Google Replaces Bard With Gemini On The Day The S&P 500 Index Breaks 5000
The S&P 500 Index, which collected the index to the top 500 market cap among companies listed on the NASDAQ New York Stock Exchange, exceeded 5000 on the last trading day of last week for the first time. This did not happen because all U.S. stocks were booming, but it was the result of four stocks, namely MS, Amazon, and Meta Nvidia, going crazy about adrenaline. Google remained at $150, similar to the previous day, and it is disappointing to see the record of a historic day when Google beat the bad by featuring the ambitious Gemini to catch up with ChatGPT.
But Gemini is much better than Bard. First of all, the underlying model is different. Bard is LLM like GPT, but if Gemini is compared to a large-scale language model (LLM) based on Pathway Language Model (PaLM), before that, if you used to find a book that needed one librarian in a huge library, now many librarians specialized in different fields split their work to bring more professional and reliable information. And unlike Bard and Chat GPT, which only learned text using multi-modal methods, they learned YouTube in large quantities. Paid version users summarize content if they tell them the YouTube address. Bard and ChatGPT were not good before. However, when I see that you can’t read YouTube well without subtitles, I still think that text is based on LLM or PaLM.
Google released Gemini in two versions: the Mu version and the paid version (Ultra), which can be used for $20, and if you pay a fee, Gemini creates more than 500 words of long sentences or coding programs, and draws pictures.
Another difference between free and paid is the ability to prioritize. Both Mini and Bard are interactive AI models, but Gemini has the ability to handle user requests based on priority. In other words, Gemini can understand the user’s intentions and handle the most important requests first, and less important requests later. This is a huge step forward. If Bard and GPT were to just go in order, Gemini would reasonably prioritize them through reasoning and solve them first.

It’s by far the best AI lecture I’ve taken this year: In addition to the advantages of AI and what changes it will make, the U.S. and Western centers point out the positive aspects of AI development. So, in the end, you also clearly state what our company and the government should do.

Summarize with 1 pointer and 1 shot below.

1. There is a big difference between those who use AI and those who do not. AI+ people are more than 20% more efficient at detecting cancer than two people.
2. In addition, more things we do can be replaced by AI. In fact, it will be replaced first by people who use AI.
3. This AI investment is concentrated in one place. 80% of the known investment is concentrated in the United States, and the U.S. is naturally expected to lead.
4. China follows scary. There are as many as 280 LLMs.
5. There is no choice but to be AI-biased led by a specific state in this way. For example, there are problems such as describing An Jung-geun as a terrorist or saying that Dokdo is a disputed area.
6. Even now, data, thoughts, and cultures of countries that do not use the Internet much are completely excluded from learning AI models. Most AI models now are Western-oriented.
7. What’s wrong with this bias? Do you know the yellow tag on YouTube? The AI model is judging by what standards? This way, the AI model can be our judge. By what standards does the model evaluate? We don’t know. We’ll just have to follow suit.
8. So we need to go to AI Nave quickly too. We need to go beyond just using the solution and improve the overall constitution with AI.
9. AI research and development requires huge investment, and the government should step up. The U.S. government also does a lot.
10. Together, we need to create a system that is transparent and unbiased. We need to be as transparent as possible and develop together. It should reflect the positions of various groups.
11. We all need to create human-centered, reliable AI together.