitive technology hub by developing AI infrastructure, data centers, and startups, recruiting talent, fostering a local AI ecosystem, and incentivizing tech companies to establish operations in the country.

💰 Top Funding News:

1. Perplexity, the AI search startup, is finalizing a $500M funding round led by Institutional Venture Partners. 

• This triples its valuation to $9Bn, a massive leap from $520M at the start of 2024.

• Since its founding in 2022, Perplexity’s revenue has surged, now on track to generate approximately $50M annually, a fourfold increase from earlier in the year.

That's all for today's email! If you want more please follow us at the social channels linked below, or check out our website!

Share our newsletter: If you like our work please share/forward this email with your friends, colleagues, and family. It's the best way to support us!

If this email was forwarded to you please sign up here to continue receiving them.

Want your content, product, jobs, or event featured in our newsletter? Reply to this email with the details, and our team will reach out to you.

Do you use AI for work? Tell us how, and you could be featured in our newsletter!

Check out our website for more resources, including a list of AI investors, products, events, and twitter follows.

For an archive of all our posts, click here.

We'd love to hear from you! You can always leave us comments or feedback by replying to this email!

Powered by AI. Curated and edited by Humans.