Market Hang panelists squared off on September 30 over the intersection of artificial intelligence, oil markets, and regulatory uncertainty in a wide-ranging discussion hosted by Turney Duff.
The panel featured Semafor Business and Finance Editor Liz Hoffman, TMX Wall Street Horizon Head of Corporate Event Research Christine Short, Kelly Intelligence Founder and CEO Kevin Kelly, and Yahoo Finance’s Julie Hyman.
OpenAI’s repeated delay of its IPO, citing safety concerns, drew immediate skepticism from several panelists, with host Duff describing the justification as worthy of an eye roll.
Hoffman argued that safety has evolved from a reputational issue for the major labs into a genuine business concern, particularly for OpenAI, which she described as a consumer-first company facing serious trust challenges.
“The pitch has to be everyone in the world is going to use us, and that is very tough when there’s trust issues around you and particularly around your CEO,” Hoffman said.
Kelly pushed back on the safety framing, arguing that OpenAI’s real concern is the competitive threat from open-source models and the enormous capital expenditure required to sustain its business over the coming decades.
Hoffman drew a pointed historical parallel, citing former Citigroup CEO Chuck Prince’s infamous pre-2008 remark that “as long as the music is playing, we have to keep dancing,” suggesting AI leaders are at least being unusually candid about the pressures they face.
Hyman referenced a Quinnipiac poll finding that roughly seven in ten Americans believe safety is more important than innovation, lending weight to the argument that public concern is a legitimate business risk for the labs.
Hoffman’s article, titled “An ‘AI OPEC,'” explored whether the major artificial intelligence labs could form a cartel-like arrangement to coordinate behavior, drawing comparisons to how early pandemic-era auto executives wanted to collectively shut down factories before their lawyers intervened.
Hoffman expressed skepticism that such an arrangement would be effective, noting that actual OPEC struggles to control production among its own members and is losing relevance as underlying models increasingly resemble a commodity.
“Given enough time and capital, all these labs are going to boil the ocean,” Hoffman said, arguing that large language models will eventually converge and that real value will sit in the application and services layer above them.
The discussion shifted to Robinhood’s newly launched AI agent for trading, which Short described as an interesting democratization of tools previously reserved for institutional investors, though she cautioned that the product had only just launched.
Panelists debated AI agent adoption more broadly, with Hoffman revealing she uses an agent called Muse to manage corporate expenses, while Hyman said she was not yet ready to hand over full authorization for financial tasks.
Kelly framed Robinhood’s move as consistent with its history of competing against established players like Charles Schwab, which recently launched its own AI assistant called Charlie, by pushing capabilities that appeal to its retail trading base.
