AI Breakfast
Posts
OpenAI’s ‘Projects’ Released

OpenAI’s ‘Projects’ Released

AI Breakfast
December 16, 2024

In partnership with

Good morning. It’s Monday, December 16th.

Holiday Special (today only): For the next 24hrs, you can become a lifetime subscriber of an ad-free version of AI Breakfast for a one-time payment of just $20?

In today’s email:

OpenAI’s ‘Projects’ Released
Meta Joins Musk in OpenAI Lawsuit
Grok-2 Upgrade
Ilya’s AI Prediction
4 New AI Tools
Latest AI Research Papers

You read. We listen. Let us know what you think by replying to this email.

Here’s Why Over 4 Million Professionals Read Morning Brew

Business news explained in plain English
Straight facts, zero fluff, & plenty of puns
100% free

See for yourself

Today’s trending AI news stories

From Chaos to Cohesion: OpenAI’s Day 7 Introduces ‘Projects’ in ChatGPT

OpenAI’s new tool "Projects" in ChatGPT, enables users to curate and structure files and conversations within customizable folders. Accessible via the ChatGPT website and Windows desktop app for Plus, Pro, and Teams subscribers, Projects provides a consolidated workspace designed for efficiency. Enterprise and Edu users will gain access in January, reflecting OpenAI’s gradual rollout strategy.

This feature integrates custom instructions, allowing users to tailor ChatGPT’s responses for specific projects. It supports file uploads, chat references, and seamless movement of conversations into Projects, creating a unified repository for ongoing tasks. OpenAI has pledged to broaden file compatibility and incorporate cloud integrations, including Google Drive and OneDrive, in the coming year.

By centralizing functionality and retaining key features like Canvas for coding and Voice Mode, OpenAI underscores its ambition to position ChatGPT as an indispensable productivity hub. Read more.

Meta Joins Musk in Legal War Over OpenAI's For-Profit Shift

Meta Platforms and Elon Musk, two unlikely allies, are uniting against OpenAI’s shift to a for-profit model. In a letter to California Attorney General Rob Bonta, Meta accuses OpenAI of weaponising its nonprofit status for financial gain, warning that this could set a dangerous precedent for the AI industry. Musk, who has filed legal injunctions, echoes these concerns, alleging that OpenAI has abandoned its original mission to develop AI for the public good in favour of investor profits.

Ironically, leaked communications suggest Musk himself once supported OpenAI’s transition to a for-profit entity, adding a layer of complexity to his legal challenge. Meta’s backing of Musk underscores shared concerns about market consolidation, particularly as OpenAI’s $157 billion valuation and connections to Microsoft amplify industry tensions.

OpenAI's new detailed blog post responds, with Sam Altman stating, "You can't sue your way to AGI," reaffirming the company's focus on advancing artificial general intelligence. The post also reveals Musk’s early push to take control of OpenAI, proposing a Tesla merger. After these attempts were rejected, Musk resigned, later accusing the company of a “bait and switch,” stating, “De facto. I provided almost all the seed, A and most of B round funding.” Read more.

Grok-2: xAI’s New Speed and Source-Citing Powerplay

xAI has rolled out an upgraded Grok AI chatbot that now operates three times faster than its predecessor. The model now rivals Sonnet 3.5 in its ability to follow prompts, with marked improvements in accuracy and language support.

A key addition is Grok’s ability to cite sources from both 𝕏 posts and legacy media, offering up-to-date, fact-checked responses—though xAI remains mum on whether it’s compensating these publishers.

The "Grok button" is also a new feature, providing additional context and image explanations. Alongside the Grok-2-1212 and Grok-2-vision-1212 models, xAI is preparing to release the Aurora image model through its API. Musk’s upcoming Grok 3 promises to push the envelope further.

In other news, Elon Musk has revealed plans for 𝕏Mail, a new email service designed to challenge incumbents like Gmail. The announcement has sparked debate among users, with many intrigued by the prospect of enhanced privacy features and deeper integration with the 𝕏 platform. However, there is notable skepticism around the need for yet another email service. Read more.

OpenAI cofounder Ilya Sutskever says the way AI is built is about to change

Ilya Sutskever, OpenAI co-founder and former chief scientist, announced a shift in AI development at NeurIPS, declaring, "We've achieved peak data and there'll be no more." He highlighted that the AI industry is approaching the limits of new data, likening it to the depletion of fossil fuels, with the internet representing a finite resource.

Ilya Sutskever @ilyasut NeurIPS 2024 full talk
— ZILIANG Peng 🏴‍☠️ (@cyberpengk)
1:00 AM • Dec 14, 2024

Looking to the future, Sutskever predicted that AI models will evolve into agentic systems capable of independent decision-making and reasoning, rather than merely pattern-matching from historical data. These models, he noted, will be more unpredictable but will possess the ability to reason from limited information.

Sutskever also compared the scaling of AI systems to evolutionary biology, suggesting that AI’s growth will parallel the relationship between brain and body mass across species. Read more.