Connect with us

AI Insights

DeepSeek Releases New Version of Model Behind Its AI Chatbot

Published

on


Chinese artificial intelligence startup DeepSeek has released a new version of the open-source AI model behind its controversial chatbot.

Citing a post on DeepSeek’s official WeChat group, Bloomberg reported that DeepSeek V3.1 is ready for testing.

The new version has a longer context window, or space for prompting, of 128,000 tokens. That’s roughly 96,000 words or about two 200-page English novels.

DeepSeek’s V3 model caused a stir in January when the startup claimed it only cost $5.6 million to train using about 2,000 of slower Nvidia chips.

That’s far cheaper than the millions it took to train frontier models from OpenAI, Google, Anthropic and others. The news wiped $600 billion of market value from Nvidia in one day. But governments soon banned the use of the DeepSeek chatbot out of concerns the data would be kept on Chinese servers.

While the startup didn’t share much more on WeChat, a post on Reddit said the latest version of the chatbot is “very, very verbose,” and also observed that the “r1 in the think button” has disappeared, indicating V3.1 could be a mixed reasoning model.

R1 is a reasoning model that DeepSeek also developed. It is offered through the three major U.S. hyperscalers AWS, Microsoft Azure and Google Cloud. The cloud providers have said the model is hosted locally so data would not be sent to China.

Developers are still waiting for R2, the next model release of R1, according to Bloomberg.

Read also: Remember DeepSeek? Many Adopt Its AI Models Despite Security Concerns

In the global AI race, only China is able to compete effectively with the U.S., Bloomberg reported. Chinese companies such as Alibaba, DeepSeek and Moonshot have developed AI models that have capabilities approaching the best ones in the U.S.

While the U.S. has banked on largely closed, proprietary AI models, China has pushed open-source models that generally are free to download and use. China is sacrificing short-term profits to ensure Chinese AI is adopted globally, according to Bloomberg. China’s 14th five-year blueprint for development, released in 2020, favored the open-source approach. Some Chinese artificial intelligence startup managers also believe the fastest way to enter new markets and compete with U.S. models is to offer open AI models.

Read more: DeepSeek Upgrades AI Reasoning Model to Rival OpenAI and Google



Source link

AI Insights

“A pastor should never use artificial intelligence to write a sermon”, Evangelical Focus

Published

on


The Graham family were very present at the Fourth European Congress of Evangelism in Berlin, where over a thousand evangelical leaders from more than 55 countries came together for four days of training, encouragement and challenges towards evangelistic mission in Europe.

Will Graham, son of Franklin Graham and grandson of Billy Graham, follows in family footsteps in answering the call to evangelistic preaching throug the Billy Graham Evangelistic Association (BGEA) .

In Berlin, he briefly spoke with Spanish news website Protestante Digital about his overall vision for Europe, and more specifically for Spain, where churches in Madrid are working with the BGEA to organise the Festival of Hope, due to take place in May next year.

“The need for bold, biblical proclamation evangelism is the same today—if not greater—than it was in my grandfather’s day”, he told a few weeks before the event.

During the convesation in Berlin, he addressed one of the issues that often came up in informal conversations among congress participants: the impact of technology, and AI in particular, on the development of the ministry.

 

Question. You have had several evangelistic experiences in Europe. Could you tell us about them?

Answer. I wouldn’t say there have been many. Most of my ministry has been in the UK, preaching in different churches. When it comes to evangelistic crusades or city-wide outreaches, I have only been part of a few: one in Thurrock, two in Scotland, and then Porto, which was an unforgettable experience. I have preached in other places too, but not in large-scale evangelistic events.

 

Q. Do you see a difference between the way Europeans respond compared to Africans or Asians?

A. Honestly, I see people from all walks of life. In Europe especially, you get a real mix of backgrounds. It really feels like the center of the world — close to Africa, near Asia, and even Americans are only a few hours away. So at the events, you meet a very diverse group of people. In Africa, we have also seen large crowds and God working powerfully. Wherever it is, it is always amazing to see who God brings.

Will Graham preached on the third day of the Fourth European Congress on Evangelism./ BGEA

 


Q. You mentioned  your grandfather Billy Graham at the press conference prior to the Congress. How has he influenced your preaching and style?

A. There are similarities, of course. I never tried to copy him, but being his grandson, growing up in the same part of the U.S., even sharing the same accent, people sometimes say I sound like him.

Still, I have my own way of preaching. I usually follow a biblical story and develop the message from there, whereas he often preached topically, although he loved stories too, like the Prodigal Son, one of his favorites and also one of mine.

But in the end, the goal is the same: to preach the gospel clearly and invite people to make a decision for Christ.

 

Q. Technology, artificial intelligence, social media… all of this is transforming the way we live. Do you think they are also changing the way mission is done today?

A. Technology has been a huge help in spreading the gospel, but it can also be a hindrance.

Take AI, for example: there are good things it can do, and also risks. A pastor should never use AI to write a sermon, that has to be between him and God. Preparing a sermon means wrestling with God’s Word, asking: “What do You want me to say to Your people?” AI cannot replace that.

That said, there are good uses for AI and other technologies in ministry. The key is to use them wisely, not letting them replace the message of the gospel.

At the end of the day, AI is made by people, and there are always human agendas behind it. So we have to be careful, but it can be useful in certain contexts.

“A pastor should never use artificial intelligence to write a sermon”

Praying time the Fourth European Congress on Evangelism./ BGEA

 

 

Q. Next year, the BGEA will hold the Festival of Hope in Madrid. What are your expectations for Spain?

A. I have never been to Spain, so I do not have direct experience yet. But Spanish people are always lively and welcoming. I am very excited about Madrid.

I hope to be there with my father, not to preach, but to see what God will do. And maybe in the future God will open doors in other Spanish cities, whether for him or for me.

Spain is special because it is one of the countries my grandfather Billy Graham never visited. He once had an invitation, but it fell through. This will actually be my father’s second time in Spain, which is historic. There is great expectation about what God will do.

Plus, I will get to practice my Spanish, though I feel sorry for the people who will have to figure out what I am really trying to say.

Published in: Evangelical Focuseurope
– “A pastor should never use artificial intelligence to write a sermon”



Source link

Continue Reading

AI Insights

A Software Engineer Explains Why He’s Not Afraid AI Will Take His Job

Published

on


Doug Steinberg, 46, is a software engineer who lives in Coral Springs, Florida. He isn’t worried that artificial intelligence will steal his job because he’s able to use it to be far more productive. The following has been edited for brevity and clarity.

In the last year or so, I’ve been using AI-assisted coding, and I’ve gotten really used to it. I don’t want to go back to the old way or be at a company that doesn’t think that you should be using it.

AI takes a lot of the load off of you. As an example, one thing that every developer has to do is make commit messages. You write some code, you do your work, and you save it, basically, to a checkpoint. You write a message about what you’re saving. It seems no one has ever really put a lot of effort into that. Now, you just say, “Hey, write me a good message,” and AI writes incredible, detailed comments.

In the past, you’d always see messages like “work in progress” or totally unusable stuff. Even if it was kind of useful, AI adds a lot more detail. Now, I never have to think about what I have to write for that. That’s only one tiny slice of the whole big picture with AI.

AI will make you better

Using AI, it almost feels like you have another person with you at all times that you can ask anything you need to or try to brainstorm with. Before, it was all on you. It’s crazy to me to think about going a day without using it, even though it’s only been a year or two since this came out.

I can write an entire app now in days, where before it would have taken me months. For example, my wife and I have a side business developing software for sales agencies in the lighting industry. We’re not using AI for it yet, but we plan to.

I started writing an app for our company back in 2023. It took me months. There’s a similar app, complexity-wise, that I made more recently for our daughter’s school to track volunteer hours, and, using AI, I made the thing in a few days. My wife says it’s like changing out your shovel for an excavator.

What really turned my thoughts about AI was when I discovered Claude Code. It seemed to do at least 5x better than anything I’d seen before. The first website I vibe-coded that actually worked out was for my wife’s company for a conference. In the past, they would just make a simple site to let people know about the event.

This time, I ended up making this whole conference platform where guests could register and get notified, manufacturers could register, and manufacturers could scan guests’ name tags to see who visited their booth.

Without AI, I would have never invested the amount of time it would take to build a thing like that. It would have easily been a couple of weeks. In this case, I got it up and running in just a few days, and looking really nice.

AI can do visual design way better than I can. That’s not something I’m good at, and it can make things look pretty decent. If I were good at design, it could probably make things look amazing. I’m good at software development, but AI will make you better at what you’re good at.

Amplifying productivity

Software engineers work for companies that build software, and their whole thing is they want to keep on putting out features as fast as they can and doing things for their customers. I just see AI amplifying it. I see there being more features — faster and better.

Ultimately, the customers are going to get more out of whatever tools they have. In some cases, maybe they’ll think they don’t need some software products anymore because they can just do it themselves. Honestly, I still think it’s going to be a net-positive. As a maker, I’m not afraid of people like me being negatively impacted.

When it comes to my career, I think the only reason I’m any good is because I had to go through all that pain to learn how things actually work. If I didn’t have to do that, I don’t know if I would have had the discipline to actually learn it. If you don’t really know how things work, you’re not going to get that far. It’s going to be a house of cards that will eventually fall over.

It’s possible that companies won’t need as many people, but I’m not that worried about it because I can make things and could make money on my own. I’ve always wanted to make my own apps, but I’ve never had enough time or been fast enough to do it.

Now, it’s almost like you have another you. You can give it a prompt, walk away, and come back later, and this whole feature is built. There’s just never been anything like it.

Do you have a story to share about your career? Contact this reporter at tparadis@businessinsider.com.





Source link

Continue Reading

AI Insights

Purple Lexity, an American artificial intelligence (AI) search company, cited Mail Business as an ex..

Published

on


Exploration loop case applied to the Mail Business homepage. Under the body of the main article, AI proposed follow-up questions, which were designed to keep readers curious. Capture the Mail Business Home Page

Purple Lexity, an American artificial intelligence (AI) search company, cited Mail Business as an excellent example of media innovation and improved independent experiences. Purple Lexity recently introduced the number of searches and users in just a few months after Maekyung introduced its search API “Sona” through an introduction to excellent collaboration cases on its official website.

Specifically, in a post titled “How Maeil Business Newspaper provides readers with in-depth search functions in the site with Sona API,” Purple Lexity evaluated that Maekyung has built an AI-based search function to increase readers’ accessibility and utilization at the same time.

In particular, the key strength was that the source of the original article was clearly informed of the answer through a question-oriented search in the form of a sentence rather than a simple keyword match.

Maekyung has been pushing for search innovation since last year to find the answer that readers want easily and accurately in more than 600 articles a day. Existing keyword searches were easy to miss context, and early AI solutions were less reliable due to hallucination problems. In response, Maekyung worked with Purple Lexity, which specializes in AI search, to explore the entire archive of articles that Maekyung has accumulated over the past 60 years when readers enter questions, presenting answers that fit the intention of the question and marking the sources of related articles together.

Purple Lexity introduced "Maegyeong AI Knowledge Search" on its website as an excellent example of collaboration. Purple Lexity
Purple Lexity introduced “Maegyeong AI Knowledge Search” on its website as an excellent example of collaboration. Purple Lexity

Purple Lexity introduced Maekyung’s principle that “independent trust is secured when the generated answers are accurately linked to the source of internal articles.” It explained that it provided immediate services to questions that required extensive information processing and built services so that readers with various expertise could understand them, improving the speed and quality of information access at the same time. The performance before and after the introduction of AI search was confirmed by figures. As of July this year, the total search volume on the website of Maekyung surged 64.4% compared to the beginning of the year, and the number of search users increased 25.0% during the same period. In particular, AI knowledge search in Maegyeong accounted for about half of the total search volume as of July.

There was also a clear change in the usage behavior of readers every day. “AI presents related questions in addition to answering questions in the AI knowledge search window, and the proportion of clicking on them accounts for about a quarter of all searches,” said an official at MaekyungAX. “Readers have begun to actively utilize the follow-up search path proposed by AI beyond simply reading articles.” An official from Maekyung also said, “Readers are using it not only as a context-based search but also as a tool for financial technology and life-oriented analysis while obtaining detailed and reliable answers even with simple questions.”

The actual Maekyung site function was also reorganized accordingly. Reading the body of a paper article creates a Chain of Search that automatically generates relevant recommendation questions to continue curiosity, and an AI search box is placed separately in each article to make it easier for readers to ask customized additional questions. In addition, Purple Lexity emphasized that this case is a model of how the collaboration between media companies and AI companies innovates their own experiences.

[Reporter Ko Min-seo]



Source link

Continue Reading

Trending