OpenAI releases the next ChatGPT model, GPT-5
Lightcap says that GPT-5 combines two things into this one model. Previously, depending on the type of query asked, Most users would let the app use the default standard mode. The user could change modes to the one that allowed ChatGPT to reason. As Lightcap said, “Sometimes you choose a thinking model, sometimes you wouldn’t. And that was, I think, a confusing experience for users.” Once you type in or verbalize your question, GPT-5 decides which mode is best for the user. It’s a win-win for users that is sure to result in better responses.
“So if it’s faster, for example, we think that’s actually indicative of it being better. If it can give you a better answer per unit of time thinking, we think that’s an improvement that’s an important vector to measure also.”
-Brad Lightcap, Chief Operating Officer, OpenAI
The interviewer, Alex Kantrowitz from Big Technology, asked the OpenAI executive why he led with the news about the reasoning vs. non-reasoning modes instead of talking about the increase in intelligence expected with the new GPT-5 model. Lightcap’s response gives us an insight into how ChatGPT works. The COO says that intelligence is based on the amount of time that the model is thinking.
OpenAI releases GPT-5 for the ChatGPT app. | Image credit-PhoneArena
In the same vein, Lightcap says about the new model, “Typically, the longer it thinks, the better an answer it can give you. So when we test the model on certain benchmarks and evals, and we allow it to think, it will dramatically outperform any of our existing models by far. He also states that GPT-5 does deliver “a dramatic improvement in intelligence,” although he goes on to say that the time the model has to reason is more important.
While GPT-5 scored better on all of the academic evaluations it was put through, OpenAI made the decision to have it score higher on health benchmarks, which Lightcap says makes it “better at medical reasoning.” Instead of going to Dr. Google for a diagnosis, perhaps a visit to Dr. ChatGPT is the better idea. Actually, a visit to a real doctor is your best plan, and surely anyone at OpenAI would tell you the same thing.
Siri, on the left, and ChatGPT respond to the same query. | Image credit-PhoneArena
Will the average user see and appreciate the improvements to the new model? Probably not, if the OpenAI executive is right. He points out that “things like structured thinking, problem solving, tool use” are measured by the company. But even taking that into consideration, these things and others that take place under the hood are not necessarily easy for users to see and appreciate. Still, with that in mind, Lightcap says that all of the aforementioned things are better on GPT-5 than they were on previous models.
How to top the responses you get from Siri and other digital assistants
While some consumers are still frightened about AI taking over their jobs or even taking over the world after a war against humans, at least at the current stage, AI is a tool that is like a super-intelligent digital assistant on steroids and speed. If you have yet to use ChatGPT as a replacement for Siri, you will be amazed. Check out the images we embedded into this article, asking Siri and ChatGPT’s GPT-5 “What can you tell me about my favorite phone of all time, the HTC One (M8)?
Keep in mind that if you do replace your digital assistant with ChatGPT, Gemini, or another AI-based app, you will have to use your current assistant to set timers and alarms. Or, you can do those tasks manually. It’s a small price to pay to receive much better responses to your queries.
#rid #digital #assistant #replace #app
