Boom Times For ChatGPT, OpenAI’s Deep Research, AI Super Bowl

Episode Summary

Executive Summary: The episode argues that AI competition is shifting from models to products and distribution. ChatGPT’s traffic has surged again, DeepSeek has exposed how quickly a strong product can challenge OpenAI, and OpenAI’s Deep Research shows reasoning/agentic tools may be the next major use case. The hosts also debate whether AI ads and marketing are wise, while noting rising capex and job displacement concerns.

Main Topics: ChatGPT’s renewed growth surge (Priority: 5/5): Using SimilarWeb data, the hosts discuss how ChatGPT moved from a long flatline to a major traffic rebound, likely driven by product improvements, broader AI hype, and the launch of voice capabilities. DeepSeek as a competitive wake-up call (Priority: 5/5): DeepSeek’s rapid rise is presented as alarming for OpenAI because it gained substantial usage quickly, exposed reasoning outputs, and showed consumers can switch tools fast when a better product appears. OpenAI’s product strategy and open-source pressure (Priority: 4/5): OpenAI’s Reddit AMA responses are examined for signs the company is rethinking open source, chain-of-thought visibility, and the future of GPT-5 amid stronger competition. Deep Research and the rise of reasoning agents (Priority: 5/5): The hosts praise OpenAI’s Deep Research as genuinely useful on day one, especially for research-heavy work, and frame it as evidence that reasoning plus tool use may be the next AI paradigm. Compute, capex, and the AI infrastructure arms race (Priority: 4/5): OpenAI, Google, Meta, and NVIDIA are discussed in the context of massive spending on compute, with executives arguing more compute still equals better products despite DeepSeek’s cheaper success. AI branding, ads, and the Super Bowl (Priority: 3/5): The hosts debate why OpenAI and Google are buying Super Bowl ads, concluding AI still has a branding problem and that the commercials must feel genuinely useful or funny, not sentimental. AI and labor displacement (Priority: 4/5): The conversation turns to Klarna, consulting, banking, and research workflows, arguing that AI is already eliminating grunt work and will likely reduce hiring even if total jobs evolve over time.

Key Arguments: ChatGPT’s growth rebound suggests OpenAI has found a product improvement—especially voice and better UX—that finally reaccelerated usage after a year of stagnation. The surge is not just an OpenAI story; it reflects that generative AI has become mainstream shorthand for many users, with ChatGPT acting as the household brand. DeepSeek’s rapid adoption shows consumer AI behavior is more fragile than many assumed: if a compelling alternative appears, users can switch quickly. OpenAI’s willingness to talk about open source and showing more chain-of-thought indicates it feels pressure from DeepSeek and broader open-source progress. Deep Research is singled out as a real, day-one useful product because it performs internet-based synthesis, asks follow-up questions, and produces well-cited outputs. Reasoning models plus agents are likely to reshape research, consulting, and knowledge work, even if the term “agentic” is being used too loosely. The AI infrastructure race remains intact: major companies will keep spending aggressively because compute is still a core strategic advantage. AI advertising needs to overcome skepticism; the best campaigns will likely be playful, self-aware, or demonstrably useful rather than emotionally manipulative. AI will not only create new roles; it will also shrink demand for many existing tasks, especially repetitive research, document prep, and synthesis work.

Data Points: ChatGPT web visits per month: ~2 billion flatlined, then rose to ~4 billion - SimilarWeb web traffic chart discussed as ChatGPT reaccelerated after a long plateau ChatGPT users growth (OpenAI-reported): 100 million to 300 million - Referenced as OpenAI’s own reported growth during the rebound period Bing web visits per month (Feb 2024): 1.5 billion - Compared against ChatGPT’s traffic to show uneven AI adoption Bing web visits per month (Oct 2024): 1.85 billion - Shows Bing barely grew despite the AI wave ChatGPT web visits per month (Feb 2024): 1.6 billion - Starting point for the later surge ChatGPT web visits per month (late 2024): 3.7 billion - Shows nearly a doubling in traffic over the period DeepSeek visits (Jan. 28): 49 million web and mobile visits - Used to show how fast DeepSeek gained traction ChatGPT visits (Jan. 28): 139.3 million web and mobile visits - Benchmark against DeepSeek’s spike OpenAI Pro subscription price: $200/month - Price paid by the hosts to access unlimited ChatGPT, Operator, and Deep Research OpenAI Stargate infrastructure plan: $500 billion attempted build - Discussed as the large-scale compute/infrastructure project Google capex plan for 2024: $75 billion - Mentioned as part of the industry-wide compute spending race Google capex spent in 2023: $32.3 billion - Compared with the 2024 capex plan to show the jump in spending Deep Research draft length: 13 pages / 3,778 words - Cited from Ethan Mollick’s example of a Deep Research output Deep Research citations: 6 citations plus additional references - Used to illustrate source quality and depth Klarna automation example: 2 weeks and 16 bankers vs. 5 minutes - Referenced as an example of AI collapsing a once-manual workflow

Pivotal Quotes: "We are on the wrong side of history on open source" — Sam Altman: OpenAI Reddit AMA response about model weights and publishing research "Deep research is fantastic. So basically market research-oriented questions... it does an incredible job with this." — Ranjan Roy: Assessment of OpenAI’s new Deep Research product "It really pushes the point home that whatever happened last week... they are about to be passed by open source and they have to embrace it." — Alex Kantrowitz: Reaction to DeepSeek and OpenAI’s response

Implications: AI competition is increasingly about product usefulness, distribution, and trust—not just model performance. Expect more reasoning/agent tools, heavier capex, sharper open-source pressure, and more automation of knowledge work.