Hi, this is Wayne again with a topic “OpenAI launches its talkative new model GPT-4o | TechCrunch Minute”.
Open ai’s latest model is a lot more talkative if you’ve been following the headlines around open AI you’ve probably seen stories about how the company is working on a search product to compete with Google. It was even supposed to announce that product this week, which didn’t happen. That means a search product is probably still in the works. What open AI just announced is arguably even more impressive. Its latest model GPT 40. The O stands for Omni, which is a reference to all the different ways you can interact with it. The biggest change is the addition of Voice open AI previous lead model have been trained on images and text now, GPT 40 adds voice into that mix. Voice is not an entirely new addition to chat GPT but chat. Gpt’S previous voice capabilities were powered by speech to text, so the interaction wasn’t really that differ from typing a question and reading an answer.
So, with these new capabilities chat, GPT is getting closer to becoming a real science fictional Voice Assistant, one that you can actually talk to, as if it were a human assistant. For one thing, you can now interrupt chat GPT in the middle of an answer. It’S also supposed to be able to detect the emotion in your voice and answer with aemotion of its own, and it can even sing. Gpt 40 also improves chat, gpt’s Vision capabilities, so you could share a photo or your desktop screen and ask it what’s going on with this software code or what kind of shirt is this guy wearing open a also released a bunch of videos showing what GPT 40 Can do, of course, a pre-planned demo isn’t the same thing as testing a product in the wild, but those videos are pretty darn impressive in one you can see: GPT 40 tutoring a student in math, offering guidance and feed feedback as they work through a math problem.
In another GPT laughs at and compliments the user’s dad jokes, which is kind of sad but hey, we’ve all wanted that kind of affirmation. One video hints at the accessibility implications of improved vision and voice where GPT describes and answers questions about what it sees to a user whose vision impaired. Maybe the craziest, are the videos where two phones with GPT 40 interact with each other.
For example, with one GPT asking the other to describe what it’s seeing and then the two of them singing together, unlike some other open AI announ ments, where the company only provides limited access to the new technology, GPT 40 is available in the free tier of chat. Gpt right now, making these models available to everyone was a big theme in the presentation. So it’s also announcing a new chat, GPT UI on the web and the fact that features like remembering preferences are now available to free users.
We’Ll see you tomorrow, .