Sponsored by

Microsoft Hires Inflection Co-Founder to Run Consumer A.I.

Mustafa Suleyman, co-founder of DeepMind, is leaving his start-up Inflection AI to join Microsoft, aiming to bolster its consumer AI division. Tasked with expanding Microsoft's AI offerings like the Copilot chatbot, Bing, and Edge browser, Suleyman, along with his team and co-founder Karén Simonyan, will work under CEO Satya Nadella. This move reflects Microsoft's intense focus on AI, leveraging Inflection AI's technology and expertise.

Inflection AI, after raising substantial funding, is shifting its focus from consumer applications to commercial AI services, discontinuing its expansion of the Pi assistant. The company's transition includes licensing its technology to Microsoft, aligning with its new strategy. Sean White from Mozilla has been appointed as the new CEO to guide Inflection AI's reoriented mission.

Suleyman's career has been marked by his contributions to AI's evolution, emphasizing both its potential benefits and risks. His perspective on AI's societal impact, as discussed in his book "The Coming Wave," advocates for a balanced approach to AI development, stressing the importance of not centralizing control over AI technologies. The ongoing debate on open-sourcing AI reflects the broader discussion on ensuring the safety and accessibility of AI advancements.

Artificial Intelligence online short course from MIT

Study artificial intelligence and gain the knowledge to support its integration into your organization. If you're looking to gain a competitive edge in today's business world, then this artificial intelligence online course may be the perfect option for you.

On completion of the MIT Artificial Intelligence: Implications for Business Strategy online short course, you’ll gain:

• Key AI management and leadership insights to support informed, strategic decision making.

• A practical grounding in AI and its business applications, helping you to transform your organization into a future-forward business.

• A road map for the strategic implementation of AI technologies in a business context.

Get More Information

‘A landmark moment’: scientists use AI to design antibodies from scratch

Scientists have achieved a significant breakthrough by using generative artificial intelligence (AI) to design new antibodies, marking a pivotal step toward revolutionizing the therapeutic antibody market. The team, led by researchers at the University of Washington, modified an AI tool called RFdiffusion, initially developed for designing mini proteins, to create antibodies that recognize specific proteins related to diseases. This innovative approach could potentially streamline and democratize the process of antibody design, moving away from traditional methods that rely on immunizing animals or screening large molecule libraries.

The modified RFdiffusion tool, which utilizes a neural network similar to those in image-generating AIs, was fine-tuned with data on antibodies and their targets, enabling the design of antibodies against various pathogens and a cancer drug target. Although the success rate was modest, with one in 100 designs proving effective, this represents a significant advancement in using AI for protein design. The researchers successfully tested some of these designs in the lab, confirming their ability to bind to intended targets, albeit with room for improvement in binding strength and compatibility with human immune systems.

This pioneering research is an early but promising step toward leveraging AI in the development of new antibody drugs, potentially transforming the field by making the design process more efficient and accessible. While the generated antibodies are not yet ready for clinical use, requiring enhancements to mimic natural human antibodies and improve efficacy, the success of this project indicates a promising future where designing therapeutic antibodies could be as simple as pressing a button, heralding a new era in drug discovery and biotechnology.

OpenAI is expected to release a 'materially better' GPT-5 for its chatbot mid-year, sources say

OpenAI is preparing to launch GPT-5, the next iteration of its generative AI model, in the upcoming months, targeting a mid-year release. The new model is anticipated to be a significant improvement over GPT-4, with enterprise customers already having glimpses of its capabilities through demonstrations. GPT-5 is currently in the training phase and will undergo extensive safety and integrity testing before its public release, although a specific launch date has not been established.

Despite its success, GPT-4 faced criticism for performance issues, including a phenomenon known as "laziness," where the model would sometimes fail to respond adequately to prompts. The upcoming GPT-5 aims to address these concerns, with expectations running high for enhanced reliability and performance, especially among OpenAI's enterprise customers who form a significant portion of the company's revenue stream.

OpenAI's development of large language models involves training on vast datasets, with the quality of outputs dependent on the training data used. There have been discussions around the ownership and use of training data, with tech companies, including OpenAI, engaging in debates over federal regulations on data usage. As OpenAI progresses with GPT-5, it continues to refine its models to maintain its standing in the rapidly evolving AI landscape.

Other stuff

• Nvidia’s Jensen Huang says AI hallucinations are solvable, artificial general intelligence is 5 years away

• Nvidia’s AI boom is only getting started. Just ask CEO Jensen Huang

• AI nurses powered by Nvidia are now a thing

• Fitbit Using Google Gemini for New AI That Could Become Your Fitness Coach

• Google DeepMind unveils AI football tactics coach honed with Liverpool

• AI researchers have started reviewing their peers using AI assistance

Hope you enjoyed today's newsletter

⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.