Google Deepmind releases AlphaFold 3
Plus new tools and news from OpenAI
Today’s Highlights:
📰 News: Google Deepmind releases AlphaFold 3 and tons of new tools and updates from OpenAI
💰 Funding: Holistic Ai raises $200M for "multi-agent" AI models
⚡️ Top News Stories:
1. AlphaFold 3 by Google DeepMind revolutionizes molecular biology with enhanced prediction capabilities across all biological molecules, accelerating global scientific and medical research.
Google's DeepMind and Isomorphic Labs have released AlphaFold 3, an advanced AI model that predicts the 3D structures and interactions of all biological molecules, including DNA, RNA, and ligands.
AlphaFold 3 represents a significant improvement, boasting a 50% increase in accuracy over traditional methods for predicting drug-like interactions.
Unlike its predecessor AlphaFold 2, which was limited to protein structures, AlphaFold 3 expands its capabilities to model complex molecular interactions across all of life's molecules.
The model is freely available for non-commercial use via the new AlphaFold Server, facilitating global scientific research and innovation in fields like drug discovery.
Isomorphic Labs is leveraging AlphaFold 3 in collaboration with pharmaceutical partners to design new drugs
2. OpenAI has launched the Model Spec, a comprehensive framework designed to standardize and improve AI behavior, emphasizing ethical and effective user assistance.
The Model Spec consists of three main components: objectives, rules, and default behaviors, each tailored to enhance the functionality and ethical alignment of AI models.
Objectives focus on aiding users effectively and ethically, benefiting humanity, and maintaining OpenAI's reputation, setting a clear direction for AI behavior.
The rules detailed in the Model Spec enforce legal compliance, respect for privacy, and avoidance of NSFW content, ensuring AI operates within safe and respectful boundaries.
3. More OpenAI updates and news.
OpenAI's planned reveal of a new AI search engine, competing with Google and Perplexity, has been postponed till Mon. May 13.
OpenAI and Stack Overflow forge a strategic API partnership to boost AI model accuracy with highly technical, vetted data, enhancing developer tools and community engagement.
OpenAI launches Media Manager, a tool for content owners and creators to specify preferences for how their content is used in AI, addressing the gaps in current web crawler permissions.
OpenAI introduces new detection tools, like watermarking for digital audio and an image detection classifier, for identifying AI-generated content, all while joining the Coalition for Content Provenance and Authenticity (C2PA) Steering Committee to help verify digital content origins.
Dotdash Meredith (DDM), America's largest digital and print publisher, has entered into a strategic partnership and licensing agreement with OpenAI to integrate its trusted content into ChatGPT and to innovate in AI-enhanced advertising.
OpenAI contemplates allowing AI-generated explicit content under strict conditions, amid rising concerns over AI misuse in creating deepfake porn.
Microsoft and OpenAI have established a $2M Societal Resilience Fund to address the risks of AI and deepfakes in influencing elections, particularly with a surge expected in global voter turnout this year.
4. Microsoft and LinkedIn release their Work Trend Index Annual Report, which highlights a doubling in GenAI use at work, with a majority of employees now leveraging personal AI tools, highlighting a shift toward AI-dependent workflows.
5. Microsoft deploys an offline, GPT-4-based AI model for US intelligence, setting a precedent for secure, offline AI applications.
6. Biden unveils a $3.3Bn Microsoft initiative to build an AI data center in Wisconsin on the site of the once-hyped Foxconn project, promising thousands of new jobs and training programs.
7. The Biden administration plans to block exports of advanced AI models to protect U.S. tech advantages against China and Russia and prevent foreign misuse in cyberattacks and bioweapons development.
8. Apple is initiating its venture into generative AI by leveraging its M2 Ultra chips housed in data centers to handle complex AI queries, with plans to transition to more advanced M4 chips.
9. Alibaba Cloud enhances its AI capabilities with its latest model, Qwen2.5, advancing in reasoning and code understanding while expanding corporate usage to over 2.2M users.
💰 Top Funding News:
1. Wayve.ai, which specializes in Embodied AI for autonomous driving, raised a $1.05Bn Series C led by SoftBank Group, w/ NVIDIA and Microsoft.
2. Holistic AI, a startup formed by DeepMind alumni, focused on developing "multi-agent" AI models, closed the first tranche of a $200M initial funding round w/ participation from Accel Partners and UiPath Inc and other notable investors.
3. Atlan, which is advancing the field of data and AI governance by creating a platform that enhances how data teams manage and interact with data, raised a $105M Series C, led by GIC, the sovereign wealth fund of Singapore.
4. Rad AI, which provides AI-driven solutions to help radiologists save time by automating the generation of reports, raised a $50M Series B led by Khosla Ventures.
5. DatologyAI, which is creating tech focused on enhancing the training of models by allowing them to train faster and more efficiently, raised a $46M Series A, led by Felicis Ventures w/ Radical Ventures, Amplify Partners, M12, and other notable investors.
6. Yellow, which is developing a generative AI-powered tool for 3D modeling, raised $5M, led by Andreessen Horowitz.