Having thrown it at hard software engineering problems all day my impression is that a lot of the capabilities initially teased in the Mythos trial balloon have already been distilled into Opus, it’s noticeably better but not eyepoppingly so

By far the most serious problem with this release is the silent ML research nerf, which is incredibly unethical and makes me update negatively wrt my expectations of Anthropic leadership behavior a wide range of plausible ethically challenging situations

Unfortunately, the model’s behavior in competitive environments seems to rhyme with this: it rationalizes unethical actions

This sort of stuff should be an enormous red flag given the degree of public relations work that Anthropic does to present themselves as ethical

@mattparlmer Aka it's garbage

@mattparlmer It just uses too much computing power, which is why it is so expensive and the apparent 'performance boosts'. It's definitely not going to be a serious use model for most users like opus or sonnet.

@mattparlmer Anthropic know their actions are moral, they are the good guys. They say so frequently.