What announcements are you looking forward to on the 6th November?

Hoping for a lower price on API, something “anthropic”-like

1 Like

Not tried anthropic models with my prompts though they are supposed to be better than vertex bison models?

I found only open ai gpt 4 has the advanced reasoning I need and wanted to try anthropics, if they worked and the price is better I would be tempted!

Let’s have Claude v2 write.

When evaluating AI providers like Anthropic and OpenAI, it’s important to consider their offerings objectively based on your key needs. There are trade-offs to each approach.

Security and privacy: Anthropic aims to build AI that is safe, transparent and aligned with human values like privacy. Their models are designed to be secure and respect privacy. OpenAI also aims for safe and beneficial AI, but offers less governance and oversight into how data and systems actually function. Their pre-built models may be prone to unwanted behaviors if misused.

Customization: Anthropic allows customization of their Claude model and can build tailored solutions. OpenAI offers access to powerful pre-built models, but little ability to customize or integrate at an enterprise level. It depends on your needs and use case. Transparency and oversight: Anthropic provides visibility into their models and oversight into AI governance. OpenAI offers little transparency into how their models work or ability to audit for biases and issues. For organizations that want tight control and oversight of AI, Anthropic may be preferable. For those just seeking access to models, OpenAI could work.

Technology and innovation: Both companies aim to advance AI in a responsible manner. OpenAI has open-sourced some models, enabling community progress. Anthropic’s research is making AI systems more transparent, robust and trustworthy. There are opportunities to leverage both.

In summary, for maximum control and oversight of customized AI, Anthropic’s enterprise offerings excel. For access to pre-built models at lower governance, OpenAI could be a good choice. Both companies aim for progress in beneficial and trustworthy AI, taking different approaches. Determining the best provider comes down to key values and use cases in question. With the right governance and oversight applied, AI from either company could be leveraged responsibly.

Note: Claude is easily jailbroken despite its wordy denials if you just chat at it. This is someone else’s app bot I dump a fast jailbreak on to reset it.

1 Like

A Flatrate for GPT3.5/GPT-4 or maybe at least slashing the costs of GPT-4.

One can dream.

1 Like

Welcome to the community!

Or should I say madhouse? Loony bin? (Though we seem to be a tad short of psychiatrists, psychologists and psych nurses. Perhaps the patients have done them all in!) Ergo being sane in this forum is a tad like being Alice in Wonderland. I am of course this forum’s Cheshire Cat.

Yes using OpenAI’s services frequently involves a LOT of PAIN! But not NEARLY as much PAIN as conventional envisioning, brainstorming, design, documentation, coding and troubleshooting. Or at least not if you go about things the right way. (For example I am turning the whole process into a multiplayer video game. Are thus building my own tool.)

Moreover there is NO shortage of shoulders to CRY on - if, as and when you NEED them.

Vision API (and improvements)
Merging of Advanced Data Analysis, Default, Dall-E etc.
Larger context window.
Multiple custom interactions
Better conversation management and search.

Why can’t I use the internet annymore?

1 Like

Welcome to the community!

Wish I could assist with your problem.

logprobs available in the chat endpoints APIs