Meta

Winsage
May 23, 2025
OpenAI has acquired the “io” hardware division from Jony Ive’s design studio, LoveFrom, investing [openai_gpt model="gpt-4o-mini" prompt="Summarize the content and extract only the fact described in the text bellow. The summary shall NOT include a title, introduction and conclusion. Text: This week has seen a fascinating interplay between two titans of artificial intelligence, OpenAI and Google, as they both made significant announcements that reveal much about their competitive dynamics. The backdrop of this rivalry is the remarkable rise of ChatGPT, which has kept Google’s leadership on high alert. As the tech community anticipated Google’s I/O event, there was palpable tension regarding whether Sam Altman of OpenAI would attempt to overshadow Google’s showcase, reminiscent of last year’s pre-I/O event where OpenAI unveiled ChatGPT’s advanced voice capabilities. This year, however, OpenAI chose to make its splash a day later, with news that has quickly shifted the narrative. OpenAI's acquisition of the “io” hardware division from Jony Ive’s design studio, LoveFrom, is a strategic move that some might view as a clever act of SEO sabotage. While the name “io” is said to stand for “input output” and was chosen well in advance, the timing of the announcement has undeniably redirected attention from Google’s strong presentation at I/O. The contrasting announcements highlight the strengths and weaknesses of both companies: while Google boasts technically superior models that are more widely utilized, OpenAI excels in capturing public interest and generating buzz. OpenAI's Strategic Moves Delving deeper into OpenAI’s recent announcement reveals a commitment to innovation and talent acquisition. The company is investing .5 billion in equity to onboard approximately 55 individuals from LoveFrom, including notable ex-Apple design leaders such as Evans Hankey, Tang Tan, and Scott Cannon. These new recruits will work under the guidance of Peter Welinder, a seasoned product leader at OpenAI. Meanwhile, the remaining LoveFrom designers, including renowned figures like Mike Matas, will continue their work with Ive, who is currently engaged in designing the first electric Ferrari and collaborating with Airbnb CEO Brian Chesky. According to OpenAI’s press release, Ive and his team will take on significant design and creative responsibilities within the organization. Once their existing client commitments are fulfilled, they will focus exclusively on OpenAI while maintaining their independence. OpenAI is also actively seeking new talent for its “future of computing” initiatives, with roles that promise competitive compensation, including a senior research engineer position offering between 0K to 5K plus equity. The timeline leading to this collaboration is noteworthy. Altman and Ive first met two years ago, and discussions about hardware collaboration began last year. The io division was established at LoveFrom to work closely with a select group of OpenAI employees, and investments from OpenAI and Laurene Powell Jobs were made towards the end of 2024. This partnership is poised to yield a voice-first AI device, with prototypes already in existence, suggesting a product that could be as compact as an iPod Shuffle and potentially wearable as a necklace. Additionally, there are plans for AI earbuds, with Altman envisioning a suite of AI-first products designed to reduce reliance on Apple and Google for distribution. Google's Resilience On the other side, Google demonstrated its own resilience during I/O, with the rollout of AI Mode in Google Search. While its product strategy may appear fragmented compared to OpenAI’s cohesive approach, Google is leveraging its vast troves of personal data to enhance the capabilities of its Gemini model. If Gemini can effectively integrate with services like Gmail, Workspace, and YouTube, it could create a compelling reason for users to remain within the Google ecosystem, much like how Meta successfully integrated Stories into Instagram, curtailing Snapchat's growth. Conversations with Google employees revealed a sense of confidence within the organization. While there is an acknowledgment of the challenges posed by OpenAI’s growth, Google’s Gemini is nearing 500 million monthly users. The company appears to be modernizing its search capabilities more swiftly than anticipated, positioning itself for continued leadership in model development. The latest Gemini models are impressive, and Google is preparing for a future that extends beyond traditional smartphones with initiatives like Project Astra. Industry Developments Anthropic also made headlines this week by unveiling its Claude 4 models, which it claims are the best for coding. In a competitive landscape where major players like OpenAI, Google, and Meta vie for dominance, Anthropic is positioning itself as a key provider of AI models. Microsoft Build faced disruptions during its event due to protests against its business dealings with Israel, overshadowing several noteworthy announcements, including the introduction of Elon Musk’s Grok model on Azure. Google’s smart glasses prototype was also showcased, featuring basic voice interactions and navigation capabilities. While still in early development, Google’s partnerships with eyewear brands signal its commitment to advancing augmented reality technology. Personnel Changes YouTube has appointed Justin Connolly from Disney as its head of media and sports, a move that has sparked legal action from Disney. Tinder CEO Faye Iosotaluno is stepping down, with her responsibilities being assumed by Match Group CEO Spencer Rascoff. Vladimir Fedorov, a veteran engineering executive from Meta, has joined Github as CTO. Will Robinson, formerly Coinbase’s VP of engineering, is now the CTO at Plaid. Stephen Deadman, Meta’s VP of data protection in Europe, is departing due to organizational changes." max_tokens="3500" temperature="0.3" top_p="1.0" best_of="1" presence_penalty="0.1" frequency_penalty="frequency_penalty"].5 billion to onboard approximately 55 individuals, including notable ex-Apple design leaders. OpenAI aims to develop a voice-first AI device and AI earbuds, with prototypes already in existence. Google introduced AI Mode in Google Search during its I/O event and is leveraging personal data to enhance its Gemini model, which is nearing 500 million monthly users. Anthropic unveiled its Claude 4 models, claiming they are the best for coding. YouTube appointed Justin Connolly as its head of media and sports, Tinder CEO Faye Iosotaluno is stepping down, Vladimir Fedorov joined Github as CTO, Will Robinson became the CTO at Plaid, and Stephen Deadman is leaving Meta.
AppWizard
May 22, 2025
Instagram is testing a new feature that introduces an immersive 3D experience for users interacting with photos on the platform, utilizing Meta's Reality Labs and the Meta Quest VR headset. A select group of testers will soon experience this functionality, which aims to enhance visual exploration and engagement with Instagram's content.
Winsage
May 22, 2025
Microsoft's Build developer conference experienced protests for two consecutive days due to the company's ties to Israel. During a keynote by Jay Parikh, a Palestinian tech worker interrupted to express concerns about Microsoft's contracts with the Israeli government, calling for an end to the collaboration. Security removed the protester, who was supported by the group No Azure for Apartheid. The previous day, Microsoft employee Joe Lopez also interrupted a keynote to speak against the company's contracts with Israel, urging colleagues to speak out. Microsoft is conducting an internal review of its technology's role in the Gaza conflict, maintaining that its relationship with Israel's Ministry of Defense is commercial and that its technologies are not misused. This follows previous instances of dissent within Microsoft, including disruptions at a 50th-anniversary event.
AppWizard
May 22, 2025
Telegram has launched a contest with a prize pool of 0,000 inviting participants to create a viral video that showcases Telegram's technological innovations compared to WhatsApp. The contest was announced on May 19, 2025, and submissions are due by May 26, 2025. It is open to global participants, with videos required to be in English and suitable for platforms like TikTok and Instagram Reels. The contest allows the use of AI tools and evaluates entries based on clarity, visual impact, meme potential, and virality. Pavel Durov, the founder of Telegram, has publicly accused WhatsApp of smear campaigns and emphasizes Telegram's commitment to privacy and political neutrality, rejecting government requests to censor political content. The contest aims to expand Telegram's user base and engage content creators by promoting its unique features, which include end-to-end encrypted chats, large group chats, bots, and advanced privacy controls. Winners will be announced in June 2025.
AppWizard
May 14, 2025
Nextcloud, a European software vendor, has raised concerns about Google's treatment of its Android Files application, which has over 800,000 users. The issue centers on the "All files access" permission, which was revoked by the Play Store in 2024, impairing the app's functionality. Nextcloud argues that alternatives like the Storage Access Framework (SAF) and MediaStore API are inadequate for their needs. The app has been able to read and write all file types since its launch in 2016 without security concerns from Google until the recent revocation. Nextcloud claims that Google's policies are stifling competition and that they have faced bureaucratic inefficiencies in addressing their complaints. Despite having a fully functional version on F-Droid, the Google Play version is restricted. Nextcloud has previously lodged a complaint with the EU regarding Microsoft's anti-competitive behavior, and they express frustration over the lack of action taken. They believe larger tech companies are trying to suppress smaller competitors.
Winsage
May 13, 2025
The current landscape of artificial intelligence interactions includes cloud-based tools like ChatGPT and Copilot, but some users, especially developers, prefer running large language models (LLMs) locally. Ollama provides a solution for this preference. To run LLMs effectively, hardware requirements include a GPU, with larger models needing more computational power. For example, Google's Gemma 3 has a 1 billion parameter model requiring 2.3GB of VRAM and a 4 billion parameter version needing over 9GB. Meta's Llama 3.2 has similar requirements. A modern PC with at least 8GB of RAM and a dedicated GPU can utilize Ollama. To install Ollama on Windows 11, users download the installer from the official website or GitHub and follow the installation process. Once installed, it operates in the background, indicated by an icon in the taskbar, and can be accessed via localhost:11434 in a web browser. Ollama primarily uses a command-line interface (CLI), requiring users to use PowerShell or WSL. Key commands include "ollama pull" to install LLMs and "ollama run" to execute them. For instance, to install the 1 billion parameter Google Gemma 3 LLM, users would enter "ollama pull gemma3:1b". Running the models opens a chatbot interface for user interaction, and exiting can be done by typing "/bye". Setting up Ollama is user-friendly and requires minimal technical expertise.
Winsage
May 12, 2025
Microsoft's annual Build developer conference will occur from May 19 to 22, focusing on new AI integrations, services, and applications for Windows. Last year's conference featured the integration of Copilot into Microsoft Teams and the introduction of AI-powered PCs. This year, deeper integrations of Copilot generative AI are expected, catering to consumer and enterprise needs. Keynotes will be streamed online for free. Microsoft raised prices on various software products by 5% last November and may use new AI features to justify these increases. Copilot is anticipated to have significant enhancements, with reports suggesting Microsoft is considering alternative AI models from xAI, Meta, Anthropic, and DeepSeek. Copilot may gain “agentic” capabilities for executing basic PC tasks, with references to an Action button found in the Copilot client for Windows. Microsoft is rumored to have developed a suite of models called MAI to compete with OpenAI’s offerings, potentially to be unveiled at Build. There are also rumors about a successor to the Azure Maia 100 AI Accelerator, possibly named “Maia 2,” with Marvell contracted for its production.
Search