LA Metro is installing AI-powered cameras on city buses to automatically catch drivers illegally parked in dedicated bus lanes and stops. Violations won't be purely automated—footage gets sent to the transportation department for human review before tickets are issued. Testing kicks off this summer.
SenseTime just launched SenseNova 5.0, a major model trained on over 10TB of tokens with a 200,000 context window. Alongside the cloud model, it rolled out an edge-side matrix that hits 78.3 words per second on flagship chips and cuts enterprise inference costs by 80%. Why it matters: SenseTime is pushing frontier multimodal AI directly into phones, cars, and edge devices.
Microsoft just launched Phi-3 Mini, a lightweight 3.8B-parameter AI model trained on synthetic bedtime stories. Microsoft claims it rivals models 10 times its size, like GPT-3.5, while running on phones and laptops. It is available now on Azure, Hugging Face, and Ollama, with 7B and 14B versions coming soon.
Apple reportedly acquired Datakalab, a Paris-based startup specializing in low-power computer vision and on-device AI compression algorithms. Finalized in December, the deal brings Datakalab's patents and engineering talent into Apple's ecosystem. Why it matters: As Apple prepares its AI push for iOS 18 and Vision Pro, running powerful models directly on your hardware without draining battery or risking privacy is the core battleground.
The International Olympic Committee just unveiled its Olympic AI Agenda, laying out plans to use AI across judging, talent scouting, and broadcasting. Demos featured algorithms analyzing diving mechanics in under 0.1 seconds and Intel using phones to scout young athletes in Senegal, with initial rollouts planned for Paris 2024.
Meta just launched Llama 3, delivering 8B and 70B open-access models trained on over 15 trillion tokens. The smaller 8B variant can easily run and fine-tune on consumer GPUs down to 4-bit precision, backed by an expanded 128k vocabulary. Why it matters: Meta is resetting the benchmark for open AI, giving developers frontier-level capabilities without proprietary API lock-in.
NVIDIA just rolled out full-stack optimizations for Meta's brand-new Llama 3 across its data center GPUs, edge devices, and PCs. In early benchmarks, a single H200 GPU churned out roughly 3,000 tokens per second on the 70B model, enough to handle 300 users simultaneously. The 8B model is also landing on RTX PCs and Jetson Orin modules.
Apple is reportedly planning to run its initial batch of iOS 18 AI features entirely on-device without cloud servers. Basic capabilities like smart text generation in Messages, Safari, and Siri will work offline to ensure privacy, while heavier tools may tap cloud partners. Apple is expected to detail the plans at WWDC on June 10.
PaintCam just launched Eve, an AI-powered home security camera that actively fires paintballs and tear gas at intruders who ignore its warnings. The system uses computer vision to distinguish friends from foes, but turning backyard defense into autonomous armed deterrence opens up a minefield of legal and ethical questions.
Hugging Face just released Idefics2, an open-source 8B vision-language model licensed under Apache 2.0. It preserves native image resolutions, significantly upgrades OCR, and punches way above its weight class on visual QA benchmarks against 30B+ rivals. Why it matters: Developers get an efficient multimodal base model ready to fine-tune directly in Transformers.
Google just released CodeGemma in collaboration with Hugging Face, bringing a family of open-access code specialists built on Gemma. Trained on an extra 500 billion tokens, the lineup includes an infilling-focused 2B model that runs on consumer hardware alongside 7B base and instruct variants. Why it matters: It offers developers fast, private code completion without needing massive cloud infrastructure.
Apple just made a big push to bring Vision Pro into corporate workspaces. Powered by on-device ML and M2's Neural Engine, the headset integrates with Microsoft 365 Copilot for spatial productivity and taps Nvidia Omniverse APIs to stream complex 3D digital twins in real time. Why it matters: Apple is positioning spatial computing as a serious enterprise AI tool, moving far beyond casual consumer entertainment.
Stability AI just launched Stable LM 2 12B, a multilingual model trained on 2 trillion tokens across seven European languages. Available in both base and instruction-tuned flavors, it is built to handle function calling and RAG workflows on standard hardware. The compact 1.6B model also got an upgrade with better conversation skills and tool use.
Opera just added experimental support for 150 local LLM variants to its Opera One developer build, letting users run models like Llama, Gemma, and Mixtral completely on-device without sending prompts to cloud servers. Each model takes 2 to 10 GB of local storage and replaces the built-in Aria assistant during chats. Why it matters: It turns local AI setup into a one-click browser download instead of a complex terminal hassle.
Apple researchers just unveiled ReALM, a lightweight AI model that decodes your screen context by converting visual layouts directly into text. Despite having far fewer parameters than GPT-4, it substantially outperforms OpenAI's flagship on screen-specific tasks—offering a blueprint for a lightning-fast, on-device Siri that can truly take action on what you are looking at.
Ola Electric CEO Bhavish Aggarwal just teased the Ola Solo, a driverless electric scooter concept packed with self-balancing tech. While the April Fools video poked fun with joke specs like an 'LMAO9000' chip, Aggarwal clarified the underlying autonomous two-wheeler tech is real and actively in prototyping for future vehicles.
Microsoft is preparing to run parts of its Copilot AI locally on PCs to cut latency and boost privacy, according to Intel. The catch? Next-gen AI PCs will reportedly require NPUs pushing over 40 TOPS, a bar no consumer chip on the market currently clears.
Apple just announced that WWDC 2024 will run from June 10 to June 14. The main event will be iOS 18, expected to mark Apple's official push into generative AI with a smarter Siri, automated summarization, and AI code generation inside Xcode. Why it matters: Cupertino is finally ready to show its AI hand.
Stability AI just rolled out Stable Code Instruct 3B, an instruction-tuned coding LLM that beats CodeLlama 7B despite having less than half the parameters. It handles natural language queries, math, and code generation across popular languages. Why it matters: Developers can run powerful coding assistants on consumer-grade hardware, with weights ready to download on Hugging Face.
Microsoft just unveiled its first AI-powered Surface PCs built exclusively for business: the Surface Pro 10 and Surface Laptop 6. Powered by Intel Core Ultra chips with dedicated NPUs, both machines feature a dedicated Copilot key for one-touch access to AI features. The hardware arrives this April alongside upgraded cloud IT management tools.
