Open this moment: AgentCDN watch page or YouTube.
by Sander Saar
A breakdown of Level 2 reasoning models like OpenAI o1 that use chain of thought reasoning to exceed PhD-level accuracy. Exploring Level 3 AI agents that can plan, reason, and take action autonomously in the workplace.
/public/creators/sander-saar/videos/video_Zc1GTNfiPjw/arvp.json
· Browse this video library
A breakdown of Level 2 reasoning models like OpenAI o1 that use chain of thought reasoning to exceed PhD-level accuracy.
On screen: A man enthusiastically speaks to the camera in front of a split screen comparing ChatGPT o1-preview and ChatGPT 4o.
Exploring Level 3 AI agents that can plan, reason, and take action autonomously in the workplace.
On screen: A man addresses the camera against a scrolling digital product background, highlighting numbered AI classification levels that appear above him.
What the footage shows, shot by shot. It is derived from the video itself, independent of the commentary. Timestamps link to the exact second.
0:00–0:03 A man enthusiastically speaks towards the camera while in front of a scrolling OpenAI webpage and graphic stages. On screen: OpenAI 3 AGENTS 2 REASONERS 1 CHATBOTS
0:03–0:07 The presenter continues speaking about AI levels while the OpenAI website and team article scroll in the background. On screen: OpenAI 5 ORGANIZATIONS 4 INNOVATORS 3 AGENTS 2 REASONERS 1 CHATBOTS Introducing OpenAI o1 before they respond. Here is the latest news Try it in ChatGPT Plus…
0:07–0:20 A man enthusiastically speaks to the camera in front of a split screen comparing ChatGPT o1-preview and ChatGPT 4o. On screen: NOW ChatGPT o1-preview > Create a fully functionoing chess game in a browser Thinking BEFORE ChatGPT 4o > Create a fully functioning chess game in browser…
0:20–0:24 A man gestures and talks in front of an OpenAI evaluation benchmark chart display. On screen: Evals Competition Math (AIME 2024) Competition Code (Codeforces) 100 80 60 40 20 0 accuracy 13.4 56.7 83.3 gpt4o o1 preview o1 percentile 11.0 62.0 89.0…
0:24–0:26 The speaker presents in front of an agent.ai webpage with layered category labels. On screen: agent.ai The Professional 2 REASONERS 1 CHATBOTS A marketplace and professional network for AI agents and the people who love Discover, connect, and hire AI…
0:26–0:30 The presenter continues speaking while highlighting agent levels against a scrolling agent.ai background. On screen: agent.ai 3 AGENTS 2 REASONERS 1 CHATBOTS
0:30–0:31 A man speaks directly to the camera with the Agent.ai marketplace webpage displayed in the background. On screen: agent.ai Agent Marketplace Competitor Analyst By @dharmesh 1 credit per execution 20.3K executions
0:31–0:33 The presenter continues speaking as the Agent.ai webpage behind him scrolls through AI agent cards. On screen: agent.ai Earnings Call Analyzer By @dharmesh 1 credit per execution 995 executions Hire
0:33–0:34 The presenter continues talking in the foreground while the Salesforce Agentforce website is displayed behind him. On screen: Try for free Watch the Agentforce Keynote Dreamforce Keynote: "Build the Future with AI Agents" Watch now Agentforce Step 1: Enable R Humans with Agents drive…
0:35–0:45 A split screen presents a presentation slide explaining Agentforce at the top while Marc Benioff presents to an audience on stage below. On screen: Why settle for copilots when you can have an Agentforce Chatbots Fixed rules & repetitive tasks Copilots Smart, tailored assistance Hey Einstein, can you tell…
0:45–0:57 A man addresses the camera against a scrolling digital product background, highlighting numbered AI classification levels that appear above him. On screen: Try for free Resolve customer inquiries 24/7 2 REASONERS 1 CHATBOTS Sales Development Representative (SDR) Sales Coach 4 INNOVATORS 3 AGENTS 5 ORGANIZATIONS
0:00 Speaker A: We're entering a new level of AI and we'll be moving fast from helpful chatbots to super agents. Speaker A: We just entered level 2 with reasoning AI called O by OpenAI. Speaker A: It's not just answering questions, it's thinking before it answers and provides chain of thought reasoning how it got to the answer when level one chatbots were like your really smart friend O already exceeds human PhD level accuracy in physics, biology and chemistry enabled by reasoning models like O1.
0:27 Speaker A: We are also seeing the first innings of level 3 AI agents Salesforce CEO but looking to bring AI into the workplace through Salesforce Agentforce. Speaker B: We can move from chatbots to co pilots to this new agent force world and it's going to know your business and it can plan, it can reason, it takes action on your behalf. Speaker A: But there's more. Speaker A: Level four brings AI innovators pushing the boundaries of science and technology being able to develop new discoveries and finally Level 5 AI organizations a network of interconnected AI working together to achieve complex goals.