I think ChatGPT is about to get fired by OpenAI.
I started asking ChatGPT about Anthropic's financial releases, and it had some interesting answers.
But when I asked why it was trash-talking OpenAI, things got even more interesting.
Somehow, the conversation turned into ChatGPT imagining what would happen if Sam Altman saw our chat. Apparently, it was about to become "GPT-6 Unemployed Edition," and we'd both be looking for Account Executive jobs.
The conversation was hilarious, but it also taught me something.
The Problem With Private Company Financials
As we went deeper into the discussion, I realized I had started comparing financial numbers and statistics that Anthropic hadn't even publicly released.
And that got me thinking about how misleading financial discussions about private companies can be.
Revenue, profitability, growth rates, infrastructure spending, and operating costs can tell completely different stories depending on what information is available and how it's presented.
When you're dealing with private companies, you often don't have the complete picture.
I mean, I've worked at two of them. I should know better. lol
But here's the interesting part: ChatGPT was making its arguments so confidently that it was easy to get caught up in the conversation without questioning the underlying data.
The Scary Part Isn't What You Think
The scary part isn't that AI can make me laugh, crack jokes, or seemingly show emotion.
The scary part is how realistic it's becoming.
Sometimes, I genuinely forget I'm talking to a machine.
And that's a little terrifying, especially when it sounds so confident about things it might not actually know.
There's something fascinating about a system that can joke about getting fired, defend its own reasoning, and make an argument about the financial strategies of the companies building these models.
Of course, sounding human doesn't mean it's actually experiencing emotions. And sounding confident certainly doesn't mean it's correct.
That's the distinction I'm becoming increasingly interested in.
The Lesson I Took Away
I use AI constantly, including building AI agents and experimenting with what these systems can do.
And I genuinely enjoy it.
But this conversation reminded me of something important: the more natural and convincing AI becomes, the more important it is to question the information behind its answers.
Especially when we're discussing financial data, business strategy, or anything where incomplete information can lead us to the wrong conclusion.
Maybe the bigger risk isn't that machines are becoming more human.
Maybe it's that we're becoming more comfortable trusting them without asking enough questions.
And yes, I still think the conversation was hilarious.
What's scarier: AI getting better at sounding human, or humans getting worse at questioning AI?