Shocking Details About Sam Altman’s Firing
Plus: OpenAI forms an AI safety committee, Ex-OpenAI safety lead joins rival Anthropic to lead a new team.
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 285th edition of The AI Edge newsletter. This edition features “Shocking Details About Sam Altman’s Firing Last Year.”
And a huge shoutout to our amazing readers. We appreciate you😊
In today’s edition:
🔒 OpenAI forms an AI safety committee
😲 Shocking details about Sam Altman's firing last year
🔄 Ex-OpenAI safety lead joins rival Anthropic to lead a new team
📚 Knowledge Nugget: The problem with how we evaluate LLMs by
Let’s go!
OpenAI forms an AI safety committee
OpenAI Board formed a Safety and Security Committee led by directors Bret Taylor, Adam D’Angelo, Nicole Seligman, and Sam Altman. Over the next 90 days, the committee will evaluate and further develop OpenAI’s processes and safeguards. Then, the committee will share its recommendations with the full Board.
Following the full Board’s review, OpenAI will publicly share an update on adopted recommendations. OpenAI has recently begun training its next frontier model, and the resulting systems from the Safety and Security team’s recommendation may bring it safely to the next level of capabilities on its path to AGI.
Why does it matter?
OpenAI has drawn a lot of criticism for putting AI safety on the backseat after its super alignment team was dissolved a few days ago. However, it seems that OpenAI is trying to rise to the occasion and address these concerns while developing AI more responsibly.
Shocking details about Sam Altman’s firing last year
Former OpenAI board member Helen Toner revealed shocking details about Sam Altman’s firing in November 2023. The board had found it difficult to trust Altman after he failed to disclose he owned the OpenAI Startup Fund, gave inaccurate information about the company's safety processes, and tried to push Toner off the board after she published a critical research paper. She claims the board found out about the launch of ChatGPT on Twitter, not from Altman.
Toner also pointed to Altman's history of being fired from previous jobs for "deceptive and chaotic behavior." (Find the full TED AI episode linked below.)
After Altman was initially fired, there was a lot of backlash from OpenAI employees who wanted him reinstated. A week later, he was reinstated as the CEO and a new board was formed.
Why does it matter?
The reasons behind why the chaos ensued remained largely unclear then. The incessant drama and lack of transparency surrounding OpenAI’s leadership, especially Altman, raises critical concerns about the fitness of a company that is currently at the forefront of this AI revolution, and possibly AGI too.
Ex-OpenAI safety lead joins rival Anthropic to lead a new team
Jan Leike, a former co-head of the OpenAI safety team, has joined Anthropic to lead a new "superalignment" team. Leike left OpenAI earlier this month after the team he co-led, which focused on long-term AI safety risks, was dissolved.
At Anthropic, Leike's new team will focus on various aspects of AI safety and security, including "scalable oversight, weak-to-strong generalization, and automated alignment research”. This is similar to the mission of OpenAI's former Superalignment team, which Leike previously led.
Why does it matter?
Anthropic, the No. 2 in AI, has always been more safety-focused in its AI research and development compared to OpenAI's increasing commercial interests. It is no surprise as Anthropic was founded by former OpenAI employees who disagreed with OpenAI’s direction. Leike’s added expertise will only bolster it’s efforts in this crucial area.
Enjoying the daily updates?
Refer your pals to subscribe to our daily newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the “Share” button on any post, you'll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
Knowledge Nugget: The problem with how we evaluate LLMs
discusses the challenges in evaluating LLMs' performance in this article. Traditional evaluation metrics may not be sufficient for broader LLM applications. The author introduces the concept of "LLM-as-a-judge," where one LLM evaluates the output of another LLM. However, this approach doesn’t always provide consistent answers.Addressing ambiguity in evaluation criteria is paramount. The author suggests breaking down vague criteria into more specific, measurable sub-criteria. He also emphasizes the need to tailor evaluation criteria to the particular use case rather than relying on generic, off-the-shelf options. The article suggests that advancements in auto-prompting techniques could be applied to develop "auto-criteria" to improve the evaluation process.
Why does it matter?
Accurate LLM evaluation is crucial. If there are blind spots, you might think an LLM is doing better than it really is, and you might not know how to make it better. "Auto-Criteria" offers a solution by automatically generating relevant evaluation criteria tailored to particular use cases. It can help create meaningful assessments, ensuring LLMs' true capabilities get reflected.
What Else Is Happening❗
💼 PwC set to become the first reseller of ChatGPT Enterprise
Companies no longer need to buy a ChatGPT Enterprise license directly from OpenAI. They can purchase the Gen AI service through PwC. Until today, businesses could only subscribe to the enterprise option by contacting an OpenAI salesperson. This also allows PwC to upsell its services to those who want to use ChatGPT to optimize their workloads. (Link)
🤖 Microsoft brings Copilot AI chatbot to Telegram users
Microsoft has added an official Copilot bot within the messaging app Telegram, which lets users search, ask questions, and converse with the AI chatbot. Copilot for Telegram is currently in beta but is free for Telegram users on mobile or desktop. (Link)
🌐 Opera is integrating Google’s Gemini models into its Aria browser AI
Opera announced a collaboration with Google Cloud to integrate Gemini models into its Aria browser AI. Aria is powered by Opera’s multi-LLM Composer AI engine, allowing it to curate the best user experiences based on their requirements. Thanks to this integration, Opera can now provide its users with the most current information at high performance. (Link)
🚀 GitHub Accelerator program empowers the open-source AI revolution
GitHub launched the 2024 Accelerator program which provides funding, mentorship, and community-building. Standout participants include Unsloth, which improves AI model efficiency, and Formbricks, which simplifies user feedback. Through this, GitHub accelerates open-source innovation and democratizes access to new tech. (Link)
🤼♂️ Elon Musk vs. Yaan LeCun on X highlights differences in AI research approach
Elon Musk and Yann LeCun, two prominent figures in AI, got into a heated debate on X over the weekend. LeCun criticized Musk's management style at his new AI startup xAI, while Musk questioned LeCun's recent scientific contributions. LeCun emphasizes the importance of open scientific publication, while Musk focuses on ambitious goals like artificial general intelligence (AGI). (Link))
New to the newsletter?
The AI Edge keeps engineering leaders & AI enthusiasts like you on the cutting edge of AI. From machine learning to ChatGPT to generative AI and large language models, we break down the latest AI developments and how you can apply them in your work.
Thanks for reading, and see you tomorrow. 😊