Apple launched two new open models: FastVLM and MobileCLIP2

The models are up to 85x faster and 3.4x smaller than previous work

Suitable for high-res image processing tasks like OCR, image captioning, visual question answering, and image understanding

Microsoft just launched a larger 10B parameter version of its VibeVoice text-to-speech model

Available under MIT license, it generates multi-speaker podcasts (going up to 45 minutes) in minutes

Supports up to 32K context

Tencent launched two new Hunyuan models:

—HunyuanWorld-Voyager, an ultra-long-range world model with native 3D reconstruction and memory

—Hunyuan-MT-7B and Hunyuan-MT-Chimera-7B, a joint AI translation system that outperforms rivals across 33 languages

OpenAI is reportedly planning a Stargate data center in India with at least 1GW of capacity

The company is also planning to open its first office in Delhi, with CEO Sam Altman slated to visit soon

If you use canva - https://t.me/HackerS_AsyLum/3230

📊 OpenAI buys Statsig for $1.1B

Statsig boss joins OpenAI as CTO of Apps → runs ChatGPT + Codex.
Statsig keeps its own brand.

Other changes:
• New CTO B2B Apps = Srinivas
• New “AI for Science” lead = Kevin

Why it matters: OpenAI wants more top talent → buying teams is faster.

🍌 Create ads with Nano Banana

Google’s Nano Banana (Gemini 2.5 Flash Image) makes pro ads cheap → ~4¢/image.

How:
1. Go gemini.google.com → Tools → turn on “Create Images.”

2. Upload product + prompt: “Create ads to fit various environments with the attached image — subway, train station, bus stop, luxury magazine”

3. Nano Banana auto-fits product into real ad scenes.

4. Upload sketches → it colors + keeps shape.

5. Turn on “Create videos with Veo” → ads become short videos.

Pro tip: Test full ad campaigns for pennies before real spend.