Apple is reportedly negotiating a mega-deal with Google to license Gemini for new AI-powered iPhone features landing later this year, alongside talks with OpenAI. Apple plans to use an external partner for cloud-heavy tasks like image creation and writing help while running smaller models on-device. It shows Apple is feeling the pressure to catch up, even if it means handing its rival another win.
Apple just acquired Canadian startup DarwinAI to bolster its artificial intelligence division ahead of iOS 18. DarwinAI specializes in making AI models smaller and more efficient, a key technical piece as Apple pushes to run generative AI features directly on user devices for privacy.
Motorola just teased its upcoming Moto X50 Ultra, positioning the F1-inspired device as its first true AI-powered phone to rival Samsung's Galaxy S24. The teaser hints at a reveal tied to the F1 China Grand Prix on April 21, sporting a leatherette back. Why it matters: Phone makers are rushing to make artificial intelligence the headline feature of 2024 flagships.
Microsoft just detailed DirectSR, a new Windows API built with Nvidia, AMD, and Intel to streamline AI super resolution across PC games. Instead of writing custom integrations for DLSS, FSR, and XeSS separately, developers can now support all three through a single code path. It rolls out soon via an Agility SDK preview.
Apple CEO Tim Cook just told shareholders that the company will "break new ground" in generative AI later this year, right as reports confirmed Apple shuttered its electric car project to shift staff to AI. The move signals an aggressive push to catch up with rivals by baking generative intelligence directly into Siri and built-in system tools.
Meta CEO Mark Zuckerberg just met with LG Electronics leadership in Seoul to discuss teaming up on next-gen extended reality devices and on-device AI. Facing fresh heat from Apple's Vision Pro, Meta is looking to marry its large language models with external hardware muscle. Zuckerberg's South Korea itinerary also includes talks with Samsung and President Yoon.
Google just released Gemma, a new lightweight family of open models built on the same research as Gemini. Available in 2B and 7B parameter sizes, it runs directly on laptops or desktops, beats rival open models at its scale, and comes with permissive commercial licensing. It gives local and open-source AI developers a massive new alternative to Llama.
OpenAI founding member and prominent AI researcher Andrej Karpathy has departed the ChatGPT creator. Karpathy, who rejoined OpenAI in 2023 after leading Tesla's Autopilot efforts, noted on social media that his immediate plan is to work on personal projects. His departure marks the loss of one of OpenAI's most influential research voices.
NVIDIA just released Chat with RTX, a free tech demo that lets you run customized chatbots completely offline on Windows. By connecting local files and YouTube links to open-source models like Mistral and Llama 2 via TensorRT-LLM, you can query notes and documents without sending data to the cloud. You will need a GeForce RTX 30 Series GPU or higher with at least 8GB of VRAM.
Microsoft just rolled out a new Copilot experience in the latest Windows 11 Canary builds, making the assistant far more context-aware. When you copy text or images, the taskbar icon animates to offer instant actions like summarizing or explaining, while Notepad is also getting right-click AI deciphering for code and logs. It's an early look at how Microsoft plans to reshape Windows for its 2024 "AI PC" push ahead of the 24H2 update.
Tim Cook just confirmed that Apple is putting “tremendous time and effort” into generative AI, with features officially rolling out to customers later this year. While Cook refused to spill specifics on the earnings call, the rare preview aligns with reports that iOS 18 could be a massive update. Apple rarely tips its hand early, signaling it's ready to counter AI moves from Google and Samsung.
Google just rolled out its "Circle to Search" AI feature to Pixel 8 and 8 Pro phones, following its debut on the Galaxy S24. Users can long-press the home button or navigation bar within any app, circle or highlight anything on screen, and refine their search with text prompts without jumping between apps.
Samsung just tapped Baidu's Ernie model to power its Galaxy AI features on the Galaxy S24 in China, ditching Google's Gemini. Baidu will handle live translations, summarization, and local search features. Why it matters: As Samsung struggles to regain lost ground in the Chinese smartphone market, relying on local AI champions is a necessary compromise.
Samsung just unveiled the Galaxy S24 lineup, packing Google's Gemini models directly into its latest flagships. The devices feature Circle to Search, real-time call translation, and generative photo editing powered by both on-device and cloud AI. It is a major push to put everyday generative AI tools into the hands of millions.
Security firm Trail of Bits just disclosed a vulnerability dubbed LeftoverLocals impacting GPUs from Apple, AMD, and Qualcomm, allowing attackers to siphon up to 180 megabytes of data directly from local GPU memory. The flaw exposes user queries and model weights from locally run LLMs. Apple patched it on M3 and A17 chips, but millions of older devices remain vulnerable.
Microsoft just unveiled a dedicated Copilot key for Windows 11 keyboards, marking the first major layout tweak in nearly 30 years. Tapping the button instantly launches Windows Copilot, putting everyday AI assistance right next to the spacebar. The new key starts rolling out this month across partner laptops and upcoming Surface devices.
Samsung just announced an in-person event on Jan. 17 in San Jose to unveil its next generation of Galaxy smartphones, promising an "all-new mobile experience powered by AI." The devices are expected to leverage Qualcomm's Snapdragon 8 Gen 3 chips to run generative AI models directly on the phone without an internet connection. Why it matters: Samsung wants on-device AI features to give it an edge over Apple in the high-end smartphone race.
Apple just detailed how it plans to fit heavy AI models onto everyday devices without blowing past hardware limits. In a paper titled 'LLM in a flash', researchers store parameters on flash memory and stream them to DRAM on demand, running models twice the size of available RAM while speeding up GPU inference by up to 25x. Why it matters: It lays the technical groundwork for reliable on-device generative AI.
Microsoft just released Phi-2, a 2.7B parameter language model trained on textbook-quality synthetic and web data. Despite its compact footprint, it matches or outperforms models up to 25x larger like Llama-2-70B on complex reasoning and coding benchmarks without RLHF. Why it matters: it demonstrates that strategic data curation can unlock emergent reasoning even in lightweight models.
The Stable LM team just released Stable LM Zephyr 3B, a 3-billion-parameter chat model that is 60% smaller than typical 7B alternatives. Tuned with Direct Preference Optimization (DPO), it delivers responsive instruction-following and Q&A on consumer hardware without high-end GPUs. It is currently available under a non-commercial license.
