3 ms·
That is my best guess too, which is an even better case for better-than-Opus-4.8 models being available from OpenAI without drawing the attention of the adminis
by furyofantares 4mo ago
That is my best guess too, which is an even better case for better-than-Opus-4.8 models being available from OpenAI without drawing the attention of the administration.
I do think their marketing hype was a factor though even in this narrative. Andrew Jassy tells a vengeful admin that there's this jailbreak for a model Anthropic said was dangerous, and vengeful admin seizes the opportunity to take vengeance.
- ethbr1 4mo agoI think any optimism about future OpenAI models is discounting the avarice of the current US administration. Having succeeded with Anthropic, it's reasonable to expect their next attempt will be: "Nice new model, OpenAI, would be a shame if national security concerns tangled up a public release. Now let's talk about equity stakes..."
- furyofantares 4mo agoMy belief that OpenAI will release a model better than Opus 4.8 globally should not be interpreted as a claim that they will avoid a shakedown. I think if they encounter corruption, they will engage with it.