AI-generated games are getting seriously popular. Astrocade, a platform for creating games with AI, now has 10 million monthly users, up from around 5 million four months ago. Much of its popular content is Minecraft clones, copyrighted characters, and Roblox knock-offs, with some games reaching millions of plays. CEO Amir Sadeghian says the tech has become so good that it is very hard to tell these games were made with AI. He also claims AI is now “writing code better than a lot of engineers.”

On this week's livestream, Arthur Soroken sits down with Amir Sadeghian, PhD and Ali Sadeghian, the brother-duo and co-founders behind Astrocade, to discuss how generative AI and Gemini multimodal models are redefining game development.

👀 Sneak peek at my next game in @PlayAstrocade , Heroes Unleashed! All character animations were created with GPT-6 Astra using Blender MCP. Before Astra, I simply couldn’t get AI to pull off 3D animation like this. This feels like a whole new level of breakthrough. 🔥 #GameDev# #VibeCoding# #AI# #Astrocade# #GPT6# #Astra#

We signed a launch contract with @astroscale_JP for the ADRAS-J2 active debris removal mission, marking our second launch agreement with @astroscale_HQ. Proud to support this pioneering mission shaping the future of sustainable space. Read more:

🃏 ┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈ ✧ Trial of the Astrolabe Season 7 is Open! ✧ ┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈┈ ⌛️ Starting this season, in addition to Mindscape materials, you can also earn Meta Jewels! Complete specified stages to earn Meta Jewels! ✨ If you haven't taken the Trial yet, this is the perfect time to give it a shot! Trial of the Astrolabe Season 7 📅 Available Until: July 29, 7:59 PM (UTC) #P5XEN#

Ever wondered what it’s like to navigate the Mediterranean using an ancient astrolabe? @FryRsquared sets sail to test the navigation tool first developed by Greek astronomers around 150 BCE. Watch Infinite Explorer

🎬 Distilled autoregressive video models are fast but tend to drift from human preferences. Astrolabe answers that challenge by doing RL alignment in the forward process, with no re-distillation and no reverse-process unrolling. Title: Astrolabe: Steering Forward-Process Reinforcement Learning URL: 📝 Overview Astrolabe is a reinforcement learning framework that aligns distilled autoregressive (AR) video models with human visual preferences. Its defining feature is doing RL in the forward process rather than via conventional reverse-process optimization. It is a large 53-page, 37-figure study. ❓ Challenges Solved Distilled AR video models suit efficient streaming generation but tend to misalign with human preferences. Worse, existing RL doesn't fit these architectures naturally: it typically needs either expensive re-distillation or solver-coupled reverse-process optimization, both heavy and hard to scale. 💡 Methodology & Proposed Approach It rests on three innovations. ・Negative-aware fine-tuning contrasts positive and negative samples at inference endpoints to establish an implicit policy-improvement direction without unrolling the reverse process ・A streaming training scheme generates sequences progressively via a rolling KV-cache, applying RL updates only to local clip windows while keeping long-range coherence through prior-context conditioning ・A multi-reward objective integrates uncertainty-aware selective regularization and dynamic reference updates to mitigate reward hacking, the collapse where only the apparent score rises 🎯 Use Cases It fits real-time streaming video generation where you want to align an efficient distilled model with preferences while preserving its speed. It applies across multiple distilled AR video models and raises quality without sacrificing inference efficiency. 📊 Significance and Results ・By avoiding the heavy paths of re-distillation and reverse-process unrolling, it addresses computational efficiency bottlenecks ・Combining forward-process negative awareness, streaming updates, and reward-hacking mitigation, it provides a robust, scalable alignment solution ・It demonstrates effectiveness across several distilled AR models, with detailed quantitative evaluation and ablations #VideoGeneration# #ReinforcementLearning#

Here is every holding Tema Space Innovators ETF $NASA Rocket Lab $RKLB 10.54% SpaceX SPV (private) 10.29% Planet Labs $PL 6.41% Filtronic $FTC 5.80% Intuitive Machines $LUNR 5.79% Firefly Aerospace $FLY 5.05% 5N Plus $VNP 4.15% OHB SE $OHB 4.01% BlackSky Technology $BKSY 3.97% Viasat $VSAT 3.94% AST SpaceMobile $ASTS 3.89% MDA Space $MDA 3.34% Universal Microwave Technology 3.12% York Space Systems $YSS 3.10% Nlight $LASR 2.95% Iridium Communications $IRDM 2.56% Astroscale Holdings 2.38% QPS Holdings 1.92% Voyager Technologies $VOYG 1.77% EchoStar $SATS 1.76% Sky Perfect JSAT 1.70% Vishay Precision Group $VPG 1.47% Lightpath Technologies $LPTH 1.41% KVH Industries $KVHI 1.03% Sphere Corp 0.94% GomSpace $GOMX 0.89% CPS Technologies $CPSH 0.88% Win Semiconductors 0.83% NextNav $NN 0.83% Synspective 0.75% Velo3D $VELO 0.62% Intellian Technologies 0.60% Gilat Satellite Networks $GILT 0.52% Avio $AVIO 0.46% Redwire $RDW 0.44% Scanway $SCW 0.17% Disclosure: @Temaetfs is a WOLF Financial Partner