Everyone, even those who haven’t made an account, will be able to use OpenAI’s most popular talking AI. Of course, all of your chats will still be used for training unless you choose not to, so it won’t be quite the same.
You won’t have to log in to chat.openai.com anymore when you visit, at least not right away in a few places. This change will be slowly rolled out to the rest of the world. Instead, you’ll be thrown right into a chat with ChatGPT, which will work the same way for people who are logged in.
You can chat all you want, but keep in mind that you’re not getting as many benefits as people who have accounts. With a persistent account, you can save and share chats, use custom directions, and do other things that you normally can’t do without one.
You can still choose not to have your chats used for training, though, which seems to go against the whole point of the company doing this in the first place. Click the little question mark in the bottom right corner, then click “settings,” and turn off the feature there. This useful gif is available from OpenAI:
It’s even better that this version of ChatGPT will have “slightly more restrictive content policies.” What does it mean? I asked, and a spokesperson gave me a long, mostly pointless answer:
The safety features that are already built into the model, like not making dangerous content, will help the signed-out experience. In addition to these existing protections, we are also adding more safeguards that are meant to deal with other types of content that may not be appropriate for a signed out experience.
Taking into account what we know about GPT-3.5’s features and the risk studies we’ve done, we thought about the bad ways that a service that is logged out could be used.
In other words, I have no idea what these stricter rules are. We will definitely find out soon, when a huge number of randos visit the site to test drive this brand-new product. “We understand that more work may be needed and would like to hear your thoughts,” the representative said. And they will receive it in large amounts!
At that point, I also asked if they had a plan for how to deal with the attempt to abuse and use the model as a weapon on a size that has never been seen before. Inference is still pricey, and even the GPT-3.5 model, which is refined and has low lift, needs power and server room. It will get hammered for all it’s worth.
They Also Had S Long, Vague Snswer For This Threat:
We’ve also given a lot of thought to how we can find and stop people from abusing the signed out experience. The teams whose job it is to find, stop, and respond to abuse have been involved in the design and implementation of this experience and will continue to do so in the future.
Take note of how little information there is that is actually useful. They likely don’t know what people will do with this thing any more than anyone else, so they will have to respond instead of being proactive.
Also Read: The Deals That Openai Has Made With Companies Could Be Bad for Competitors
It’s not clear which places or groups will be able to use ChatGPT for free first, but it started today, so keep checking back to see if you’re one of the lucky ones.
What do you say about this story? Visit Parhlo World For more.
