Calm Down, Twitter: OpenAI's Altman Says No AGI Next Month

OpenAI CEO Sam Altman moved to tamp down spiraling speculation about artificial general intelligence (AGI) on Twitter, telling his followers to "cut your expectations 100x." While hinting at upcoming developments, Altman was unequivocal that the company has neither built AGI nor plans to deploy it next month, pushing back against recent social media hype. This marks another instance of OpenAI leadership attempting to manage public expectations around artificial intelligence developments.

Google's Game-Changing Move in AI Search: What You Need to Know 🚀Big news: Google just partnered with AP to feed real-time news directly into its Gemini AI app. Here's why it matters: This solves one of AI's biggest problems - outdated information. While other AI chatbots often work with stale data, Gemini will now have fresh, verified news from one of the world's most trusted sources. The timing is perfect. Microsoft's Bing, OpenAI's ChatGPT, and others are all racing to make their AI more current. Google just beat them to it by leveraging its deep news industry connections. What's next? Expect Microsoft and others to scramble for similar partnerships. The AI search race just got more interesting.

AI is History's New Unreliable Gatekeeper

Peter Turchin and team unveiled a fascinating project at NeurIPS this week: the first rigorous test of AI's historical knowledge. While machines master coding and math, their grasp of human history remains largely unverified. This new benchmark, developed by historians and computer scientists together, aims to change that—just as these systems begin shaping how we teach and learn about our past. More at NeurIPS 2024.

🧬 Breakthrough: AI Just Revolutionized Protein Engineering

EvolutionaryScale's AI has engineered a jellyfish-inspired glowing protein, bypassing what researchers estimate would take 500 million years of natural evolution.

Here's why this matters: Novel Creation: Their ESM3 AI model didn't just modify existing proteins – it created esmGFP, a functional fluorescent protein sharing only 58% similarity with its closest natural relative.

Design Control: The team demonstrated precise protein engineering by starting with a faint glowing protein and iterating to create one matching natural brightness level.

Practical Applications: Fluorescent proteins are crucial tools in biotechnology, helping scientists track protein behavior in living cells. This breakthrough shows we can create new biological tools from scratch.

Future Impact: This technology opens new possibilities in medicine research and energy solutions . Thesustainable technology that excites me most is how it bridges the gap between computer science and biology. We're not just analyzing life anymore – we're programming it.

Model News

DeepSeek-R1's new reasoning models rival OpenAI-o1, offering open-source innovation with state-of-the-art performance across math, code, and reasoning benchmarks.

Runway unveils a new image generation model offering high-quality photorealism and precise stylistic control.