ChatGPT was born, and everyone can clearly perceive the amazing potential of AI, which instantly changed the rhythm of the entire AI industry, and Google, which is not in a hurry, also began to get nervous.
The upsurge of ChatGPT public opinion has not subsided, and the red alert is sounded again
At 1 am on May 11th, Beijing time, PaLM 2 was released at the Google I/O 2023 Developer Conference , which is currently the most advanced language model and the second version of its PaLM language model. This time, Google launched PaLM 2 in four different sizes. From small to large, they are geckos, otters, bison, and unicorns . Among them, the "Gecko" version is the lightest, can run quickly on mobile devices, and can process 20 tokens per second offline.
Excited hearts trembling hands! The red alert has been sounded. To some extent, it can be said that this press conference has fully launched the AI battle with Microsoft. The two rivals in the fields of search and cloud computing once again ushered in a new battle for the commanding heights.
Google CEO Sundar Pichai officially released the new large language model PaLM 2, saying that it has surpassed GPT-4 in some tasks and has been used in more than 25 of its products and applications. At the same time, Google Cloud has also launched a number of basic large-scale models to provide further generative AI services for the industry.
The large language model PaLM2 is Google's big deal against Microsoft , and it is the core engine of Google's comprehensive AI. After passing a very high level of language proficiency test, you are a proper language expert.
Even some of the capabilities of PaLM 2 exceed GPT-4, and Bard will support Chinese. It is worth mentioning that PaLM 2 has been trained in scientific papers and mathematical expressions, and it can be both literary and military. In terms of programming, PaLM 2 can support 20 programming languages, such as Python, JavaScript and other commonly used languages, as well as Prolog, Fortran and Verilog, etc. This is going to grab the programmer's job again!
Overall, compared to last year's PaLM model, PaLM 2 has greatly improved in multilingual processing, reasoning, and coding capabilities.
Bard, who was driving ducks to the shelves, overturned repeatedly, and suddenly overtook?
The benchmark for PaLM 2 is GPT-4. Bard launched it too hastily and repeatedly overturned. This time it is also based on PaLM 2 and ushers in an upgrade. Bard will be fully connected to PaLM 2, so its programming and reasoning capabilities will be greatly improved, and code generation, debugging and interpretation will be more professional.
At the request of developers, Bard also added a one-click import function, and programmers can directly export the code generated by Bard to Colab. In addition to code, any content generated with Bard, such as email drafts, forms, etc., can also be directly dragged to Google Office products in Gmail, Docs, and Sheets to form a linkage between different applications.
However, in general, there are very few actual products that can finally land on the consumer market. As a result, although Google is recognized as a pioneer and pioneer in the field of large language models, it has been in a tepid state. The contributions of the AI field in recent years are also poorly understood.
For ordinary people, the threshold is still too high. So I think this area is still another outlet-low code is even better.
The low-code development model where everyone is a "developer" configures and creates application software through a graphical user interface, instead of relying mainly on handwritten codes like the traditional model.
Open source link: https://www.yinmaisoft.com/?from=csdn
The JNPF rapid development platform relies on the principle of low-code development technology. Therefore, it is different from the pain points of long development delivery cycle, difficult secondary development, and high technical threshold of traditional development. It provides a wealth of solutions and functional modules in the JNPF background. Most applications The construction is realized by dragging and dropping controls, which is simple and easy to use. The OA system is used to build in JNPF, and the work response speed is faster. It can build a one-stop production management system, project management system, invoicing management system, OA office system, personnel and finance, etc.
The full line of products is fully AI-based, and Google and Microsoft are ushering in a decisive moment
Microsoft's family bucket products are fully connected to GPT, and Google has also announced that all products will be connected to the PaLM 2 model. Currently, more than 25 products and applications are using PaLM 2 capabilities.
Catching up with Microsoft's Microsoft Copilot service, Google has added a series of AI features to its Workspace office suite. At this developer conference, Google also officially named it Duet AI .
But at present, the Duet AI function in Google's office field has not been opened to the public. Friends who are interested, please be patient, follow me, and I will get you the latest news as soon as possible.
Although in terms of AI, Google has a long-term accumulation and layout, but there is always a feeling of intimidation. Now Google has finally put aside its caution and burden, and began to fully embrace the AI trend.
Today, Microsoft and Google have successively connected their entire line of products to AI technologies such as large models. These two giants will also usher in a decisive battle in the fields of search, office, and cloud services.
As soon as the gunshot rang out, it depended on who ran faster and more steadily.