Hacker News Reader

Comment by XenophileJKO | Hacker News Reader

I think their current advantage is willingness to risk public usage of frontier technology. This has been and I predict will be their unique dynamic. It forced the entire market to react, but they are still reacting reluctantly. I just played with Gemini this morning for example and it won't make an image with a person in it at all. I think that is all you need to know about most of the competition.

View on HN

How about Anthropic?

Aren't they essentially run by safetyists? So they would be less willing to release a model that pushes the boundaries of capability and agency

From what I've seen, Claude Sonnet 3.5 is decidedly less "safe" than GPT-4o, by the relatively new politicized understanding of "safety".

Anthropic takes safety to mean "let's not teach people how to build thermite bombs, engineer grey goo nanobots, or genome-targeted viruses", which is the traditional futurist concern with AI safety.

OpenAI and Google safety teams are far more concerned with revising history, protecting egos, and coddling the precious feelings of their users. As long as no fee-fees are hurt, it's full speed ahead to paperclip maximization.

Not to dispute your particular comment, which I think is right, but it's worth pointing out we're full steam ahead on paperclips regardless of any AI company. This has been true for some 300 years, longer depending how flexible we are with definitions and where we locate inflection points

This has not been my experience. Twice in the last week I've had Claude refuse to answer questions about a specific racial separatist group (nothing about their ideology, just their name and facts about their membership) and questions about unconventional ways to assess job candidates. Both times I turned to ChatGPT and it gave me an answer immediately

As an AI model I can't comment on this claim.

I think Anthropic is a serious technical competitor and I personally use their product more than OpenAI, BUT again I think their corporate cautiousness will have them always +/- a small delta from OpenAI's models. I just don't see them taking the risk of releasing a step function model before OpenAI or another competitor. I would love to be proven wrong. I am a little curious if the market pressures are getting to them since they updated their "Responsible Scaling Policy".

Reply on news.ycombinator.com