Morris C.

We've spoken with hundreds of ad creatives, marketing designers, filmmakers, and animation teams — and heard the same thing: the outputs look great… until they don't 😅. When they fail, it's incredibly hard to tell why. Is it the prompt, the model, or the world itself quietly breaking? That ambiguity is the real bottleneck. Physion-Atlas 1.0 introduces a more objective, diagnostic way to evaluate video world models — moving beyond high-level comparisons to surface what actually matters. It disentangles prompt misalignment from physical and visual inconsistencies, grounding every judgment in explicit spatiotemporal evidence. Not just which output is better, but what breaks, when, where, and why. From abstract comparisons → diagnosable reality 🔍 📄 Blog: physionlabs.ai/blog/physion-a… 📝 Evaluate your model: docs.google.com/forms/d/e/1FAI…

Everyone can build 😃 , but not everyone can merge 😅 . Everyone on our team can build now. That's not a metaphor. Our designers, PMs, and growth team are turning ideas into working code with Claude Code, Codex, and Cursor. Real features. Real prototypes. Real PRs. But here's what we learned fast: Everyone can build. Not everyone can merge. AI makes it easy to generate something that works on your machine, in your test case, for your specific prompt. It's much harder to generate something that's: - Safe to run at scale - Easy for the next engineer to maintain - Consistent with architecture decisions made six months ago - Not going to quietly break something else on a Tuesday The gap isn't effort. Everyone is trying. The gap is engineering judgment, and that's still hard to prompt your way into. So as an AI-native team, we've had to build a new layer: Not just "who can build," but "what is merge-ready." The bar for a PR is no longer: "Did you make it work?" It's: "Does it belong in production?" That distinction is one of the most important things we've figured out this year.

It's crazy to realize how my life depends on AI right now. I filed my 2025 taxes using AI, read and write contracts with it, managed the process of buying, remodeling, and designing an apartment, and have it ingrained pretty much everywhere. I know these models are unreliable and hallucinate all the time, so I try to be as methodical and careful as possible. Rule Numero Uno: Never trust a single model. For anything high-stakes, I compare answers across multiple models. I use Claude, ChatGPT, Gemini, Gamma 4, and lately Qwen as well. Sometimes it's annoying to copy and paste across multiple models, but 100% worth it. I think I found a better process. The @cueyofficial team reached out to me with a solution for this. They built a Chrome extension that runs alongside any AI tools you already use. When you send a prompt, Cuey checks other models in the background and gives you the stronger answer if one is available. No need to copy/paste across models. Cuey does that behind the scenes. A few notes: • It works inside ChatGPT, Claude, and Gemini directly. You don't need to change your workflow. • It supports 30+ models across tiers, from fast/cheap to advanced reasoning models. • It's positioned specifically for judgment-heavy work: research, strategy, contract review, tradeoff analysis. • Privacy-first: they don't log your prompts or use them for training. You'll find the link below.

Gemini 3.0 is now live on Forge. You can call Google’s newest model through the same unified TensorBlock API you already use across providers. One interface, many models. Switch providers without changing your application logic. Try it here → forge.tensorblock.co

@13yearoldvc @DavideCrapis @ethereumfndn Congrats on a fantastic move!

I am joining @DavideCrapis on the dAI team at @ethereumfndn as an AI Advisor & Coordinator to help shape the ERC-8004 ecosystem and the agent economy on Ethereum. If you’re building toward autonomous commerce or "Stripe for agents", reach out 😃 → ai@ethereum.org

@cyodyssey 确实有效

最近找了除了扫地，拖地之外第三个寻找内心平静的方式，刷杯子。

At KBW 🇰🇷, we joined @googlecloud, @eigencloud & @saharaai on the “Decentralizing Intelligence” panel at Hack Summit by @hack_vc. Conversations highlighted the accelerating shift toward open AI and the growing demand for infrastructure to support agentic intelligence.

@PhalaNetwork @ConfidentialC2 Congrats!

Phala has joined the Confidential Computing Consortium (CCC). The @ConfidentialC2 brings together industry leaders to accelerate the adoption of TEE technologies and standards. CCC powers privacy-forwarding infrastructure, with members like AMD, Intel, ARM, NVIDIA, huawei and more.

@karminski3 从隐私角度看，在 Mac 上运行模型推理是合理的；但若要在 Mac 上训练模型，就完全不make sense了。

还记得年初那个可以让多台 Mac 组成集群运行大模型的框架 exo 吗？貌似已经停更了。 从github上看，exo项目最后一次提交还是3月份，而我去看了他们的 discord，也有人说官方已经好几个月没出现了。目前他们的 X 账号还在更新，看上去是全都铺在了新的框架 kpop 上。好像还是打算搞 Mac 训练大模型。 说实话，我觉得他们可能一开始的创业方向就是有问题的。Mac 上运行大模型进行推理是个很不错业余与项目。但是作为商业项目很明显毫无盈利空间（人们不会因为这个付钱）。于是他们就想当然的想用 Mac 来训练大模型.... 这是歪的离谱了... Mac 在推理上还算行，但在训练大模型上跟老黄的卡比那真的是毫无性价比了。

@DavideCrapis This is awesome, congrats Davide!

We’re starting a new AI Team at the Ethereum Foundation (the dAI Team). Our mission: make Ethereum the preferred settlement and coordination layer for AIs and the machine economy. The team will focus on two main areas: - AI Economy on Ethereum = giving AI agents and robots ways to pay, coordinate, and follow rules without middlemen. - Decentralized AI Stack = making sure the future of AI doesn’t rely only on a handful of entities but has open, verifiable censorship-resistant alternatives. We believe Ethereum can be as useful for today’s AI developers as it will be for the sci-fi future. That’s why we’ll work closely with ecosystem projects to accelerate progress and push the boundaries of research and innovation at the intersection of AI and blockchains. Connecting two communities that have too often worked in parallel. We’ll work side by side with both the Protocol and Ecosystem teams at the EF. Linking protocol improvements with the needs of AI builders, and funding innovative public goods that will make Ethereum the best home for AI. Ethereum makes AI more trustworthy, and AI makes Ethereum more useful. The more intelligent agents transact, the more they need a neutral base layer for value and reputation. Ethereum benefits by becoming that layer and AI benefits by escaping lock-in to a few centralized platforms. Acting with purpose and urgency, we’ll continue our recent work on ERC-8004—a standard for proving who an AI agent is and whether you can trust it— and continue to support new upgrades and standards. Inspired by the d/acc philosophy and Ethereum values, we’ll support projects that are leveraging Ethereum to build a more resilient and secure AI ecosystem where humans can flourish with AI. Ethereum + AI is about making sure humans keep agency and AI can reach its potential. Neutral, verifiable, censorship-resistant infrastructure means AI works for the people, all of us. The team starts with myself as lead, and we’re just getting started. Explore our job postings and resources below, and join us in building the future of decentralized AI on Ethereum.

@DavideCrapis 100% agree

A bold mission requires a dedicated and focused team.

We’re excited to work alongside @WalrusProtocol, the global data layer, to build the modular infrastructure for the next wave of AI-driven systems. Together, we’re making intelligence open, composable, and ready to scale.

@tensorblock_aoi 🤝 @Zai_org GLM-4.5, Z.ai’s most advanced LLM, is now live on TensorBlock Forge. Just connect your Z.ai API key to access it seamlessly through Forge’s unified interface. This collaboration brings state-of-the-art AI capabilities to users—enabling smarter code generation, reasoning and stronger agentic functionalities. Learn more and get started here: forge.tensorblock.co

Don’t sleep on the next one 👀. Packed VIP founders dinner w/ @hack_vc & @krakenfx during SBC in Berkeley. Wall-to-wall founders, researchers, & frens. The kind of room where deals spark and legends meet.

We're honored to be featured by @startuptofollow! Thank you for the thoughtful interview and spotlight. Check out the article here: 👉 startuptofollow.com/article/buildi…

@karminski3 Cerebras把Qwen3-coder推到了2000 tokens/s是真猛啊 x.com/CerebrasSystem…

Kimi 新发布了个型号 kimi-k2-turbo-preview 应该只是 kimi-k2-0711-preview 的不同部署，这个 turbo 部署在了更快的集群上或者配额更高一些。速度从 10 token/s 提升到 40 token/s 这个速度还是不太够看，建议搞到120 token/s 啊, 这才是具备战斗力的速度.

@cerebras 2000 tokens/s is crazy! We tested Qwen3-coder + Gemini-2.5-flash on Claude code and the result is pretty good. It is time to test again! x.com/tensorblock_ao…

🟧🟨QWEN3 CODER is LIVE on Cerebras 🟨🟧 2,000 tokens/s - 20x faster than Sonnet 0.5s time-to-full-answer 131K context $2 per M input/output tokens Available in Cline, Windsurf & more

@jamwt Absolutely. Fast, cheap, and iterative — with human-in-the-loop and feedback, you gain intelligence and maximize impact. We tested Qwen3-Coder + Gemini-2.5-flash on Claude code via TensorBlock Forge. It is fast and good! x.com/tensorblock_ao…

Why isn't everyone just using Kimi K2 on @GroqInc for codegen? Sure, it's only 90% as good as sonnet. But it's like 3x+ faster and 5x+ cheaper. Run it twice and you're ahead. This is the future: use validation, embrace failure, and run quick/cheap until success.

🎉🏆 Winners of the last week's DevHunt!🎉 Our top 3 champs are: 1⃣🥇 Supabricks by @DarioDiCarlo2 - best Dev Tool of the week!!! ▶️ devhunt.org/tool/supabricks 2⃣ @tensorblock_aoi 🥈 ▶️ devhunt.org/tool/tensorblo… 3⃣ @ossalternative 🥉 ▶️ devhunt.org/tool/openalter… Keep up the great work, everyone! 👍

@tensorblock_aoi @claude_code @AnthropicAI The full context is x.com/AnthropicAI/st…

🔥 @claude_code Nerfed? We Got You Covered. @AnthropicAI just dropped tighter weekly rate limits. Auto-mode? 24/7 agents? Multi-threading? Say goodbye. But with Forge, you stay unstoppable: ✅ Mix Claude Code with any model ✅ No code change ✅ Beat the cap. Stay in flow. 🛠 Try now: github.com/TensorBlock/cl… Best performance. Full control. Let’s build. 💥 #vibecoders #claudecode #LLM #AItool #DevTools