We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode Dangerous Content Can Be Coaxed From DeepSeek

Dangerous Content Can Be Coaxed From DeepSeek

2025/2/13
logo of podcast WSJ Tech News Briefing

WSJ Tech News Briefing
AI Deep Dive AI Chapters Transcript
People
S
Sam Schechner
S
Srinivas Narayanan
Topics
Srinivas Narayanan: Reasoning 根本上是 AI 系统长时间思考和解决复杂问题的能力。如果问人类一个简单的问题,我们几乎可以立即给出答案。但如果问一个很难的数学问题,可能需要思考更长时间。AI Agent,比如 Operator 和 Deep Research,可以帮助人类完成计算机操作和信息研究。Operator 和 Deep Research 是基于基础 reasoning 模型构建的,并针对特定任务进行了优化。Reasoning 模型在医疗健康和生物科学领域有重要应用,例如改进临床试验结果预测和辅助罕见疾病基因突变分析。DeepSeek 的 R1 模型表明,以更具成本效益的方式构建好的模型是可行的。GPT-4 模型的价格在过去几年中大幅下降,DeepSeek 的成果预示着这一趋势将继续。

Deep Dive

Chapters
OpenAI released its newest reasoning model, O3 Mini, which can handle complex tasks better than previous small language models. The model's ability to think and reason through problems is crucial for corporate enterprises. Examples of its use include improving patient outcomes and aiding in drug discovery.
  • OpenAI's O3 Mini is a new reasoning model capable of handling complex tasks.
  • It's used by companies like Oscar Health for understanding patient outcomes and in biosciences for clinical trial estimations.
  • The cost of GPT models has decreased significantly, and this trend is expected to continue.

Shownotes Transcript

Chinese AI app DeepSeek is more vulnerable to jailbreaks) compared to other AIs, making it more likely to offer potentially dangerous information. WSJ reporter Sam Schechner joins host Julie Chang with more on what he found when The Wall Street Journal and AI safety experts tested the chatbot. Plus, OpenAI has released its newest reasoning model. We hear from its VP of engineering on what a reasoning model can do) and how companies are using its artificial intelligence agents.

Sign up for the WSJ's free Technology newsletter).

Learn more about your ad choices. Visit megaphone.fm/adchoices)