About Grok 4: Elon Musk's New Chatbot That's "Smarter Than All Graduates"
The new model came over 24 hours after xAI was forced to remove inappropriate Grok 3 posts from X that included antisemitic comments and replies to users. Grok 3 came under scrutiny after it seemingly went off the rails following an update over the weekend that instructed the chatbot to "not shy away from making claims which are politically incorrect, as long as they are well substantiated."
"We need to make sure that the AI is a good AI," Musk said at the launch of Grok 4 on Wednesday, without acknowledging the offending Grok 3 posts or controversy. The xAI team introduced the new chatbot during a livestream on X, calling the release an era of "Big Bang Intelligence."
Introducing Grok 4, the world's most powerful AI model. Watch the livestream now: https://t.co/59iDX5s2ck
— xAI (@xai) July 10, 2025
Grok 4 is "smarter than almost all graduate students, in all disciplines, simultaneously," according to Musk. He said he team has "run out of test questions to ask...Reality is the ultimate reasoning test."
About Grok 4
The new AI chatbot includes improved voice conversations, with the xAI team touting benchmarks showing the new AI system scoring higher than OpenAI and others.
"At times it may lack common sense, and it has not yet invented new technologies or discovered new physics, but that is just a matter of time," Musk said.
Grok 4 is trained on xAI's Colossus supercomputer and promises real-time web access through its DeepSearch tool and advanced scientist-level reasoning. It's a coding-focused variant and aims to debug, write, and explain code with greater accuracy, like OpenAI's Code Interpreter or GitHub's Copilot.
"You can cut and paste your entire source code file into the query entry box on http://grok.com and @Grok 4 will fix it for you! This is what everyone @xAI does. Works better than Cursor," Musk said in a post on X.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Indian Express
23 minutes ago
- Indian Express
What is Comet, Perplexity's agentic AI-powered browser?
Perplexity, known for its AI-powered search engine, earlier this month introduced its new AI-first browser, Comet, which is designed to help users assign various web tasks to AI agents. The browser is still in early access under the Perplexity Max plan, which costs $200 a month. From user responses and early reviews, Comet seems to be offering a glimpse into the future where web browsing will be a completely different experience. It will be like users delegating a task to an agent who would browse for them to accomplish it. In essence, a user will not be directly interacting with websites, but an AI agent will be doing it in the browser for them. How is Comet different from just using the Perplexity browser? Why does the Aravind Srinivas-led AI company need to build an entirely new web browser? What can Comet do? Comet looks a lot like Google Chrome, and there is a good reason for that. The browser is built on the Chromium framework, which is an open-source architecture framework for web browsers maintained by Google. This framework is also the backbone of popular browsers such as Chrome, Microsoft Edge, and Opera. Comet would support Chrome extensions, bookmarks, and even sync with a user's settings if they import them. The AI-first browser is built from the ground up with Perplexity AI agents that are integrated directly into the browser. It not only helps users with search but also acts on their behalf across tabs, services, and platforms — this is why Perplexity describes Comet as an 'agentic browser'. Its search capabilities come from Perplexity's answer engine that is based on a mixture of foundational LLMs like GPT-4o and Claude 4.0 Sonnet, along with its proprietary model Sonar. Perplexity AI was launched a month after OpenAI's ChatGPT created ripples across the world. The AI-powered search engine rapidly grew in popularity, which stems from its unique positioning as an 'answer engine', which is radically different from a traditional search engine or AI chatbots. Perplexity was among the first to combine AI with traditional web search with accurate and real-time answers to any question. However, Comet is local and standalone in what seems like platform independence. In case Perplexity chose to build on top of Safari or Chrome, then it would likely be second to Google or Apple's native AI tools. As an independent browser, Comet gives Perplexity full control. When it comes to local context access, cloud-based browser agents would require one to log in and start from scratch. But, with Comet, the agent already knows what tab a user is viewing and what they are doing. This eliminates the need for copying, pasting, or reloading. Since it's local, the agent can interact with open pages and logged-in services instantly without additional authentication. The futuristic browser comes with some powerful features that make web browsing efficient and more productive. With Comet, you browse less and prompt more, and the AI takes care of the rest. It comes with a plethora of use cases that could likely make it a daily essential, much like what Google Chrome is to millions of users. From scheduling meetings from Gmail to comparing products across tabs, Comet is more than a browser. This is possible because it can read an email, find a suitable time for a user's calendar, draft a reply and create a tentative event with a Google Meet link. And all of these from within Gmail, without switching tabs. Similarly, if a user is shopping for a camera online, they can ask Comet to summarise and compare specs like recording quality or frame rates. If they have multiple tabs open, Comet will pull context from all these open tabs and offer a clean, structured summary to help them make an informed decision. One of the most useful use cases is the ability to summarise videos or articles. Comet can reportedly summarise an open YouTube video or web article in seconds. The video summaries are generated using transcripts and content on web pages. Some users with early access have also revealed that Comet can also accept LinkedIn requests. It can review pending requests and bulk-accept the ones it deems most relevant, saving users time. Comet's built-in assistant is placed at the top-right corner. The assistant can help in summarising the page a user is on; perform actions like clicking, filling forms or navigating; look up things across open tabs, and act simultaneously across sites. Users can even @mention tabs by name for specific tasks at open pages. For example, @YouTube – find the top comment. It needs to be noted that Comet will ask a user for confirmation before posting and notify them of the outcome. However, in some cases limitations imposed by platforms may prevent Comet from posting or engaging with comments. After years of stagnant innovation in web browsing, Comet seems to be signalling a huge shift in how the world will use the internet. Users will no longer be required to go on manually clicking through dozens of tabs, as they can now delegate tasks to agents that will understand context, objectives, and user preferences. Web browsing backed by AI agents is crucial in times when there is an overload of AI-generated content. Browsers like Comet could help filter through the clutter and deliver accurate information faster. The biggest catch with Comet for now is its availability, as it is a part of Perplexity Max, which is priced at $200 a month. Over time the company may lower the barrier or release a limited free tier if rivals OpenAI or Google move quickly with their own agentic browsers. Perplexity will roll out invite-only access slowly to its waitlist over the summer, and new users will receive a limited number of invites to share. Bijin Jose, an Assistant Editor at Indian Express Online in New Delhi, is a technology journalist with a portfolio spanning various prestigious publications. Starting as a citizen journalist with The Times of India in 2013, he transitioned through roles at India Today Digital and The Economic Times, before finding his niche at The Indian Express. With a BA in English from Maharaja Sayajirao University, Vadodara, and an MA in English Literature, Bijin's expertise extends from crime reporting to cultural features. With a keen interest in closely covering developments in artificial intelligence, Bijin provides nuanced perspectives on its implications for society and beyond. ... Read More
&w=3840&q=100)
First Post
23 minutes ago
- First Post
Trump explored ways to cut contracts with Elon Musk's SpaceX after fallout but failed: Report
Following the fallout between Elon Musk and US President Donald Trump, the administration explored the possibility of cutting contracts awarded to Elon Musk's SpaceX, but concluded that it was not possible as SpaceX was involved in critical work for the Department of Defense and Nasa and there were few to no alternatives, according to a report. read more US President Donald Trump talks to the media, next to Tesla CEO Elon Musk with his son X Æ A-12, at the White House in Washington, DC, US, March 11, 2025. (Photo: Reuters) After the public break-up of Elon Musk and US President Donald Trump, the administration explored ways to cut contracts to Musk-owned SpaceX, but concluded that it was not possible, according to a report. After Musk mounted personal attacks on Trump over the 'One Big Beautiful Bill', Trump had suggested he could cut subsidies to Musk's companies like SpaceX and Tesla. Over the years, the US government has given several billions in contracts, subsidies, and tax credits to Musk's companies — most notably SpaceX. STORY CONTINUES BELOW THIS AD Days after Trump's statement, a senior administration initiated a review of government contracts to SpaceX to see if and how they could be cancelled, as per the Wall Street Journal. However, the review concluded that the Department of Defense and Nasa are so deeply dependent on SpaceX for a host of purposes, ranging from military connectivity to satellite launches, that contracts to SpaceX could not be cancelled, the newspaper reported. As per the latest information available publicly, ongoing US government contracts to SpaceX total around $18 billion. This does not include the money that SpaceX or other companies owned by Musk have got in subsidies or tax credits in ongoing or past contracts. Public estimates say that around 30-40 per cent of SpaceX's annual revenue comes from the US government — mostly from services to Nasa, military, and intelligence services. Trump administration realises US depends too much on Musk The Journal reported that the review of Musk's contracts, which was led by Josh Gruenbaum of the General Services Administration's (GSA), asked the Defense Department, Nasa, and around half a dozen other federal agencies to fill out a spreadsheet with all of their current SpaceX contracts and other transaction agreements, as per the newspaper. The spreadsheet also asked whether a competitor could do the job that had been contracted to the SpaceX, the report said. After receiving the information, the Trump administration realised it could no cut contracts to SpaceX for two reasons. Firstly, SpaceX was doing some of the most critical work for the US military, Nasa, and spy agencies like the National Reconnaissance Office (NRO). Such work cannot be disrupted or stopped. Secondly, even if the Trump administration would have gone for short-term disruption for long-term solution or punishment to Trump, they would have found few alternatives. In most of the cases, the technology and launch capacity of SpaceX is simply unmatched. There are few to no companies that can do the work that SpaceX does — particularly at the scale of SpaceX. STORY CONTINUES BELOW THIS AD
Time of India
26 minutes ago
- Time of India
OpenAI, Oracle deepen AI data center push with 4.5 gigawatt Stargate expansion
Synopsis OpenAI announced that Oracle will provide 2 million chips to help scale its AI data center infrastructure. This partnership aims to boost computing power for OpenAI's advanced models like ChatGPT, reflecting growing demand for AI capabilities and the need for massive processing resources to support training and deployment.
