A new open-source toolkit called Gemma Multimodal Fine-Tuner now lets developers fine-tune Google's Gemma AI models on text, images, and audio directly on Apple Silicon Macs, requiring no NVIDIA hardware and supporting terabyte-scale cloud datasets while keeping data fully private on-device.
Spotify is rolling out AI-powered playlist generation for podcasts, letting premium users in select markets build personalized podcast feeds using natural language prompts, with customizable update schedules and AI-generated explanations for each recommendation.
Google launches 'AI Edge Eloquent,' a free iPhone dictation app that uses on-device AI to automatically strip filler words from speech while keeping all voice data completely private.
Granola launches a free AI meeting notes app for Apple devices, automatically transcribing meetings and phone calls while generating smart summaries with highlights, decisions, and action items — earning early praise for accuracy despite minor UI bugs.
Kintsugi, a California startup that spent seven years developing AI software to detect depression and anxiety from speech patterns, is shutting down and open-sourcing its technology after FDA regulatory delays drained its funding before approval could be secured.
Google launches free AI avatars, Veo 3.1 video generation, and AI music tools in Google Vids just as OpenAI shuts down Sora, positioning itself as the dominant force in AI video creation for both consumers and enterprise users.
Microsoft launches three in-house AI models — MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 — claiming best-in-class performance across speech and image generation while undercutting rivals like OpenAI and Google on price, built by teams of fewer than 10 engineers using half the GPUs of competitors.
Microsoft launches three powerful AI models — MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 — undercutting Google and OpenAI on price while reaffirming its multi-billion dollar OpenAI partnership.
A global 'frugal AI' movement is gaining momentum, delivering small, energy-efficient models that run on cheap, offline hardware to serve communities excluded from mainstream AI, including preserving endangered Indigenous languages using minimal data on sub-$50 devices.
Four beginner-friendly AI tools—Wispr Flow, Comet or Atlas browser, Perplexity Discover, and ChatGPT Voice—are quietly revolutionizing daily productivity, with voice dictation accuracy hitting 95% and potential 3-4x efficiency gains, even as public skepticism keeps many users on the sidelines.