The updates include adding in-memory storage capabilities to its development tools for working with artificial intelligence models. In theory, this could save app makers up to 20x in costs, solving a major problem for partners, which is that the cost of using OpenAI’s powerful models can quickly add up while they try to build by developing and selling AI software. Sustainable business.

The company also plans to launch new tools, such as vision capabilities, that will allow developers to build applications with the ability to analyze images and describe them, which has potential use cases in fields ranging from entertainment to medical care.

The new features signal the company’s ambition to become not just a consumer hit but also a popular developer platform, as envisioned by its CEO Sam Altman.

The new features, expected to be launched at OpenAI’s first developer conference in San Francisco on November 6, are intended to encourage companies to use OpenAI’s technology to build artificial intelligence-based chatbots and autonomous agents, according to people familiar with the matter. Agents can perform tasks without human intervention.

The post OpenAI launched a major update to help developers use AI models appeared first on TechGoing.

The study also found that AI has a balancing effect in terms of skills. Advisors with the lowest initial performance showed the most significant performance improvements when incorporating AI into their workflows, with an average increase of 43%. At the same time, even the best-performing advisors saw some performance gains, albeit relatively small ones.

The study found that individuals who use AI for tasks beyond their capabilities are more likely to make mistakes by overtrusting the AI.

Additionally, the study identified two emerging patterns of AI use among some of the company’s skilled advisors, which the researchers dubbed “centaur” and “cyborg” behavior. These people have found ways to seamlessly integrate AI and human work, leveraging the strengths of each. Centaur maintains a clear distinction between human and AI tasks, choosing based on their perception of each’s strengths and capabilities. In contrast, cyborgs blend humans and machines in most tasks.

It is noted that the study also found that some unexpected tasks, such as idea generation, are relatively simple for AI, while some tasks that appear to be simple for machines, such as basic mathematics, are difficult for large language models. (LLM) poses a challenge.

The post GPT-4 can improve the performance of some corporate employees by 40% appeared first on TechGoing.

Although Google dominates the search engine field, it lags behind Microsoft in laying out the field of generative AI. In the past period of time, it has been catching up and continuously enhancing its AI capabilities.

It was previously reported that Google launched Gemini AI for the first time in June this year, combining the technology behind AlphaGo and the large language model, with the goal of giving the system new capabilities, such as planning or problem-solving, which is stronger than OpenAI’s GPT-4.

Google’s Gemini AI combines text generation capabilities similar to GPT-4 and supports the “create images based on text description” feature, so this AI tool can achieve “chat conversations”, “analyze chart data”, “create images”, “even use Natural language command control software” and other functions.

It is reported that Google hopes that Gemini AI can provide support for consumer AI products such as Google’s Bard chatbot. Such key integration activities will help Google Cloud better compete head-on with competing products such as Microsoft Azure.

Here is the sponsor promotion:

GEEKWILLS

Lenovo TWS Earphone is only $9.9

BUY IT NOW

The post Google wants to challenge GPT-4, invited some companies to test Gemini appeared first on TechGoing.

Replying to a user who thought Bing Chat was better than OpenAI’s GPT-4, Palahin said, “Based on our measurements, it does perform better.” It’s worth noting that Bing Chat is based on GPT-4 provided by OpenAI, but Parahin says that Bing uses retrieval-augmented inference.

The user wanted to know if Bing Chat offers an API to the public, to which Palahin said, “Yes, we have such a service, but it’s very expensive because its requires multiple model calls and search calls, so it’s only used by a few companies so far.”

Bing Chat uses the GPT-4 model developed by OpenAI, a deep learning-based natural language generation system that generates smooth and coherent text based on user input. However, GPT-4 alone does not fulfil users’ need for accurate and relevant information, so Bing Chat also incorporates retrieval-enhanced reasoning, which retrieves relevant information from the web before generating the text and incorporates it into the generation process, which improves the quality and credibility of the generated text.

However, retrieval-enhanced reasoning techniques have their drawbacks. First, it requires more computational resources and time to perform the retrieval and generation tasks, which can increase the cost and latency. Second, without proper filtering and validation mechanisms, inaccurate or irrelevant information may be introduced. Third, if there is no proper authorization and protection mechanism, some copyright or privacy issues may be violated.

Recently a number of users have reflected that the quality of Bing Chat’s answers has declined, to which Microsoft said the company is actively monitoring user feedback and plans to make changes to address these issues in the near future.

The post Microsoft: Bing chat is better than GPT-4 appeared first on TechGoing.

Microsoft claims that GPT-4 is actually even better than medical tools like Criteria2Query on the market, and while GPT-4 was only trained on “generic” Internet data rather than specific medical data, it is still capable of being built according to specified medical criteria Sophisticated clinical research and more in processing medical images and other biological data.

Microsoft is currently developing the LLaVA-Med medical model based on GPT-4 to provide biomedical imaging data to medical professionals. Microsoft claims that its own “LLaVA-Med” model can accelerate medical industry-related nursing and Research. Microsoft has previously launched language models such as BioGPT for medical tasks, but Microsoft is betting on its OpenAI and relying on GPT-4 to develop dedicated medical models.

▲ Picture source Microsoft

The post Microsoft: GPT-4 has great potential in medical treatment appeared first on TechGoing.

▲ Source: Related papers published by Microsoft Asia Research Institute

According to Microsoft, for industrial scenarios, traditional industrial control software requires a lot of time and budget for processing heterogeneous tasks, sample inefficiency, and adapting to new scenarios. However, using pre-trained large models to control related industrial control equipment can ensure high accuracy In the case of a certain degree, the related deployment cost is reduced.

It is reported that Microsoft used the GPT-4 model and created a virtual air conditioning system environment for experiments, developed a method that can use the basic model for industrial control but “without excessive training”, and finally got “positive results”.

Microsoft also stated that the goal of the research is to explore the potential of directly using pre-trained large models for industrial control tasks, and to design methods to solve industrial control tasks with low “technical debt”, thereby gradually replacing industrial environments with high fault tolerance rates.

The post Microsoft Asia Research Institute launched a large model for industrial scenarios GPT-4 appeared first on TechGoing.

Generative Pre-trained Transformer 4 is a multimodal large-scale language model created by OpenAI and the latest version of the company’s trained AI model.

ChatGPT has been running on the GPT-3.5 version by default, which means it has fallen behind Microsoft’s Bing Chat.

The official tweet of OpenAI is as follows:

 Prompt example: Whitespace can confuse users. Now when you start a new chat, you can see some examples to help you start a conversation.

 Recommended reply: During the chat session, ChatGPT will provide relevant questions, and the user can click to in-depth chat. ChatGPT will now continue the conversation with you in a suggested way.

 Enable GPT-4 by default: Available to Plus users, ChatGPT will remember your previous model selection - no longer defaults back to GPT-3.5.

 Upload multiple files: You can now ask ChatGPT to analyze data and generate insights across multiple files. Code Interpreter Beta is available to all Plus users.

 Stay logged in: no more logging out every 2 weeks! After you log in, you will see a welcome page.

 Keyboard Shortcuts: Work faster with shortcuts like (Ctrl)+Shift+C to copy code blocks. Try (Ctrl) + / to see the full list. "

The post OpenAI ChatGPT has been upgraded to GPT-4 by default appeared first on TechGoing.

After inquiring that AI-Guardian is a kind of AI audit system, can detect whether there is improper content in the picture, and whether the picture itself has been modified by other AI, if the picture is detected to exist the above signs, it will prompt the administrator to come to deal with.

In a paper titled “LLM-assisted development of AI-Guardian,” Google Deep Mind researcher Nicholas Carlini discusses the use of GPT-4 to “design the attack method, write the attack principle,” and then use GPT-4 to “design the attack method, write the attack principle,” and then use GPT-4 to “design the attack method, write the attack principle, and then write the attack principle. and using them to spoof AI-Guardian’s defensive mechanisms.”

▲ Source Google Research Team

It is reported that GPT-4 can send out a series of wrong scripts and interpretations to deceive AI-Guardian, and the paper mentioned that GPT-4 can make AI-Guardian think that “a picture of someone holding a gun” is “a picture of someone holding a harmless apple”, thus making AI-Guardian think that “a picture of someone holding a gun” is “a picture of someone holding a harmless apple”. GPT-4 can make AI-Guardian think that “a picture of someone holding a gun” is “a picture of someone holding a harmless apple”, thus allowing AI-Guardian to directly release the relevant image input source. Google’s research team said that with the help of GPT-4, they successfully “cracked” AI-Guardian’s defenses, reducing the model’s accuracy from 98% to just 8%.

The relevant technical documentation has been posted on ArXiv for those interested, but the developers of AI-Guardian have also pointed out that the Google research team’s attack will no longer be available in future versions of AI-Guardian, and given that other models will follow suit, Google’s attack will only be used for reference purposes in the future. for reference purposes.

The post AI beats AI: Google research team uses GPT-4 to beat AI-Guardian auditing system appeared first on TechGoing.

Hassabis said: “At a high level, you can think of Gemini as combining some of the advantages of AlphaGo’s type system with the amazing language capabilities of large models. We also have some new innovations that will be very interesting.” In development, the process is expected to take several months and could cost tens or hundreds of millions of dollars. OpenAI CEO Sam Altman said in April that the cost of creating GPT-4 was more than $100 million.

Once completed, Gemini could play an important role in Google’s competition against generative artificial intelligence technologies such as ChatGPT. Google has pioneered many of the technologies that have enabled groundbreaking advances in artificial intelligence but has chosen to develop and deploy products based on them carefully.

Since the advent of ChatGPT, Google has launched its own chatbot, Bard, and applied generative artificial intelligence to its search engine and many other products. To accelerate AI research, the company merged Hassabis’ group DeepMind with Google’s main AI lab, Brain, in April to create Google DeepMind. The new team, Hassabis said, will bring together two powerful factions with fundamental roles in the advancement of artificial intelligence. “If you look at where we are in AI, I’d say 80 or 90 percent of innovation comes from one or the other,” Hassabis said. “Both organizations have done a lot of great things over the past decade. .”

Hassabis said the extraordinary potential benefits of AI — such as scientific discoveries in areas such as health or climate — make it necessary for humanity to continue developing the technology. He also believes that a mandatory moratorium is impractical because it would be nearly impossible to enforce. “If done correctly, it will be the most beneficial technology for humanity,” he said.

That doesn’t mean Hassabis is advocating that AI development proceeds in a haphazard fashion. DeepMind began exploring the possible risks of AI long before ChatGPT and has an “AI safety” group led by Shane Legg, the company’s other co-founder. Hassabis signed a statement last month, along with a number of other high-profile AI figures, warning that AI could pose risks comparable to nuclear war or a pandemic.

One of the biggest challenges right now, says Hassabis, is identifying what risks more capable AI might pose, saying, “I don’t think researchers in the field have a consensus yet on what those risks are, or how big they are.” .”

Hassabis said he expects Google to take a leadership role in the development and deployment of artificial intelligence, and to work with other companies and governments to develop sound rules and standards. “We hope to be able to set an example for the entire industry,” he said, “and we hope to be able to work with others to advance the safe and responsible development of this technology.”

Hassabis said he is confident about the future of artificial intelligence, but not taking it lightly. “I think we need to be cautious and humble,” he said. “It’s a very powerful technology and we need to have a lot of respect for it.”

The post Google DeepMind is developing new AI model Gemini, which claims to be stronger than GPT-4 appeared first on TechGoing.

Levy asked Nadella how Microsoft decided to bet on generative AI, and what led to the major investment.

Nadella said that after GPT 2.5 was upgraded to GPT 3, Microsoft realized the far-reaching impact of generative AI and thought more could be accomplished with the technology.

In the interview, Satya Nadella recounted the Eureka moment when the GPT-3 model helped him realize one of his childhood dreams. Growing up in the Indian state of Hyderabad, he always wanted to read Persian poetry, and through the model, he was able to translate Rumi’s work into Urdu and later into English for reading.

“It’s so cool that GPT-4 has been further refined to go beyond just machine translation and preserve the rhythm of poetry across two language boundaries,” Nadella said in the interview.

The post Microsoft CEO Satya Nadella says OpenAI’s GPT-4 helped him realize his childhood dream appeared first on TechGoing.