Connect with us

Apps

How Auto-GPT Outperforms ChatGPT: A Breakthrough AI Tool

Published

on

You may be one of the many people who are amazed by the potential of ChatGPT and other large language models (LLMs) such as the new Bing or Google’s Bard. If you haven’t heard of them yet – which is unlikely since ChatGPT is said to be the fastest-growing app ever – here’s a brief overview:

LLMs are software algorithms that learn from massive text datasets, allowing them to comprehend and produce human language in a realistic way. The most famous example is ChatGPT, a chatbot interface that uses the GPT-4 LLM to create stunning conversations. ChatGPT can talk like a human and generate anything from blog posts, letters, and emails to fiction, poetry, and even computer code.

LLMs are impressive, but they have one major limitation. They can only do one thing at a time, such as answering a question or generating a text, before needing more human input (called “prompts”). This means they are not very good at more complex tasks that require multiple steps or depend on external factors.

That’s where Auto-GPT comes in – a technology that tries to solve this problem with a simple approach. Some think it could be the next step towards the “holy grail” of AI – the development of general, or strong, AI. Current AI applications are usually made to do one thing, getting better at it as they get more data. Some examples are image analysis, language translation, or self-driving cars. Because of this, they are sometimes called “specialized AI,” “narrow AI,” or “weak AI.”

A general AI is one that can theoretically do many different kinds of tasks, even ones it wasn’t built to do, much like a naturally intelligent being (such as a human) can. Sometimes it’s called “strong AI” or “artificial general intelligence” (AGI). AGI is maybe what we imagined when we thought of what AI would be like before machine learning and deep learning made weak/narrow AI a common reality around the beginning of the last decade. Think of the science fiction AI shown by robots like Data in Star Trek that can do almost anything a human can do.

So what is Auto-GPT?
The easiest way to understand it is that Auto-GPT can do more complicated, multi-step tasks than existing LLM-powered applications by making its own prompts and using them on itself, creating a loop.

You can think of it like this: To get the best answers from an application like ChatGPT, you need to think carefully about how you ask the questions. So why not let the application make the question itself? And while it does that, also let it ask what the next step is – and how it should do that … and so on, creating a loop until the task is done.

It works by splitting a big task into smaller sub-tasks and then creating separate Auto-GPT instances to work on them. The original instance acts like a kind of “project manager,” coordinating all of the work done and putting it together into a final result.

Besides using GPT-4 to make sentences and prose based on the text it has learned, Auto-GPT can also search the internet and use the information it finds there in its calculations and output. In this way, it’s more like the new GPT-4 enabled version of Microsoft’s Bing search engine. It also has a better memory than ChatGPT, so it can make and remember longer chains of commands.

Auto-GPT is an open-source application that uses GPT-4 and was made by one person, Toran Bruce Richards. Richards said that he wanted to make it because traditional AI models “, while powerful, often have trouble with tasks that need long-term planning, or are not able to improve their methods based on real-time feedback.”

It is one of a group of applications that are called recursive AI agents because they can use the results they make to create new prompts, linking these operations together to finish complex tasks.

Some applications of Auto-GPT and AI agents are:

  • Software development: While apps like ChatGPT are good at generating code, they are usually limited to short and simple programming and software design. Auto-GPT, and maybe other AI agents that work like it, can make software applications from beginning to end.
  • Business optimization: Auto-GPT can also help businesses to increase their net worth by themselves by looking at their processes and giving smart suggestions and insights about how they can be better.
  • Internet research: Unlike ChatGPT it can also go online, meaning you can ask it to do market research or other similar tasks – for example, “find me the best smartphones for under AED 1500.”
  • Humanity destruction: One very disruptive task it has been given is to “destroy humanity” – and the first sub-task it made for itself to do this was to start researching the most powerful atomic weapons ever. Since its output is still only making text, its maker says it won’t actually do much with this task – hopefully.
  • Self-improvement: Auto-GPT can also use itself to make itself better – its maker says it can make, check, review, and test changes to its own code that can make it more able and efficient.

It can even make better LLMs that could be the basis of future AI agents, by speeding up the model-making process.

What could this mean for the future of AI?
Ever since generative AI applications came out, it’s been clear that we’re only at the beginning of a very long journey, in terms of how AI will change and affect our lives and society.

Are Auto-GPT and other agents that use the same methods the next step of that journey? It seems likely. At the very least, we can expect AI tools that let us do much more complex tasks than the relatively simple things that ChatGPT can do, to start to become common.

Soon, we will see more creative, advanced, varied, and useful AI output than the simple text and pictures that we’ve become used to. These will no doubt have an even bigger impact on the way we work, play and communicate.

Other possible positive impacts include lower cost and environmental impact of making LLMs (and another machine learning-related activity) as autonomous, recursive AI agents to find ways to make the process more efficient.

However, we also have to think that by itself it doesn’t really fix any of the problems with generative AI. These include the unreliable (to say it nicely) accuracy of the output it makes, the potential for misuse of intellectual property rights, and the chance of it being used to spread biased or harmful content. In fact, by making and running many more AI processes to achieve bigger tasks, it could possibly make these issues worse.

 

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Apps

Saudi Pro League Partners with TikTok

Published

on

The Saudi Pro League (SPL) announced TikTok as its ‘Digital Partner’ for the 2024-25 season. In this official partnership, TikTok will introduce a dedicated SPL Exclusive Hub, a central space where fans can access content from the league’s clubs and partners. This hub offers an immersive experience, bringing fans closer to their favourite teams and players in innovative and interactive ways. This collaboration sets new standards for sports entertainment by leveraging TikTok’s creative community and dynamic platform to bring the SPL closer to its expanding fanbase.

Through this partnership, TikTok will elevate the excitement of Roshn Saudi League (RSL) matches like never before. SPL will offer exclusive content available only on TikTok, giving fans unique insights into the league, its clubs, and players. From behind-the-scenes footage to in-depth player interviews, the content will create an engaging and immersive experience for fans.

Additionally, the SPL will provide TikTok creators with exclusive opportunities to explore and showcase new facets of the league, such as access to training sessions, match-day experiences, and interactions with top players. These opportunities will not only enhance engagement but also drive creative storytelling that resonates with a global audience. With SPL’s innovative experiences and thrilling football content, TikTok football fans can look forward to a diverse range of content and engagement opportunities, all exclusively on TikTok. With more than of 44.2 million likes and a follower count of 2.5 million, the SPL’s presence on TikTok has created a centralized digital space for fans to engage with the league like never before.

Omar Mugharbel, CEO of the Saudi Pro League, said, “The partnership between the Saudi Pro League and TikTok is a groundbreaking development for our sector in the region and represents a strategic leap into the future of sports entertainment. This is more than a partnership, it’s a digital revolution. We are harnessing TikTok’s power to transcend traditional engagement and offer our fans a canvas for creative expression and a suite of experiences tailored to their passion for football. We’re creating a realm where each fan has a front-row seat to the spectacle of the Roshn Saudi League, powered by the boundless creativity that TikTok embodies. The SPL Exclusive Hub on TikTok provides an immersive and interactive digital home for our passionate football fans. We are shaping the future of fan engagement and setting a global benchmark right here in the Saudi Pro League.”

Kinda Ibrahim, TikTok’s GM for Operations in the Middle East, Turkey, Africa, Pakistan and South Asia, said, “Saudi Arabia’s passion for football runs deep in its history. By partnering with the Saudi Pro League, we are tapping into this strong affection for the sport and aligning with the league’s strategic vision to become one of the top football leagues globally. Through modernising the fan experience with a dedicated space for league content, clubs, and partners, we aim to strengthen the community of sports enthusiasts by providing them with more ways to connect and engage with exclusive content online.”

Continue Reading

Apps

Age of Empires Mobile to Release Around the World on October 17, 2024

Published

on

Mobile gaming fans around the world can prepare for a brand-new take on a strategy classic as Age of Empires Mobile will officially launch on the Apple App Store and Google Play on October 17. Co-developed by TiMi Studio Group (Call of Duty: Mobile, Pokémon Unite) and World’s Edge, an Xbox Games Studio, Age of Empires Mobile is a fresh and innovative experience for the beloved series, bringing familiar elements from the iconic Age of Empires games with brand-new and mobile-specific gameplay to mobile players.

On the road to global launch, players will have the opportunity to participate in on-site events at the Xbox booth, and the stage event with fun engagement activities at the Level Infinite booth, during Gamescom in Cologne, Germany. Media and creators will be invited to preview the game in the business area. Players can pre-register now for the official launch on the App Store and Google Play, join the game’s official Discord or head to the game’s website for more information.

“We expect players to mystify, mislead and surprise their enemies through the realistic, immersive, medieval gameplay featured in Age of Empires Mobile,” said TiMi Studios team leader and studio general manager Brayden Fan. “We’ve already seen alliances form and alliances crumble through our multi-month playtesting. Through it all, players have shown strategic wit, savvy teamwork and a passion for an entirely new Age of Empires experience.”

Built from the ground up for mobile and social play, Age of Empires Mobile includes a variety of game features:

  1. Multiple single-player modes built on the Age of Empires lineage and featuring iconic elements from the original series
  2. Build your empire in a lively, realistic medieval world featuring stunning civilizations, imperial cities and historical figures
  3. Immersive battlefields offer one-of-a-kind original gameplay with large-scale castle sieges featuring realistic weapons taking on multi-dimensional defences in a fight with thousands of players worldwide
  4. Be strategic as you encounter realistic and interactive terrain and weather and manage multiple troops with real-time micro-control

Built to expand on the series’ lineage, Age of Empires Mobile offers players the opportunity to select from some of the world’s most legendary historical figures to lead their armies. Players choose from dozens of these leaders, such as Barbarossa, Darius the Great, Hammurabi, Joan of Arc and Leonidas I, each featuring unique individual talents and synergies with each other to unleash on your foes.

“Our goal at World’s Edge is to bring the magic of Age of Empires to as many different types of players around the world as possible. Partnering with TiMi Studio Group means we can combine their expertise in mobile gaming with the strength of our franchise and history – to bring a brand-new take on Age of Empires to even more gamers, in even more places,” said Michael Mann, Studio Head at World’s Edge.

Continue Reading

Apps

WhatsApp to Intro Real-Time Voice Conversations with Meta AI

Published

on

WhatsApp is set to revolutionize user interaction with the introduction of real-time voice conversations powered by Meta AI. This new feature, currently in development, will allow users to engage in natural, spoken dialogue with the AI chatbot.

Key Features:

  1. Personalized Experience: Users can select from various voice options for Meta AI, tailoring the interaction to their preference.
  2. User Control: The voice chat mode will be optional, ensuring users have complete control over their conversations.
  3. Privacy Focus: WhatsApp is implementing a privacy indicator to inform users when their microphone is active, safeguarding user data.
  4. Accessibility: Captions and transcripts will be available, enhancing accessibility and providing a visual record of conversations.
  5. Customizable Responses: Users can choose between brief or full responses from Meta AI, adapting the AI’s output to their needs.

This exciting development promises to make AI interactions more engaging, accessible, and personalized for WhatsApp users.

Continue Reading
Advertisement

Latest Reviews

Follow us on Facebook