Trending December 2023 # Openai’s Chatgpt Update Brings Improved Accuracy # Suggested January 2024 # Top 16 Popular

You are reading the article Openai’s Chatgpt Update Brings Improved Accuracy updated in December 2023 on the website We hope that the information we have shared is helpful to you. If you find the content interesting and meaningful, please share it with your friends and continue to follow and support us for the latest updates. Suggested January 2024 Openai’s Chatgpt Update Brings Improved Accuracy

OpenAI, the company behind the popular conversational AI model, ChatGPT, has released an update to improve the chatbot’s accuracy.

Following an extended period of downtime on Tuesday, ChatGPT is up and running with a new model.

This is the first update to ChatGPT this year and the second update to the model since its launch in November.

A popup message lists the changes in what OpenAI calls the “Jan 9 version” update.

“We made more improvements to the ChatGPT model! It should be generally better across a wide range of topics and has improved factuality.”

In addition, you can now stop ChatGPT while it’s in the process of generating a response, which is a heavily requested feature.

Details in the release notes are limited, so I tested several prompts in ChatGPT to learn more about the new update. Here’s what I discovered.

Data Still Limited To 2023

While ChatGPT may be more accurate now, its knowledge is still limited to 2023 data.

When asked what’s in the new ChatGPT update, the AI tells me:

“ChatGPT is a large language model developed by OpenAI that is trained on a dataset of conversational text. Since it’s an OpenAI model, It’s updated regularly with new features and improvements. However, as my knowledge cut off is in 2023 I don’t have the latest updates and I cannot provide you a specific information on which new features or improvements have been added to the model in recent updates.”

When asked another way, I was able to get more insight into what’s involved in improving the accuracy of an AI model:

“In general, when an update is made to an AI model like ChatGPT, it can be aimed to improve its performance in one or multiple aspects. For example, it can have a bigger dataset to train on, or improvements in its architecture that allows it to generate more fluent and accurate responses. And such update also make it better in factuality, i.e its ability to understand and generate responses based on factual information.”

Accuracy Remains Questionable

OpenAI claims ChatGPT can now provide more factual answers.

I tested that claim by consulting a GitHub repository of ChatGPT failures and running several prompts to see if it would produce different answers.

Related: ChatGPT Examples: 5 Ways SEOs and Digital Marketers Can Use ChatGPT

Test One: Failed

Previously, ChatGPT could not accurately identify how many times Argentina won the FIFA World Cup.

Disregarding the 2023 World Cup win, because ChatGPT’s knowledge is limited to 2023, it should say Argentina has won it two times. Once in 1978 and again in 1986.

As shown in the tweet below, ChatGPT didn’t always return the right answer:

— indranil sinharoy (@indranil_leo) December 29, 2023

I ran the prompt through the updated version of ChatGPT, and it returned a different but still incorrect answer.

Test Two: Failed

Previously, ChatGPT was unable to provide a correct answer when asked who is the taller basketball player between Shaq and Yao Ming.

I ran the prompt through the updated version of ChatGPT, and it confidently returned the same incorrect answer.

Going through the ChatGPT failures linked above, I found it continues to struggle with the same prompts.

It’s difficult to pinpoint the areas in which ChatGPT can return more accurate responses. It would be helpful if OpenAI could provide specific details in the release notes of future updates.

That said, be careful when using ChatGPT as a source of information. Although it provides correct answers to many questions, it’s currently not dependable enough to replace Google.

Source: OpenAI

Featured Image: CHUAN CHUAN/Shutterstock

You're reading Openai’s Chatgpt Update Brings Improved Accuracy

Rackspace Brings Openstack To Datacenters

Rackspace is moving beyond its own walls to help others deliver and deploy cloud solutions. Rackspace is one of the leaders behind the open source OpenStack cloud platform effort.

The Rackspace Cloud Edition will provide datacenters with an OpenStack cloud that has the operational support and managed services backing of Rackspace.

Cloud Storage and Backup Benefits

Protecting your company’s data is critical. Cloud storage with automated backup is scalable, flexible and provides peace of mind. Cobalt Iron’s enterprise-grade backup and recovery solution is known for its hands-free automation and reliability, at a lower cost. Cloud backup that just works.


“Rackspace Cloud Private Edition is a set of reference architectures based on our own real world experience operating one of the largest clouds out there,” Mark Collier, vice president of marketing and business development at Rackspace told chúng tôi “On an ongoing bases we’ll also have managed services for OpenStack, helping people to operate and run clouds based on our experience.”

In just over a year of existence, OpenStack has become a major force in the world of cloud computing. Initially OpenStack was started by Rackspace together with NASA and today has over 90 contributing member companies. The most recent OpenStack release, codenamed Diablo, was unveiled in September and includes new networking and scheduling capabilities.

While OpenStack is a software release, the Rackspace architecture does mention both server and networking hardware. Collier noted that the reference architecture recommends using Dell C-Series servers and Cisco for the networking gear. That said, Rackspace is a member of the Facebook-led Open Compute initiative as well, which is building out open hardware to help improve datacenter efficiency.

“We’re trying to take our commitment to openness beyond just the software and really start to go out to the market and identify other hardware configurations that we know will work,” Collier said. “So when we turn around and help people operate OpenStack clouds, we’re confident that we can do our job as the managed services for OpenStack piece.”

On the managed services side, Collier said that over the years, Rackspace has developed a lot of tools to help remotely manage clouds. He noted that remote management is something that Rackspace understands well from their own day-to-day operations.

“The vast majority of Rackspace employees never set foot in our datacenter,” Collier said. “We own the datacenters and we control the facilities but because of the abstraction of the cloud model, there is less need to actually have a physical presence in the datacenter.”

For example, Collier said that on the storage side there is enough replication within a datacenter cloud environment that if there is a storage failure it’s not a critical event. An administrator can take their time to replace the failed device.

“As we see the cloud revolution taking off and that technology reducing the need to physically be in the datacenter, we can take those types of tools that we used today to operate our cloud and use them in customer environment to help them run their own clouds,” Collier said.

Sean Michael Kerner is a senior editor at chúng tôi the news service of chúng tôi the network for technology professionals.

Flash Tricks For Improved Search Engine Rankings

Flash Tricks for Improved Search Engine Rankings

Let’s first take a look at how search engine indexing can cause you problems on your web site.

Most web sites are built up of menus and context areas. The menus are frequently text based, making them easy to update or change. The content is dependent on our writing creativity. Both of these can lead to search engine indexing trouble.

Search Engines look through the text on your pages, menus as well as content and they create their index on what they find. So far so good. But just how do the search engines do this? They can’t look at your page and decide which is the main content area visually, so they simply start at the top of the code and work down.

If your site follows the standard pattern of a navigation bar on either the top or down the left side of the page and uses a table structure to achieve this, then your whole nav bar will be read and indexed before your main content area. If your site has a lot of variation then this shouldn’t be a problem. But what if your site is focused on one subject and your navigation bar tends to repeat words? As an example you may have a site that sells watches and your nav bar may read like this: Men’s Watches, Ladies Watches, Sport Watches, etc. You can see how easy it is to repeat that word Watches.

Search engines like to give points to sites that contain valuable content that is easily categorized and recognizable to visitors, but they also take away points for keyword spamming. In the above Watch example, the nav bar could easily cause your page to be listed as a keyword spammer.

Here is the first Flash Trick to improve your ranking. Create the navigation bar in Flash. This way all those repeating words are now hidden from the search engine spiders. As an added benefit the code taken up by the Flash will probably be less than the code used in the text based nav bar. This will help the search engine spiders to focus on the main content area of your page.

Let’s now look at another common problem with search engine indexing. In this example consider a shopping site selling the same watches as in our previous example. Each watch page will have a description of the individual watch, and that is fine. But each page may also have “boiler plate” text as well. There may possibly be a standard description for a particular watch brand, or possibly warranty or shipping information included on the page.

Another red flag that goes up for the search engine spiders is text repeating from page to page. The more distinct each page is the more likely the search engines will consider the text as relevant. If there is too much repeated text, the search engines may even drop all the pages that they believe have duplicated text. Not a good situation, especially if you don’t want to be forced into creating completely original text for every page on your site.

Here is Flash Trick number two. Keep all the distinct content on your pages as html text and convert any repeating text areas into Flash files that are placed into the pages. This way, only the distinct text is visible to the search engines and your repeating text is hidden in the Flash file. Any text that you tend to repeat from page to page is a prime candidate for the Flash treatment.

George Peirson is a successful Internet Trainer and is the author of over 30 multimedia based tutorial training titles covering such topics as Photoshop, Flash and Dreamweaver. To read his other articles and see his training sets visit HowToGurus.

Interpreting Loss And Accuracy Of A Machine Learning Model

Machines are getting more intelligent than ever in the modern world. This is mostly brought on by machine learning’s rising significance. The process of teaching computers to learn from data and then utilize that information to make judgments or predictions is known as machine learning. Understanding how to judge the performance of these models is essential as more and more sectors start to rely on machine learning. In this blog article, we’ll examine the machine learning concepts of loss and accuracy and how they can be used to evaluate model efficacy.

What is Loss in Machine Learning?

In machine learning, loss refers to the error between expected and actual data. A machine learning model’s objective is to reduce this error or loss function. The loss function is a mathematical function that measures the discrepancy between output values that are expected and those that are actually produced. The performance of the model improves with decreasing loss. The gradients required to update the model’s parameters during training are calculated using the loss function, which is a crucial step in the training process. Depending on the issue being addressed, several loss functions are employed, such as cross-entropy loss for classification problems and mean squared error for regression problems. Since increasing prediction accuracy is the ultimate aim of every machine learning model, minimization of the loss function is essential. Developers and data scientists can build better models and boost their performance by grasping the idea of loss in machine learning.

What is Accuracy in Machine Learning?

In machine learning, accuracy is a crucial parameter for gauging how well the model predicts the future. It is calculated as the proportion of accurate forecasts to all of the model’s predictions. The performance of the model improves with increasing precision. When solving classification issues, accuracy is crucial since the model must accurately categorize examples into several groups. For instance, the proportion of emails that are accurately categorized as spam or not spam in a spam detection system serves as a gauge of the model’s accuracy. In many applications, maximizing accuracy is essential since poor forecasts might have serious repercussions.

Interpreting Loss and Accuracy Context of the Problem Being Solved

In machine learning, it’s essential to comprehend the context of the issue being handled in order to interpret a model’s performance. Different issues call for various accuracy and loss trade-offs. For instance, reducing false negatives is more crucial than reducing false positives in a medical diagnosis system. Maximizing accuracy is more significant in a fraud detection system than maximizing recall. Developers and data scientists can construct relevant metrics for evaluating the performance of the model by first understanding the context of the issue.

Trade-off Between Loss and Accuracy

In machine learning, loss, and accuracy are frequently trade-offs. A model that maximizes accuracy may not always be one that minimizes the loss function and vice versa. For instance, a model that overfits the training data in image recognition tasks may have a low loss yet perform badly on fresh data. In contrast, a model that underfits can have a bigger loss yet perform better with fresh data. The trade-off between accuracy and loss relies on the particular issue being resolved as well as the limitations of the application.

Importance of Considering the Validation Set

A validation set is a crucial consideration when evaluating a machine learning model’s performance. A portion of the dataset called the validation set is left aside so that the model can be tested on new data. When a model performs well on training data but poorly on new data, this helps prevent overfitting. Overfitting can be discovered by comparing the model’s performance on the validation set with the training set. Developers and data scientists can prevent overfitting by carefully weighing the model’s hyperparameters while monitoring the model’s accuracy and loss on the validation set.


To sum up, assessing a machine learning model’s loss and accuracy is a crucial stage in the machine learning process. The model’s performance can be evaluated, modifications can be made with knowledge, and the problem can be solved as intended can all be done by developers and data scientists. A machine learning model’s performance should be interpreted by taking into account the trade-offs between loss and accuracy, the context of the issue being solved, and the use of an appropriate validation set.

Emeritus Brings Tech Upskilling To Global Fortune 500 Company

blog / Learner Stories Emeritus Brings Tech and AI Upskilling to Global Fortune 500 Company

Share link

In the case study below, we explore how Emeritus Enterprise worked with a client to solve a problem through our learning solutions.

The Client

A multinational Fortune 500 company with a workforce of more than 100,000 employees. 

The Problem

Throughout its long history, the client has constantly evolved its products and updated how it provides value to customers. In recent years, the organization has undertaken a digital transformation in part by integrating artificial intelligence (AI) and machine learning (ML) across its various lines of business.

Broadly speaking, PwC data found that 52% of companies accelerated their AI adoption plans in response to the COVID-19 pandemic. And this will continue into the future.

The client deploys AI and ML in a variety of ways. Whether in customer service, marketing, finance, manufacturing, or research and development, AI technologies have the potential to positively impact customers, employees, stakeholders, and shareholders.

Client leaders identified that managers would be crucial to an organizational culture shift that prioritized digital transformation. Those leaders needed to be able to grasp the technical aspects of AI well enough to communicate effectively with technical teams and colleagues.

After consulting with stakeholders, it was clear that the following problem statements existed:

How can we help teams navigate digital transformation and applications of AI/ML?

How can AI accelerate consumer insights or category growth?

What AI experiments should we run?

How can we avoid any detrimental uses of AI that may impact customers, employees, stakeholders, etc.?

The Solution

This client piloted a course with the University of California at Berkeley and Emeritus, enrolling a few employees to test the curriculum. The course, Artificial Intelligence: Business Strategies and Applications, was an eight-week, cohort-based online learning journey. Learners  interacted with professionals based around the world as well as with Berkeley faculty and practicing subject matter experts. 

The course covered a range of topics to help the organization’s leaders learn how to organize and manage successful AI application projects.

After pilot groups gave excellent reviews, the client decided to scale up deployment. In November 2023, a cohort of about 40 managers started the public course with learners from various other companies. But the client also partnered with Emeritus’ Enterprise team to supplement the eight-week course with two private additional sessions for its employees. These sessions provided a forum to discuss the application of AI within the client’s industry segment and specific AI projects that employees could implement within the organization. 

Led by an industry expert, the private sessions helped strengthen relationships among employees that will create ongoing learning communities. This development opportunity benefitted both the employees and employer, enhancing individual growth while building a strong, skilled, and committed workforce that benefits the future growth of the company.

Based on the success of the first cohort, a second cohort of about 40 learners was enrolled in the same AI course with Berkeley in March 2023. Three supplemental private office hours were also added. During the second cohort, leaders engaged in intimate conversations about the application of AI in their current roles and began designing capstone projects that would directly apply to their business unit.

The Results

The feedback from employees so far has been positive. In fact, the client’s research and development teams are moving forward with several of the projects that employees completed in the course. 

Learners shared that they appreciated learning the fundamentals of AI and feel confident in applying AI practices to help the larger organization in ways they had not previously thought about. The learners also said they valued the ease of self-paced independent learning and appreciated the ways complex concepts were explained with a clear structure in bite-sized pieces. 

The course included a capstone project that learners worked on individually and then received feedback on from the course leaders. 

Partner with Emeritus Enterprise

Are you ready to explore how Emeritus Enterprise solutions can help your company meet and exceed its goals? Reach out to discuss our customized online employee training programs.

Asking Bing Chat To Be More Creative Will Decrease Its Accuracy

Microsoft began rolling out the new Bing Chat response options at the end of last week. (This reporter does not yet have access to them on his personal account.) Mike Davidson, corporate vice president of Design and Research at Microsoft shared a screenshot:

Microsoft is attempting to balance what it apparently sees as Bing’s core function: a “copilot for the web.” It’s never been quite clear what that entirely entails, but, initially, it seemed like Microsoft intended Bing Chat to be a tool to supplement its traditional search engine: summarizing results pulled from a variety of sites, to save users the need to dig for those results on their own. Some of the more creative elements, such as the ability to tell stories and write poems, were apparently seen as bonuses.

Perhaps unfortunately for Microsoft, it was these creative elements that users latched on to, building on what rival OpenAI’s ChatGPT allowed. When journalists and testers began pushing the limits of what Bing could do, they ended up with some bizarre results, such as threats and weird inquiries about relationships. In response, Microsoft clamped down hard, limiting replies and essentially blocking Bing’s more entertaining responses.

Microsoft is apparently trying to resuscitate Bing’s more creative impulses with the additional controls. But there’s apparently a cost for doing so, based on my own questions to Davidson. Large language models sometimes “hallucinate” (make up) false facts, which many reporters have noticed when closely querying ChatGPT and other chatbots. (It’s presumably one of the reasons Bing Chat cites its sources through footnotes.)

I asked Davidson whether or not the creative or precise modes would affect the factual accuracy of the responses, or whether Bing would adopt a more creative or factual tone instead.

Yep. The first thing you said. Not just tone in a colloquial sense.

— Mike Davidson (@mikeindustries) February 25, 2023

What Davidson is saying is that if you opt for the more creative response, you run the risk of Bing inventing information. On the other hand, the “creative” toggle presumably is designed for more creative output, where absolute accuracy isn’t a priority.

Just to be sure, I asked for clarification. Davidson went on to say that if users want an entirely accurate response, it comes at the cost of creativity. Eliminating creative responses on the basis of inaccuracy defeats the purpose. In time, however, that may change.

With the state of LLMs right now, it’s a tradeoff. Our goal is maximum accuracy asap, but if you overcorrect for that right now, chats tend to get pretty muted. Imagine you asked a child to sing a song. Now imagine you muted every part that wasn’t perfect pitch. Which is better?

— Mike Davidson (@mikeindustries) February 25, 2023

Microsoft, then, is making a choice—and you’ll have to make one, too. If you want to use Bing Chat in its role as a search assistant, select the “precise” option. If you value more creativity and don’t care so much whether the topics Bing brings up are totally accurate, select the “creative” option. Perhaps in the future the twain shall meet.

Update the detailed information about Openai’s Chatgpt Update Brings Improved Accuracy on the website. We hope the article's content will meet your needs, and we will regularly update the information to provide you with the fastest and most accurate information. Have a great day!