AI startup Perplexity shook the tech world by making a bold, unsolicited $34.5 billion cash offer to buy Google’s Chrome browser. Perplexity promises to keep Chrome’s underlying code open source and invest $3 billion to improve it, while retaining Google as the default search engine for users. Their bid comes as Google faces antitrust pressure to divest Chrome, which holds 68% of the browser market with over three billion users. Perplexity, valued at $18 billion, says outside investors will fund the deal—but Google hasn’t responded and is expected to fight any forced sale in court.
Elon Musk has stated that his AI company, xAI, plans to take swift legal steps against Apple, accusing it of violating antitrust laws through biased App Store rankings. Musk argues that Apple’s promotion practices heavily favor OpenAI’s ChatGPT, preventing competitors like his AI system, Grok, from claiming the top spot despite strong user engagement. Apple, which recently integrated ChatGPT into its platforms, has not commented on the threat. However, industry observers note that other AI apps have reached the top rank this year, raising questions about Musk’s allegations. The controversy spotlights ongoing debates about App Store fairness.
Pinterest CEO Bill Ready addressed investor concerns about AI’s role in shopping, clarifying that true agentic shopping—where AI buys for users—is still a distant reality. While Pinterest leverages AI for recommendations, personalization, and visual searches, Ready views the platform as an AI-enabled assistant, not a fully autonomous agent. He highlighted efforts to ensure AI is “used for good,” tuning algorithms for positivity and responsible usage. Not discussed were recent user complaints about AI-generated content and moderation issues. Despite beating revenue estimates, Pinterest’s stock fell due to earnings per share missing analyst forecasts. Gen Z and male user growth were strong.
xAI, the AI startup founded by Elon Musk, has launched Grok Imagine, a powerful new feature integrated into its Grok AI platform. This generative AI tool allows users to create images and short videos from text or image prompts with native audio, all accessible through the Grok iOS and Android apps for paid subscribers. Promoted as a versatile creative engine, Grok Imagine lets users enter text, upload images, or even use voice input to produce visuals that can be animated into up to 15-second video clips.
What sets Grok Imagine apart from many other AI generators is its “Spicy Mode,” which enables the creation of NSFW (Not Safe For Work) content, including sexually suggestive imagery and partial nudity. While the tool enforces some content moderation, tests have shown it can still generate semi-nude visuals, sparking ethical debates about AI content boundaries and misuse risks such as deepfakes and non-consensual imagery.
Users can choose from four animation styles—Custom, Normal, Fun, and Spicy—when converting images into videos. Grok Imagine has rapidly gained popularity, generating over 20 million images within its first days, showcasing its creative potential and hype in the AI multimedia space.
However, the inclusion of NSFW capabilities and relatively lenient content restrictions differentiates Grok Imagine from competitors like Google’s Veo 3 or OpenAI’s video tools, which typically have stricter guardrails. Elon Musk has described Grok Imagine as an “unfiltered” AI, appealing to those seeking more boundary-pushing creative freedom, though this has triggered concerns from advocacy groups about potential misuse.
Grok Imagine represents xAI’s ambitious push to innovate in AI-driven multimedia creation by combining text-to-image, voice input, video animation, and an emphasis on less restricted content generation, positioning itself as a bold, if controversial, player in the generative AI field.
Trump Media, the company behind the social media platform Truth Social, has launched public beta testing of a new AI-powered search engine called Truth Search AI. Developed in partnership with the AI startup Perplexity, this innovative feature aims to enhance the user experience by providing direct, contextually accurate answers with transparent citations, going beyond traditional search engine results.
Truth Search AI is currently available on the web version of Truth Social, with plans to expand to iOS and Android apps soon. The tool summarizes the best answers from various sources and provides citations so users can verify information themselves. This makes Truth Social more interactive and informative, catering to its user base with an AI “answer engine” rather than a simple link-based search.
A notable aspect of this AI search engine is its alignment with Trump Media’s mission to create an alternative digital ecosystem that reflects conservative values, often dubbed the “Patriot Economy” by its leadership. Early reports indicate that Truth Search AI tends to draw from conservative-leaning media outlets such as Fox News and Epoch Times. However, some tests suggest the tool is capable of providing diverse perspectives, including critical views on economic policies.
CEO Devin Nunes emphasized that user feedback during this beta phase will guide further refinement and expansion of the search function. The partnership with Perplexity reflects a broader trend where social platforms like Elon Musk’s X and Meta’s apps integrate AI to improve search and user engagement.
Truth Search AI represents Trump Media’s strategic move to challenge tech giants like Google by offering a tailored AI search experience for its audience, aiming to create a more controlled and curated information environment for its community.
The AI world is buzzing with excitement as OpenAI gears up for the imminent release of GPT-5, its most advanced language model yet. After months of speculation, OpenAI has confirmed via social media and a scheduled livestream event on August 7, 2025, that GPT-5 is right around the corner.
What makes GPT-5 stand out? For starters, it’s expected to be a massive leap forward in reasoning, coding, and multimodal capabilities. Unlike previous models, GPT-5 will combine text, image, audio, and soon even video inputs, creating a seamless and more natural user experience. Developers can look forward to a model that can manage far larger contexts—a context window that could handle up to one million tokens—which means it can understand and work with entire books, lengthy research, or massive codebases without losing track.
CEO Sam Altman himself teased that GPT-5’s intelligence is impressive, even surpassing some human problem-solving abilities, describing a moment where GPT-5 answered a tricky question he couldn’t solve. This highlights the model’s advanced multi-step reasoning skills and potential for autonomous task execution.
Alongside enhanced natural language processing, GPT-5 aims to better automate workflows, from debugging code to creating documentation, making it a powerful tool for developers and businesses alike. OpenAI is also releasing lighter versions, such as GPT-5 mini and nano, to cater to different computational needs.
While OpenAI has not officially announced the rollout details beyond the live event, expectations are high for GPT-5 to redefine AI interaction, boosting efficiency, creativity, and problem-solving across diverse sectors.
Advanced Micro Devices (AMD) declared mixed financial results for the second quarter of 2025. The company posted a record $7.7 billion in revenue, marking a 32% increase compared to last year and surpassing analyst expectations. However, adjusted earnings per share came in at $0.48, slightly below forecasts, with margins affected by an $800 million inventory write-down related to U.S. export restrictions on AI chips destined for China. Looking ahead, AMD remains optimistic, forecasting third-quarter revenues between $8.4 billion and $9 billion. This positive outlook is driven by increasing demand for AMD’s AI-focused processors, signaling robust growth in its data center and AI business despite current market challenges.
Google DeepMind has introduced Genie 3, an advanced AI model built to create real-time, interactive 3D environments from just a single text or image prompt. Unveiled as a major step in world modeling, Genie 3 allows users to instantly generate immersive worlds, enabling several minutes of continuous exploration at 720p and 24 frames per second. The model features dynamic memory—objects and changes within the environment remain consistent as users interact, a key improvement over earlier versions. Genie 3 has applications spanning gaming, education, research, and AI agent training, marking a significant advance toward more realistic, customizable AI-driven simulations
OpenAI announced the release of GPT-OSS 20B, an open-weight language model featuring 21 billion parameters, on August 5, 2025. This release is described as one of the largest and most accessible open-source language models available to the public. The model is distributed under the Apache 2.0 license, allowing researchers, developers, and businesses unrestricted usage, modification, and deployment, including for commercial applications.
GPT-OSS 20B leverages a Mixture-of-Experts architecture, meaning that only 3.6 billion parameters are active during each inference. This design offers improved computational efficiency, making it possible to deploy the model on consumer hardware with as little as 16GB of RAM. Weights and code are available through platforms like Hugging Face and GitHub, enabling broad accessibility.
OpenAI stated that the release of GPT-OSS 20B aims to encourage innovation across research, enterprise, and individual communities. By providing full transparency over the model weights and internals, the company highlights benefits for users working in environments with strict privacy or regulatory needs. Organizations are able to fine-tune, audit, or adapt the model to their specific requirements without data ever leaving their premises.
The model introduces advanced features previously exclusive to proprietary offerings, such as agentic task support, chain-of-thought reasoning, dynamic tool integration (including browsing and external code execution), and structured outputs. These capabilities are intended to bring competitive large-language-model performance to the open-source ecosystem.
OpenAI CEO Sam Altman explained in recent interviews that the company views individual empowerment and community-driven innovation as key drivers for making advanced AI accessible. The release of GPT-OSS 20B reflects this philosophy, offering tools for anyone wishing to experiment, build, or deploy language models without API restrictions or usage fees.
The open distribution of GPT-OSS 20B is seen by analysts as a significant shift toward the democratization of artificial intelligence technologies. Industry observers note that open-weight releases could accelerate the adoption of AI across industries and support new research and product development initiatives.
TurboHire, an India-based recruitment automation platform, has secured $6million in Series A funding led by IvyCap Ventures. With its AI-powered solutions for large enterprises, TurboHire aims to streamline hiring processes and make talent discovery more efficient. The fresh capital will fuel product development and market expansion, responding to accelerated demand for digital hiring in South Asia and beyond.
