Daily AI Newsletter: Nobel Prize in Physics for the work done in foundational AI

In today’s email:

🇺🇸 OpenAI says bad actors are using its platform to disrupt elections but with little ‘viral engagement’
🤖 Rise of the robots: AI to shape UK defense review
🚙 Amazon’s new AI-powered vision tech tells drivers which packages to deliver
🧰 11 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

Geoff Hinton and John Hopfield win Nobel Prize in Physics for their work in foundational AI

Geoff Hinton and John Hopfield have jointly won the 2024 Nobel Prize in Physics for their pioneering work on artificial neural networks. The Royal Swedish Academy of Sciences recognized their contributions, which began in the late 1970s and early 1980s, as foundational in enabling modern machine learning with artificial neural networks. Their research laid the groundwork for much of the advancements in AI, which has emerged as a key driving force behind what some refer to as the fourth industrial revolution.

Hinton, often called the "godfather of deep learning," gained a Ph.D. in artificial intelligence in 1978 and went on to co-create the backpropagation algorithm, a crucial method that allows neural networks to learn from their mistakes. This work transformed how AI models are trained. In 2013, Hinton joined Google after the acquisition of his company, DNNresearch, but left the tech giant last year, expressing concerns about AI's role in spreading misinformation. He currently serves as a professor at the University of Toronto.

John Hopfield, a professor at Princeton, also made seminal contributions to artificial intelligence. He developed the Hopfield network, a type of neural network that demonstrated how computational systems could mimic the functioning of human memory by storing and retrieving patterns. His work showed how principles from biology and physics could be applied to AI, significantly advancing the field.

The Nobel Prize committee highlighted the broad impact of their work, noting that artificial neural networks are being used across numerous fields, including physics, to develop materials with novel properties. In addition to the prestigious recognition, Hinton and Hopfield will share the cash prize of 11 million Swedish kronor (approximately $1 million) and each receive a gold medal and a diploma.

Finally, a developer tool that makes digging around for the answer a thing of the past.

Other developer tools can’t tell you how your codebase works and why. Unblocked can. We augment your source code with context from Slack, Confluence, Jira (and more), so your team gets helpful answers without having to search for them.

Unblocked gives you:

Instant answers about every aspect of your codebase
Relevant, historical context for any code open in an IDE
Automated answers to questions asked in Slack, removing interruptions to you

⌛Engineers who use Unblocked save an hour or more a day.

“Every developer now has the ability to tap into past discussions and decisions to fill their knowledge gaps, regardless of their tenure. We are moving faster and making more accurate decisions as a team as a result.” - Alex Mallet, EVP of Engineering at Forto

Start your 21-day trial today

OpenAI pursues public benefit structure to fend off hostile takeovers

OpenAI is exploring a shift to a public benefit corporation (PBC) structure to protect itself from hostile takeovers and activist investors. The move, which follows the company's recent $6.6 billion fundraising round, aims to ensure that the company maintains its mission-oriented focus while balancing the interests of shareholders, employees, and society. This type of corporate structure is relatively rare, with fewer than 20 publicly traded companies in the US adopting it, but has become popular among AI companies, including Anthropic and Elon Musk's xAI. The PBC model offers flexibility to fend off acquisition attempts and demands from activist investors, giving OpenAI a potential safeguard against external pressures.

The proposed change comes during a period of growth and challenges for OpenAI, now one of Silicon Valley's most valuable companies with a valuation of $150 billion. The restructuring would retain a not-for-profit entity as part of the overall corporate structure, with that entity focusing on pursuing the mission of benefiting humanity. Sam Altman, OpenAI's CEO, is expected to lead the for-profit PBC, which will concentrate on developing technology and expanding its commercial success. Meanwhile, the not-for-profit entity would remain independent, led by a different executive.

The PBC structure could provide a "safe harbour" for OpenAI's leadership to prioritize long-term objectives over short-term profit pressures, allowing them to turn away activist demands. This setup may also allow the company to navigate challenges and scrutiny from investors and lawmakers who have expressed concern over the impact of artificial intelligence on society. Legal experts have noted that the PBC model gives management more leeway from the traditional fiduciary duty of maximizing shareholder wealth, allowing them to focus on a broader mission.

Despite the potential advantages, the PBC model is largely untested, both in court and in terms of public perception. As AI companies face increasing scrutiny from lawmakers over their societal impacts, OpenAI hopes that the PBC structure will offer an effective way to balance its technological ambitions with its stated mission of benefiting humanity. The company remains committed to creating AI that benefits everyone, with the non-profit arm continuing to play a vital role in pursuing that mission.

Why Vertical LLM Agents Are The New $1 Billion SaaS Opportunities

Jake Heller, founder of CaseText, was interviewed on an episode of "The Light Cone" podcast, where he shared the story of his company and its dramatic growth after adopting advanced AI technologies. Jake had been working on CaseText for over a decade, initially building a tool to make legal research more efficient. The legal industry, notorious for its reliance on outdated tools and time-consuming tasks, presented Jake with numerous challenges. The early stages of CaseText focused on combining legal content with better technology, aiming to improve how lawyers conducted research. Despite setbacks and pivots, including an unsuccessful attempt to involve lawyers in user-generated content, Jake and his team made steady progress by leveraging natural language processing and machine learning.

The turning point for CaseText came when they gained early access to OpenAI's GPT-4 technology. Recognizing the immense potential of this powerful language model, Jake made the bold decision to pivot the entire company within 48 hours, shifting all focus to building a new AI-powered product called "CoCounsel." This move took 120 employees off their current projects and redirected their efforts toward utilizing GPT-4 to create an AI legal assistant that could significantly improve the efficiency of lawyers by summarizing documents, performing legal research, and providing insights faster than ever before. The company's commitment and rapid iteration led to impressive results—so much so that within just two months of launching CoCounsel, CaseText was acquired by Thomson Reuters for $650 million.

During the interview, Jake emphasized the importance of thorough testing and iteration in building AI-powered solutions. He and his team applied a test-driven development approach to prompt engineering, writing thousands of tests to ensure that the AI consistently delivered accurate results. This focus on precision and reliability was crucial in gaining the trust of lawyers, who were initially skeptical of new technology that could potentially alter their workflows. Jake shared his excitement for the new OpenAI model, GPT-4 Turbo (or GPT-4.5), which he described as a significant improvement, capable of performing more detailed, nuanced tasks, and hinted at the potential of enhancing the model's performance through domain-specific expertise.