Trick 1 made the clip above. The Hero Orbit. One product photo in, an Apple-style orbit ad out. No set, no crew, no turntable. Prompt:
"Create a 10-second 16:9 cinematic product showcase in one continuous shot. [Product] rests on a glossy dark concrete floor. Camera slowly
TECHNOLOGY
-
Hero Orbit: Apple-style ad from one product photo
By
–
-
Cartesia: a pair of number one models for conversation
By
–
Cartesia did not just release two models, it released the only pair from one company ranked #1 on both ends of a conversation.
— AI Highlight (@AIHighlight) 15 juin 2026
Sonic-3.5 turns text into speech and Ink-2 turns speech into text, faster than anything else streaming.
That is the whole voice loop locked down by… https://t.co/OA6mvEx06nCartesia did not simply release two models; it released the only pair from the same company ranked number one at both ends of a conversation. Sonic-3.5 converts text to speech and Ink-2 converts speech to text, faster than anything else.
-
AI-powered oven toasts bread using camera and NVIDIA Jetson
By
–
My whole life is that way. My toast was toasted by an oven with a camera that looked at the toast, sent that to an NVIDIA Jetson card, and an AI decided how long to toast it. What a world!
-
New data shows US AI talent advantage no longer permanent
By
–
One year after their 1st white paper analyzing DeepSeek’s talent base, HAI senior fellow @AmyZegart and @HooverInst
's Emerson Johnston publish new data that shows even more clearly that the U.S. can no longer assume a permanent AI talent advantage. (2/4) https://
hoover.org/research/updat
e-deepseek-ai-and-the-great-talent-competition
… -
Future brief access to top AI models for everyone
By
–
In the future everyone will have access to the SOTA models for five minutes.
-
OrcaRouter detects hallucinations and ensures reliability
By
–
A model has a bad day and your output collapses.
OrcaRouter makes reliability structural, not a matter of luck.
If a leg hallucinates, the judge detects it and provides the answer that holds the road. -
Gemma 4 31B runs fastest on SambaCloud, over 30% faster
By
–
Gemma 4 31B is running fastest on SambaCloud ⚡
— SambaNova (@SambaNovaAI) 15 juin 2026
As verified by @ArtificialAnlys, SambaCloud delivers Gemma 4 31B over 30% faster than the next provider.
Same model. More speed. Better agent experiences.
Read more: https://t.co/Y61VK1AiOJ pic.twitter.com/QDkY6vpdAMGemma 4 31B is running fastest on SambaCloud As verified by @ArtificialAnlys
, SambaCloud delivers Gemma 4 31B over 30% faster than the next provider. Same model. More speed. Better agent experiences. Read more: https://
sambanova.ai/blog/gemma-4-3
1b-running-fastest-on-sambacloud?utm_source=x&utm_medium=organic&utm_content=blog-announcement
… -
Model distillation: the most powerful are American, interesting examples
By
–
The reality is that everyone distills from everyone, usually from the most powerful, and the most powerful right now are American models. Here are some interesting examples.
-
ElevenLabs Music v2 enables developers to build music generation workflows
By
–
Music v2 already powers ElevenMusic and ElevenCreative.
— ElevenLabs (@ElevenLabs) 15 juin 2026
Now developers can build the same music generation and editing workflows directly into their own products, unlocking new experiences like Around the world by ElevenLabs.https://t.co/6ctp9EXJJG pic.twitter.com/A2wOcwIkIrMusic v2 already powers ElevenMusic and ElevenCreative. Now developers can build the same music generation and editing workflows directly into their own products, unlocking new experiences like Around the world by ElevenLabs. http://
elevenexperiments.com/around-the-wor
ld
… -
Build a GPT-style transformer from scratch without high-level libraries
By
–
Train your own LLM from scratch. This repo builds a GPT-style transformer from the ground up, without using any high-level libraries. You see exactly how attention, multi-head attention, the feed-forward block, embeddings, residuals, and layer norm fit together. And it doesn't