- AI Breakfast
- Posts
- OpenAI's "Strawberry" Model Rumors
OpenAI's "Strawberry" Model Rumors
AI Breakfast
August 28, 2024
Good morning. It’s Wednesday, August 28th.
Did you know: NVIDIA’s earning report is today?
In today’s email:
OpenAI “Strawberry” Update
Gemini Model Update
Claude’s “Artifacts”
Beijing Robotics Conference Review
4 New AI Tools
Latest AI Research Papers
You read. We listen. Let us know what you think by replying to this email.
In partnership with 1440
Seeking impartial news? Meet 1440.
Every day, 3.5 million readers turn to 1440 for their factual news. We sift through 100+ sources to bring you a complete summary of politics, global events, business, and culture, all in a brief 5-minute email. Enjoy an impartial news experience.
Today’s trending AI news stories
OpenAI reportedly preparing to launch new 'Strawberry' AI model
OpenAI is reportedly on the brink of launching its new AI model, “Strawberry,” which could advance towards the goal of artificial general intelligence. Previously known internally as Q*, Strawberry is expected to surpass its predecessors by addressing complex mathematical problems, developing sophisticated marketing strategies, and solving intricate puzzles like the New York Times’ “Connections.” It boasts a MATH benchmark score exceeding 90%, outperforming both GPT-4 and GPT-4o, and positioning OpenAI ahead of its competitors.
The information reporting on OpenAI Strawberry 🍓
Their guess is it’s an advanced reasoning mode that you’ll be able to turn on/off depending on how time sensitive your queries are.
They’re also reporting that OpenAI has been demoing this capability to the NatSec community.
— Bilawal Sidhu (@bilawalsidhu)
2:42 PM • Aug 27, 2024
The model's development has caused internal disruption at OpenAI, including a brief period of CEO Sam Altman’s removal, due to concerns about Strawberry's potential to achieve significant breakthroughs and the associated risks of AGI.
A release is anticipated for the fall, with future scenarios potentially allowing ChatGPT users to switch between Strawberry and other models based on the urgency of their needs. Read more.
Google drops 'stronger' and 'significantly improved' experimental Gemini models
Google's ongoing Gemini updates continue with the introduction of the Gemini 1.5 Flash-8B, alongside enhanced versions of Gemini 1.5 Flash and Gemini 1.5 Pro. These models demonstrate improved performance across various internal benchmarks, with Gemini 1.5 Pro excelling in math, coding, and complex prompts. The smaller Gemini 1.5 Flash-8B variant, released today, boasts significant gains and is positioned as a lightweight alternative to its predecessors.
Today, we are rolling out three experimental models:
- A new smaller variant, Gemini 1.5 Flash-8B
- A stronger Gemini 1.5 Pro model (better on coding & complex prompts)
- A significantly improved Gemini 1.5 Flash modelTry them on , details in 🧵
— Logan Kilpatrick (@OfficialLoganK)
5:09 PM • Aug 27, 2024
According to Logan Kilpatrick, Google's product lead for AI Studio, these models are designed to handle long contexts, processing high-volume multimodal inputs with unprecedented scale. The latest update also made waves on @lmsysorg's Chatbot Arena, where the new Gemini 1.5 Flash-8B surged from #23 to #6 in the rankings, based on over 20,000 community votes. The Gemini 1.5 Pro showed strong gains, particularly in coding and math. Read more.
Anthropic launches Claude Artifacts generally for all users, mobile
Anthropic has made its "Artifacts" feature widely available across all tiers and mobile platforms. This feature, initially released for the Claude family of LLMs, lets users generate and execute interactive code—think charts, games, and visualizations—directly within their chat windows.
Today, we're making Artifacts available for all Claude users. You can now also create and view Artifacts on the Claude iOS and Android apps.
Since launching in preview in June, tens of millions of Artifacts have been created. But where did it all begin?
Here's how we built it.
— Anthropic (@AnthropicAI)
4:00 PM • Aug 27, 2024
Previously a manual toggle, Artifacts is now effortlessly integrated, enhancing accessibility for users on Free, Pro, and Team plans. It supports diverse outputs such as code snippets and interactive dashboards, streamlining workflows for developers, product managers, and designers alike. Free and Pro users can share and remix these creations, fostering a global exchange of ideas, while Team users benefit from secure project-based collaboration.
This rollout highlights Anthropic's emphasis on enhancing user experience and practical applications, placing creativity and functionality above mere processing power. Read more.
China robots conference spotlights the changing face of humanoids
At the World Robot Conference in Beijing, China's progress in humanoid robotics was underscored by innovations from local companies. Wisson Technology showcased its cost-effective robotic arms, which utilize 3D-printed plastics and pneumatic artificial muscles instead of traditional motors and reducers. This approach allows Wisson to offer its arms at a fraction of the cost of conventional models, approximately 10,000 yuan ($1,404). Despite these advancements, some industry experts, like Yi Gang of Ti5 Robot, highlighted persistent reliability issues, particularly with harmonic gears, which limit production volumes.
The conference underscored President Xi Jinping’s vision for technological leadership, with Premier Li Qiang calling for vigorous efforts to stabilize supply chains and enhance the global presence of Chinese robots. Read more.
Etcetera: Stories you may have missed
4 new AI-powered tools from around the web
arXiv is a free online library where researchers share pre-publication papers.
Thank you for reading today’s edition.
Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.
Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email!