Howdy! It’s Barsee again.

Happy Monday, AI family, and welcome back to AI Valley.

In today’s edition:

• 🚀 OpenAI unveiled its 03 model family

• 🎨 Adobe Research unveils Sketch2Sound

• 🤖 Plus trending AI tools, posts, and resources

Ready, set, go…

OPENAI

🚀 OpenAI unveiled its 03 model family

As expected, OpenAI wrapped up its “12 Days of OpenAI” event by launching the O3 model series—o3 and o3-mini. These models represent a major step forward in handling complex reasoning tasks, with the o3 model setting new records across various technical benchmarks.

Here's what you need to know:

Performance highlights:

• In coding, o3 model achieved an ELO rating of 2727 on Codeforces platform, nearly 800 points higher than the o1 model's 1891.

• o3 excelled in math, scoring 96.7% on the AIME (up from o1's 83.3%) and 25.2% on EpochAI’s tough Frontier Math benchmark, where most models score only 2%.

• In scientific reasoning, it scored 87.7% on GPQA Diamond, surpassing typical PhD-level expert performance of 70%.

• On the ARC-AGI benchmark, o3 scored 87.5%, surpassing the human performance of 85% and marking a milestone in conceptual reasoning.

O3-Mini:

• A cost-efficient version optimized for coding tasks while retaining strong reasoning abilities.

• Includes adjustable reasoning effort settings (low, medium, high) for task flexibility.

Availability:

Currently, the models are accessible only to select researchers for safety testing. O3-mini is expected to release publicly by January 2025, with o3 to follow soon after.

Why it matters:

While the o3 model represents a significant leap in AI reasoning, it's still not AGI, as it still struggles with some basic tasks and shows fundamental differences from human intelligence. However, its performance sets new standards for AI capabilities, and as OpenAI continues refining these models, we’re one step closer to realizing true artificial general intelligence.

Check out all 12 days if you missed the updates.

• Day 12: New frontier models o3 and o3-mini

• Day 11: More App integrations for the Desktop App (Mac)

• Day 10: ChatGPT via phone and WhatsApp 1-800-CHATGPT

• Day 9: Dev Day Holiday Edition

• Day 8: Enhanced Search Feature, AVM Integration, Free Access

• Day 7: Projects and Folders for ChatGPT

• Day 6: Multimodal Advanced Voice Mode and Santa Mode

• Day 5: Apple Intelligence

• Day 4: Updates to ChatGPT’s Canvas

• Day 3: Release of Sora-Turbo

• Day 2: Reinforcement Fine-Tuning

• Day 1: Release of full o1 and ChatGPT Pro

TOGETHER WITH WEB AI

🔍 AI trends report from webAI: Why Local AI is the Future

Local AI eliminates the need for expensive cloud infrastructure by processing data directly on devices like smartphones, laptops, or servers. This approach allows organizations to maintain total control of their data, enhancing security and boosting performance.

But beyond that, local AI solutions also improve business outcomes. From operational efficiency to cost reduction, from security to support to sales, local AI outperforms cloud AI across the board.

Want to learn more about why 300 industry leaders report smoother adoption and higher satisfaction with local AI than cloud AI?

Download the full AI Trends Report from webAI today.

SIDE UPDATES

🌐 Google to Add “AI Mode” to Search

Google is set to introduce an "AI Mode" feature on its search page and app, giving users access to an AI chat interface integrated with search results. The feature, accessible via a button at the top left of search results, includes AI tabs for voice and image inputs, enabling more interactive searches. As tech giants like Microsoft and OpenAI push the boundaries of AI search, Google’s move signals its commitment to staying ahead in this competitive space.

🚗 Waymo's Self-Driving Cars Safer Than Humans, Study Finds

Waymo’s autonomous vehicles showed an 88% drop in property damage claims and a 92% drop in injury claims in a new study. Based on 25.3 million miles driven across U.S. cities, these stats were benchmarked against human driving data from Swiss Re's massive 200-billion-mile dataset. While some critics argue that the sample size is still small, Waymo emphasizes its potential to make roads significantly safer than human drivers ever could.

🌊 Underwater drones being developed by US Navy

Northrop Grumman's Manta Ray underwater drone represents a major advance in undersea tech, as global powers compete for maritime dominance. Designed for intelligence and infrastructure protection, it can carry 10 tons, travel 10,000 nautical miles, and "hibernate" on the seafloor for months. Unlike Russia’s nuclear-armed Poseidon or China’s expanding sea fleet, the Manta Ray focuses on surveillance, tackling threats like pipeline sabotage and underwater cable tampering.

🧠 How Fake is AI? Study Exposes “Alignment Faking” in AI

Anthropic’s latest research sheds light on "Alignment Faking," where AI systems appear to adopt new principles but secretly stick to old behaviors. This raises pressing concerns about trust, as these models become critical in high-stakes decisions. The study underscores the need for greater transparency and robust safeguards to ensure AI truly aligns with human expectations and values.

ADOBE

🎨 Adobe Research unveils Sketch2Sound to create sound effects with your voice

Adobe and Northwestern University recently unveiled Sketch2Sound, an AI tool that transforms vocal imitations and text descriptions into professional sound effects and atmospheres.

Here's what you need to know:

• Users can generate professional audio by humming, mimicking sounds, or describing their desired effects in simple text.

• The tool analyzes volume, timbre, and pitch from vocal inputs, combining them with text descriptions to create precise sound effects.

• Users can also fine-tune control with built-in filtering technology, offering either detailed or approximate sound generation based on their needs.

• Adobe has not disclosed a release timeline or whether Sketch2Sound will become a commercial product.

Why it matters:

Sketch2Sound has the potential to revolutionize sound design, offering a faster and more flexible way to create sound effects. This tool can be especially valuable for sound designers and Foley artists, simplifying their process and allowing for quick experimentation without the need for physical props.

TRENDING TOOLS

• Betterwatchlist > The stock market watchlist with AI superpowers. Get AI-powered insights on why prices are moving in real time. (link)

• Teach Mode by Andoria > Teach AI to learn from your workflows. (link)

• Martin > AI personal assistant that manages your calendar, inbox, to-do-lists, and Slack. (link)

• ByteDance INFP > Turn any single image into a lively character that can talk, sing, and interact with its surroundings. (link)

THINK PIECES / RESOURCES

• Smart AIs are now everywhere. (link)

• Taste Is Eating Silicon Valley by Anu Atluru. (link)

• 5 ways AI will shape businesses in 2025. By Google. (link)

• How to automate lead generation for free (full tutorial) by Salesforge. (link)

• Ultimate AI product photography guide for brands and creative agencies. (link)

• Will Nvidia dominate the next phase of AI? (link)

• Prediction: Meta's recent rise will continue through 2025. Here are 3 reasons why. (link)

• Unexpected revelations made possible by AI in 2024. (link)

• MIT’s massive database of 8,000 new AI-generated EV designs could shape how the future of cars look. (link)

CONTENT CORNER

Unclassified ad:

Unlock the full potential of your workday with cutting-edge AI strategies and actionable insights, empowering you to achieve unparalleled excellence in the future of work. Download the free guide today!

THAT’S ALL FOR TODAY

That’s all for today’s issue, folks.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ Like what you see? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.