OpenAI GPT-4 is coming — and Google better be ready
OpenAI GPT-3 is everywhere — and for a good reason. AI has proven superior to anything else out there, as thousands of people use it daily. But this is only the beginning: just recently, Microsoft Germany CTO, Andreas Braun, has confirmed that GPT-4 is coming. And it’s going to be an even bigger deal.
Read also: Microsoft AI Experiment: Is Bing about to become a threat to Google?
What makes OpenAI GPT-4 a big deal
OpenAI GPT-3 can do everything — at least in the realm of text. It can write songs and limericks, research scientific articles, and graduate universities. But GPT-4 will be multimodal, meaning it can give you answers in text, images, video, and speech. It’s not confirmed that it will do all these things right off the bat, but it will support video.
Another big deal (possibly even more important) is how much work is done to make AI more reliable. As we all know, OpenAI GPT-3 is prone to making mistakes. In fact, it lies pretty much all the time, requiring constant proof-checking. The new version should improve its confidence metrics to make data even more reliable.
Different languages will also get improvements, making a system genuinely versatile. You will be able to converse with AI in one language and get answers in another.
If this news is making you think that Google is failing, it’s because it is failing. Well, at least visibly: while the company does promise to release its own alternative to GPT-4, we haven’t seen anything in action yet, and those promising prototypes we saw a few years back never made it outside of closed doors. The future seems to be bright for Microsoft but a lot less so for their biggest internet rivals.