House ambassador Chase Sui Wonders offers a new take on a House emblem, wearing a Bird on a Rock by Tiffany brooch with a diamond of over 13 carats as a choker during the 83rd Venice International Film Festival. She also wears a Jean Schlumberger by Tiffany Ailsa ring in 18k gold and platinum with an aquamarine of over 8 carats and diamonds along with a Bird on a Rock by Tiffany Wings ring in platinum with diamonds. #TiffanyAndCo# #TiffanyHighJewelry#

glossy meets lip kits 👄  a NEW take on our iconic lip kit featuring a supple kiss glaze and a contrasting precision pout lip liner for glossy, defined lips. the glossy lip kit will be available in six shades on 6.24 🩷 @kyliecosmetics

Daisy Edgar-Jones and Esmé Creed-Miles lead a clever, tasteful new take on the Jane Austen classic that firmly targets Gen Z audiences, writes FT film critic Danny Leigh. Tap here for his film review:

This is the only way I can accept another Assassin’s Creed game🚀 I don’t care if it’s a “well-made game” I don’t care if it’s “the story they wanna tell” I don’t care if it follows trends and sells well I don’t care if it’s popular I don’t care if it’s a “fresh new take” I don’t care if it features a half-baked cameo or ancestry to an old character If it has no meaningful 3rd person playable modern day that has a rewarding pay-off and overall sci-fi conspiracy thriller vibe with deep connections to conspiracy theories and technologically advanced ancient civilizations, not mythological gods or fantasy RPG with a line note that justifies it as “tech”, then none of that means anything, and I really don’t want it. All those justification don’t matter to me in the slightest if it doesn’t deliver an experience like the first five games🎤

if you know anyone at a brand looking to reach Gen Z and younger millennial women in 2027, send them my way. we're building an ambitious new take on the video podcast around one of the most compelling creators of her generation, whose relationship with her audience is unlike anything i've seen. not sharing names or details just yet but starting conversations with a handful of potential brand partners with annual budgets starting in the mid six figures. DM me or tag a brand you think would be a great fit.

I've been cooking something pretty cool, and experimental to go alongside @stonkpit and @minepitderp - it uses the same POW mining as now (0 new mines or mining), tiny bit of optional extra Gas, but opens up a plethora of opportunities, It's also designed to fund into the DerpBoys and the Merchant, I'm likely to deploy and test it tomorrow, I'm not announcing what it is completely till I've tested it live on Mainnet It's a new take on an old trick, but i think it's pretty cool. - I also side missioned something else, whilst cooking the big one, which I might deploy before or after the big one - this one was sparked by a reply in discord tonight and by some long overdue ideas from the days of The Gold Mine on Apechain It's built purposefully for DERP, but could extend to every other coin on RH, with one click redeploys. Actually I just had an extra idea for it... but maybe get a MVP v1 out first Gn, derp DERP ⛏️💵

Guizhou isn't just a summer retreat in SW China—it's packed with surprises. Who knew lakeside views and horse riding could create such a thrilling sense of freedom? This is Gen Z's cool new take on mountain adventures. #TrailRiding# #HorsebackRiding# #Camping#

🔁 A way to train "think longer, get smarter" models without the gradients exploding. Title: Thinking with Looped Flows URL: Looped models that recurrently update hidden states to "think" have long struggled with unstable BPTT (backprop through time). Looped Flows fixes this by borrowing training principles from diffusion models. Here are 3 highlights. 🧠 Training recurrence without BPTT By training each step with a local loss at gradually decreasing noise levels, the model learns to keep "thinking" stably, without vanishing or exploding gradients. ⏱ More compute at inference, for free Just using a finer time grid at test time boosts accuracy, from 74.5% at 8 steps to 97.9% at 128 steps on Sudoku, with no retraining needed. 🏆 Beats prior looped models on ARC-AGI It reaches 58.8% on ARC-AGI-1 and 12.2% on ARC-AGI-2, outperforming previous looped-model approaches on both. A neat new take on test-time compute scaling: thinking longer at inference genuinely pays off. #LLMInference# #MachineLearning#