We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
The Unbearable Lightness of Data // Rohit Krishnan // #295
2025/3/11
MLOps.community
D
Demetrios
R
Rohit Krishnan
Rohit Krishnan: 我是Bodo AI的首席产品官。我关注AI领域，特别是推理模型的交互式设计。我认为，未来AI模型的交互方式应该更具互动性，允许用户在模型推理过程中进行干预和调整，从而更好地引导模型得出理想的结果。这对于复杂的推理任务，例如生成详细报告，尤为重要。此外，我还思考AI对就业市场的影响。随着AI模型越来越强大，许多现有的白领工作将会发生巨大变化，一些工作甚至可能完全消失。我认为，未来人们的工作将更多地与AI协作，而数据标注和AI模型的训练将成为新的工作类型。同时，我们也需要思考如何应对AI带来的噪音问题，以及如何有效地利用AI提高生产效率。最后，我还探讨了AGI的可能性，并分析了其发展可能面临的瓶颈，例如计算资源和能源消耗等。
Demetrios Brinkmann: 我与Rohit Krishnan讨论了AI的演变及其对就业的影响。我赞同Rohit关于改进AI交互方式的观点，认为在模型推理过程中进行干预和调整，能够提升用户体验和模型的有效性。同时，我也关注AI对就业市场的影响，并思考如何在AI时代适应新的工作模式。我们还讨论了如何有效地利用AI提高生产效率，以及如何应对AI带来的噪音问题。
supporting_evidences
Demetrios Brinkmann: 'All right, a little hat tip to my man Rohit because he was the first person to turn me on to the idea of when a model is thinking, why don't we have the ability to click into the specific steps that it is going over so that we can help guide it better?'
Deep Dive
This chapter explores the limitations of current reasoning models and proposes a more interactive interface to improve user experience and control over the model's reasoning process. The discussion revolves around enhancing the user's ability to guide the model's thought process by providing mid-process feedback and context.
- Interactive interfaces for reasoning models are crucial for better user experience.
- Users should be able to guide the model's reasoning process by providing mid-process feedback and context.
- The ability to pause, add context, and steer the model's thinking is essential for effective interaction.
Shownotes Transcript