If you want your generations to feel designed and not…generated We hid a few invite codes to @variantui in this video The first 100 get unlimited credits, reply for a hint

We've raised $7M to help companies build AI agents that actually learn and work. @Osmosis_AI is a platform for companies to fine-tune models that outperform foundation models with reinforcement learning. Better, faster, and cheaper.

The U.S. just finalized a $20B currency swap with Argentina to stabilize markets. But President Javier Milei faces political backlash at home after a major electoral defeat, corruption scandals, and stalled reforms. His economic turnaround may hinge on October’s midterms. |

. @MKBHD liked my intern project at apple!! I worked on bringing widgets to visionOS. I prototyped pinning widgets to surfaces and making them feel like 3D furniture in your home. Check it out:

We raised $5.9M to build @variantui: code generation with taste, range, and room to explore Invites start rolling out today, reply for early access

Had this feature in the very first version of https://t.co/KALWfNCfQ6 over a year ago. Surprised it hasn't gone mainstream yet

Future of the mobile OS but also LOOK MOM IM IN THE DEMO 🫨

AI is going to completely change the way we interact with our phones. Here's a first look at how we're using it to build a home screen without apps.

give me now

gpt 4.5 is available in Iris for pro subscribers, $10/mo.

Here are 3 reason why I feel like making the switch from chatgpt to iris @ronithhh @interfaceclubhq : 1. Accessibility: Iris appears at the corner of my screen with a simple key press, allowing me to view my existing work seamlessly.

also since we are getting an influx of new Iris users: I wanna clarify that this is a prototype for an upcoming release — it is not included in the current shipped version of the product

@interfaceclubhq @jasonyuandesign this is a prototype that does work fairly well with today's LLM. But we can't fully take advantage of SOTA model intelligence due to latency. there's soooo much more that will be possible when they get faster. https://t.co/51Zah1YV6T

Dynamic prompt field

Thinking…

Not all models... @variantui in one shot, no screenshots, no component libraries

Very happy to share what we’ve been working on! This came from a simple question: If DeepSeek used RL (GRPO) to train foundation models, what if we do something similar to help AI agents learn? And what if we then make it happen in real-time? Answer: big improvements! 1/

the initial design direction for our next major release @interfaceclubhq was centered around real-time shapeshifting UI that morphs on every keystroke. Unfortunately was not realistic to make with today's models. Might dig up an old demo we built and post later

@jensrst and I have made some really wild prototypes @interfaceclubhq lmao

He's disappointed the tsunami didn't hit

Yep. We are on the case @interfaceclubhq