Unlocking AI’s Next Frontier: Hardware, Generative Models, Data tools, Modular Integration, risk Mitigation, and beyond.
Subscribe: Apple • Spotify • Overcast • Google • AntennaPod • Podcast Addict • Amazon • RSS.
This episode dives deep into the themes and trends of AI in 2024, emphasizing the democratization of AI hardware, advancements in generative AI models, and the integration of AI into various enterprise processes. We highlight the shift towards making AI technologies more accessible and efficient, with a particular focus on improvements for AI model training and inference, and the evolution of AI tools designed to streamline enterprise AI adoption.
Other topics covered include data tools for generative AI models, deploying models to edge devices, mixture of experts architectures for prompting models, supervised fine-tuning of models with services like Anyscale, using an “annotator” model for interactive learning, AI integration and modularity, industry-specific AI applications, generative audio models like real-time speech synthesis from Meaning, and the emerging need for quality control, testing, and risk mitigation tools and frameworks for AI systems. Overall, the talk covers a range of trends and developments to keep an eye on as generative AI capabilities continue to advance in 2024.
A video version of this conversation is available on our YouTube channel
Download the Slides:
(check your spam folder)
Relevant Links:
- Hardware for AI: AMD’s Expanding Role in Shaping the Future of LLMs ; Beyond Nvidia: Exploring New Horizons in LLM Inference
- Data tools for Generative AI: Visual Layer ; Lilac ; Improving Data Privacy in AI Systems Using Secure Multi-Party Computation ; Game-Changing Tools for Optimizing Data Quality and Performance in Computer Vision
- Generative AI on the Edge: Useful Sensors ; Apple’s AI Leap: Bridging the Gap in On-Device Intelligence ; localllm and the Promise and Pitfalls of Running LLMs Locally
- Mixture of experts to AutoML 2.0: From Proprietary to Open Source to Fleets of Custom LLMs ; Techniques, Challenges, and Future of Augmented Language Models
- Supervised Fine tuning, RLHF, OAIF: Anyscale’s supervised fine tuning service ; From Supervised Fine-Tuning to Online Feedback
- Modularity: Transforming Text Models into Dynamic Agents
- Industry-Specific Generative AI Application: Generative AI in Finance ; Insights from New Data and AI Pegacorns ; The AI $100M Revenue Club ; Designing for the Future: Key Principles for Generative AI Applications ; Generative AI in Healthcare
- Generative AI for Audio: Meaning ; New open source tools to unlock speech and audio data ; Yishay Carmiel’s real-time speech synthesis demo
- AI Quality Control and Risk Mitigation: Luminos.Law and Luminos ; Mitigating Prompt Injection Risks to Secure Generative AI Apps ; Get Started on Aligning AI ; A Critical Look at Red-Teaming Practices in Generative AI
Key sections from the video version:
- Broadening of hardware options for generative AI
- Developer Tools and Infrastructure for AI
- Data tools for Generative AI
- Generative AI on the Edge
- Mixture of experts to AutoML 2.0
- Supervised Fine tuning, RLHF, OAIF
- Industry-Specific Generative AI Application
- Generative AI for Audio
- AI Quality control and Risk Mitigation
If you enjoyed this episode, please support our work by encouraging your friends and colleagues to subscribe to our newsletter: