Foundations for a Better World - MIT AI Conference 2023 ✨ (Part I)
Foundations for a Better World - MIT AI Conference 2023 ✨ (Part I)
The MIT AI Conference happened on 10/21, with speakers like Sal Khan, Greg Brockman, Sam Altman, Dava Newman, and more
Duanduan had the great pleasure of volunteering at the MIT AI Conference at the Computer History Museum in Mountain View on 10/21. The theme for this year is “Foundations for a Better World - Shaping The Future With Responsibility, Creativity & Collaboration.” The panels are structured to reflect the areas that are at the forefront of AI revolutions: healthcare, transportation, asset management, enterprise, trust & safety, security, supply chain…
This is going to be Part I of a two-part newsletter. Stay tuned for more content next week! Hope you find it enriching and enjoyable! 🤩
🔗 Here is the list of speakers at the conference.
💬 Fireside Chat with Sal Khan & Greg Brockman
Sal Khan: Khan Academy / Greg Brockman: OpenAI / Moderated by: Wendi Zhang (MIT AI Conference Chair)
AI and Education:
With the advent of GPT-4, Brockman saw the potential to provide a more personalized education experience. The synergy between OpenAI's technology and Khan Academy's mission made their partnership in education a natural fit.
LLM and the future of education: While GPT-7 (I wonder how far/close it will be? 🧐) could potentially offer world-class education for free, it's also clear that just having a powerful language model is not enough. There is a need to pair these models with social structures and support mechanisms and find ways to integrate them into traditional education systems. Khan Academy is looking to rethink the paradigms of online education, leveraging AI and machine learning to challenge conventional thinking on topics like app development and internet learning.
Open vs Close AI:
While Khan Academy initially began as an open-source project, Khan believes that cutting-edge models should be in the hands of entities that can be held accountable (e.g. large companies like Google, Meta, Microsoft…). There are concerns about misuse by actors with malicious intent, particularly in areas like surveillance, misinformation campaigns, and fraud.
Greg although is still a proponent of open source, he also believed that we really need to understand where we are going. It is a tough thing to create something extremely powerful and understand how, as a creator, would feel comfortable knowing how others will use it.
Human Skills in the Age of AI:
Skills like judgment, management, and critical thinking are even more important. As AI evolves, human values, goals, and oversight will continue to be crucial.
👩⚕️ Healthcare: Improving Health with AI
Anmol Madan: RadiantGraph / Othman Laraki: Color Health / Moderated by Zeenat Patrawala (VantAI)
Transformation in Healthcare:
Healthcare, despite being data-intensive, remains an analog business. A significant portion of healthcare involves interactions with highly trained personnel. Language models can help systematize these interactions, bridging the knowledge gap in some areas.
AI can be fine-tuned more effectively than humans in some aspects. For instance, it can provide more personalized care and better adjust to individual patient needs, given that most healthcare standards are based on general averages.
Healthcare resisted AI and machine learning use cases. Today, there's more openness to partnerships with tech organizations, which can integrate AI systems into existing healthcare data infrastructures.
The timeline for realizing the full potential of AI in healthcare is seen as long-term, spanning five to ten years.
Data in Healthcare:
A plethora of unused healthcare data exists (e.g., facial cues pointing to health issues). AI can help harness these signals.
Equity & Risk in Healthcare:
AI can introduce new business models in healthcare. For instance, a globally available AI-powered service can offer world-class oncology advice, leveling the playing field between community clinics and top-tier oncology centers.
AI can democratize access to quality healthcare, addressing disparities in access (e.g. digital health tools help make health services more accessible, particularly in distant locations).
Healthcare data can contain inherent biases. AI can either amplify or help counteract these biases, based on its design and application.
📈 AI in Asset Management:
Erkko Etula: Brooklyn Investment Group / Moderated by Jon Xu (Future Advisor)
Transformation in Asset Management:
Data Collection & Analysis: Effective data collection and analysis lead to better investment decisions.
Signal Generation: Advanced models, like transformers, can predict financial returns and enhance investment strategies.
Portfolio Management: AI can automate repetitive tasks, making the process more efficient and scalable.
Compliance: A significant portion of compliance tasks, such as reviewing marketing materials or client communications, can be automated.
Advisor & Client Experience: Modern interfaces driven by AI can enhance the interaction between asset managers and financial advisors.
Role of AI:
Scalable Personalization: AI can automate tasks and provide intuitive user experiences.
Tax Efficiency: using AI to identify tax-loss harvesting opportunities is more efficient and accurate than ever before.
Cost Reduction: By directly investing in underlying assets (for example: investing directly in individual stocks and bonds - no middleman) and leveraging AI, significant cost savings can be achieved, especially in areas of risk management, compliance, and portfolio management.
For private markets: For private companies that lack the rigorous reporting requirements of public companies, language models can become essential research tools. Proper calibration allows portfolio managers to extract useful data from a sea of noisy information.
🔥 AI Deals of the Week
Keywords: CRM, Data Collection
Clearbit is a B2B intelligence platform that collects data about potential customers with AI-powered heuristics. It was valued at $250M post-money for its Series A in 2019.
Brickeye Closes $10M CAD Convertible Debt Round led by BDC Capital
Keywords: Construction, Analytics, IoT
Brickeye is a Toronto, Canada based construction risk management and productivity tool. It incorporates IoT sensors and network to drive its data platform. The funding is consisted of two tranches of $5M CAD each.
Responsiv Raises $3M in Seed Round led by Greylock
Keywords: Legal, Chatbot, Policy
Responsiv builds legal assistant to provide in-house legal teams with answers to legal research questions and policy drafts. It cuts down the cost of legal processes by reducing reliance on external legal counsels and reducing the time taken to conduct legal research from hours to 90 seconds.
Matic Raises $24M in Series A
Keywords: Cleaning, Robot, Data Privacy
Matic is a household cleaning robot that doesn’t send data to cloud. It utilizes computer vision to detect environments in the room and doesn’t require cloud connection to perform predictions. Matic is founded by ex-Google researcher Navneet Dalal and Nest’s Product Manager Mehul Nariyawala.
GlassFlow Raises $1.1M in Pre-Seed
Keywords: Data Pipeline, Event-driven Architecture
Glassflow builds event-driven data pipeline for Python, reducing the cost in training and maintenance when compared with Kafka, which is written for Java. It offers fully-managed serverless setup in AWS and GCP.
Give us a like ❤️ if you enjoyed this! Share your thoughts 💭 below on topics you'd like us to dive into in upcoming newsletters. Your feedback shapes our content!