After a few days of trying out a few of the major domestic AIs for text, I’m a bit aesthetically fatigued. Generally speaking, they are all based on ChatGPT to show what they can do. When I think of the freshness of the ChatGPT just came out, I used to play with it all night long day and night, but now I recall it as if it was a lifetime ago.
To put it simply, DeepSeek has increased its arithmetic power by changing its algorithm, which, to put it bluntly, has compressed costs and increased efficiency. It is reasonable to say that this improvement can be ultimately implemented to increase the integrity and accuracy of the data, but DeepSeek itself also admitted that the current data can not. My testing has done nothing more than prove this from another angle. I can’t compare how it compares to ChatGPT, and at the very least, it’s no more in-depth than Doubao or Wen Xiaoyin.
But as the young man in the IT industry told me, the entertainment function of the domestic AI is powerful, which is the reason why it can utilize the talents of young Chinese engineers and have a large youth market to ensure its profitability. And my American classmate who works in IT told me after reading the article that China is the world leader in audio technology.
Gee, isn’t that reflected in Doubao? It’s said that low-profile, practical AI is being developed every day, but it’s just that it’s not as high-profile and publicized as DeepSeek. Beanbag is just a popular free AI that is used more by the public and therefore has more traffic. The niche market can really be officers of the AI small professional and practical and charge.
Even though Beanbag is a popularized AI, its main audio technology still surprised me a bit.
Recorded a section to play around with.
Here’s the chat on Beanbag’s main page. As you can see when I was testing it out, the content of its answers can instantly follow up with changes as you need them, and there’s no sense of detachment or machineness. It also includes voice tones and various dialects, seamlessly. For simplicity’s sake, I skipped the other variations and just picked a dialect. It’s not quite as authentic as Shanghai, but the whisper mode and fricative mode will probably give you goosebumps. Give it a try if you’re interested.
The characterization menu has a wide range of personalities. For example, Lin Daiyu, whose character and tone are taken from the Daiyu model of the TV series Dream of the Red Chamber, can speak according to random conversations that match her character’s personality, and the algorithms in this one should be good. Let’s take a look at a real-time video
I also said I was going to see Bao Hairpin, but I ended up knocking over Lin’s jealousy. This is great for entertainment. When you are bored at home alone, you can travel through time and space to chat with Sister Lin, but of course you have to have a bit of talent to be able to chat with this genius. But what really amazes me is not the above but the fact that it can mimic anyone’s voice as long as you speak to it on demand. Don’t believe me? Check out the little video below, which shows me talking to another me.
I don’t know how you feel about it. Anyway, when I heard the other me, I immediately thought of two consequences
The first is a metaphysical bias towards negativity. What if this technique is utilized by scammers? And I’m sure there will be scammers who will take advantage of it in the Chinese context, especially after DeepSeek lowers the cost of arithmetic. If one day my wife receives such a voice from me asking her to send me money, will she believe it? I’m sure there are still people out there who have been warned about scammers’ tricks countless times, so if you add voice to the mix, the number of people being scammed will double, won’t it?
The second idea seems to be a bit metaphysical. If with the improvement of this kind of audio technology, if we can improve the brain-computer interface, and implant all my memories, when I place me on the Internet, is this me in a way my immortality? Because this me has almost all the memories, rationality, emotions, personality, and even the appearance of my face, except for the fact that I don’t have a physical body. In other words, if my parents were able to record video and voice with the help of beanbags to form a unique online presence, then in a sense they would be alive, and I would be able to access the Internet to have a conversation with them whenever I needed to, and since the content of the conversation is not solidified but randomly generated, how is it fundamentally different from a real WeChat phone call or a video?
My back is a little cold, but I know that the day may not be far away.
And this originally from the entertainment of the thing, but there is a certain possibility of changing human life. At least, our loved ones and friends somehow can always live and communicate with you.
Therefore, I am optimistic about Doubao’s expansion into audio, if it does not just stay in gaming and entertainment, but if it pursues something.
This is a backup number, in case you lose contact, follow this number. Thank you! Here’s the QR code for the main number. If you want to read more, please follow it.
This is the backup number. In case you’ve lost contact with us, you can follow this number. Thank you!
Leave a Reply