OpenAI is falling apart because we’re near the end of what large language models can do, they are intimately aware, and they are fracturing along the lines of AGI True Believers (spend the next decade on another moonshot breakthrough) versus Business People (spend the next decade returning our investment). That’s my best guess.
@kyle Maybe I'm just cynical but I think there's a huge in between, where they want a return on capital faster because they'll go broke otherwise even if there are still big breakthroughs to be had.
@mergesort At a high level I think it’s about the awareness of a technical cliff, a disagreement on how comfortable they are at their altitude, and a disagreement about how fast they need to climb. Both our narratives might fit into that.
Specifically, true believers leaving is sketchy. Similar to conspiracy thinking. If OpenAI really had AGI on the horizon, and if these people really believed that, then they would do anything to do what they thought was necessary. So one condition is false.
@kyle Complete speculation alert but I think my highest probability case is that AGI is possible, not near (say the next 5-10 years), and being underfunded by people who are prioritizing the short term business case because they don't think they'll make it to the long term outcome. That scenario doesn't imply there's a cliff, but that they wouldn't have reached their goals at OpenAI and would have only become harder over time, hence leaving.
@mergesort Yeah that is totally on the table and as likely as my speculation. I’m just hoping to will that future out of existence. So, please stop thinking about it.
Add comment