Yu-Jen Lin

Local AI models have just gotten insane. Having some fun with Flux.1 Kontext [dev] with my AI PC. Sample below. This is just scratching the surface. InPaint, outpaint, stylizing, photo compositing all possible just by prompting what to change. This is the future. I did this using ComfyUI, part of my AI Playground install, running on my Intel Core Ultra 200V (Lunar Lake) laptop. ComfyUI now provides template workflows for to do this. Very handy, worked straight away.

29

Podonos just raised $2.4 million in pre-seed funding, and the signal is loud and clear: the voice AI industry finally has someone measuring the sound, not just making the noise. Based in Los Gatos, Podonos is building the infrastructure layer that decides whether voice AI feels real, human, and ready for primetime. The industry is racing toward a projected $47.5 billion by 2034, but growth without standards is chaos. Podonos is making sure the hype actually holds up when you press play. At the center is ✦Soohyun Bae, PhD in Computer Engineering from Georgia Institute of Technology and Y Combinator W22 alumnus. His track record stretches from engineering roles at Google Maps to leading AR mapping at Niantic, plus co-founding Bobidi and TickTock AI. He has seen what happens when technology scales without proper #guardrails. With Podonos, he's betting the future of AI isn't just about how smart the models get, but how believable and trustworthy they sound. Podonos evaluates models across naturalness, similarity, emotion, recognition accuracy, pronunciation, tone, and resilience. Add in #personaconsistency, and suddenly you're not just testing machines, you're testing their ability to pass as human. The twist? They deliver results in under twelve hours, when the legacy approach drags on for two months. In AI time, that's the difference between relevance and irrelevance. Investors caught the beat. Serac Ventures, led by Kevin Moore, took the lead, backed by NAVER D2SF, the venture arm of South Korea's NAVER, and KAIST Venture Investment Holdings. Together, they've now backed Podonos twice, with this $2.4 million round stacking on top of a $750,000 raise in March 2025. Total funding sits at $3.15 million, and every dollar is pointed toward scaling #engineering, refining #analysistools, and launching into the U.S., Europe, and Southeast Asia. This isn't theory. Podonos already serves Resemble AI, Play AI, and Sanas AI, hitting six-digit ARR by November 2024 with just four employees. Their secret weapon is reach: 150,000 evaluators spread across nine languages and thirteen locales, forming a human-in-the-loop platform that blends speed, rigor, and global coverage. That reach lets Podonos deliver evaluations at a scale and pace that no competitor has matched. The strategy is ambitious: expand into high-demand sectors like #healthcare, #finance, #gaming, and #advertising, then move beyond voice into #multimodal model evaluation, spanning video AI and large language models. The mission isn't just to check if AI works, it's to decide if AI works for people. #Startups #StartupFunding #EarlyStage #VentureCapital #PreSeed #AI #AIInfrastructure #Voice #VoiceAI #VoiceTech #Infrastructure #Technology #Innovation #TechEcosystem #StartupEcosystem #Hiring #TechHiring If software engineering peace of mind is what you crave, Vention is your zen.

6

1 Comment

📣 Blackwell sets a new inference speed world record — A single NVIDIA DGX B200 server with eight #NVIDIABlackwell GPUs can generate over 1,000 tokens per second (TPS) per user on the Llama 4 Maverick model, the largest and most powerful model in the AI at Meta Llama 4 family. ⚡⏱️ Additionally, a system with eight Blackwell GPUs can also deliver up to 72,000 tokens/second in a maximum throughput scenario. 🏆 Blackwell is the first platform to achieve this model performance, demonstrating how it delivers the best combination of throughput, output speed and accuracy for LLM token generation. 🏎️🏁 See our tech blog for details.

3

In this episode, we discuss BlenderFusion: 3D-Grounded Visual Editing and Generative Compositing by Jiacheng Chen, Ramin Mehran, Xuhui Jia, Saining Xie, Sanghyun Woo. BlenderFusion is a generative visual compositing framework that enables scene synthesis by segmenting inputs into editable 3D elements, editing them in Blender, and recomposing them with a generative compositor. The compositor uses a fine-tuned diffusion model trained with source masking and object jittering strategies for flexible and disentangled scene manipulation. This approach achieves superior performance in complex 3D-grounded visual editing and compositing tasks compared to prior methods.

3

🦞 Today at NVIDIA GTC changed everything. Again. Jensen Huang didn't just talk about AI today. He brought Olaf on stage. A fully functional Olaf robot from Frozen walked out and held a conversation with Jensen — trained inside an NVIDIA simulation jointly developed with Disney. NVIDIA Blog The crowd lost it. And honestly? So did we. Disney's session at GTC is called "From the Screen to Reality via Physical AI" — and it's exactly what it sounds like. The translation of animated characters like Olaf requires balancing artistic fidelity with the constraints of the physical world, using GPU-accelerated simulation, deep reinforcement learning, and modular mechatronics to make it all work. Think about what that means for the parks. Live characters may still have a role — but robots that never get tired, never have an off day, and stay perfectly in character 24/7? That future is no longer science fiction. Walt Disney Imagineering's Kyle Laughlin said it plainly: "The BDX droids are just the beginning." OpenClaw — your personal AI agent, running locally, acting autonomously. NemoClaw — the enterprise-grade layer with security, guardrails, and scale. ClawCon — the community building real workflows on top of both while wearing lobster hats (*What if/when/how will this show up at the HumanX conference next month?) And then there's tokens. Tokens are the bane of my existence because nothing is as frustrating as being mid video generation and...running out of tokens. Jensen opened the keynote with a video about AI tokens and everything humanity is doing with them — from healthcare to space to construction. NVIDIA And he's right to lead with it. Because here's the truth nobody is saying loudly enough yet: Tokens are the new operating expense. They belong in your business budget, your project budget, and your hiring budget — right now. Every prompt, every agent loop, every autonomous task your team runs costs tokens. If you're not tracking and planning for token spend the way you track compute, software licenses, or headcount — you're already behind. At Lions Rest Studios, we're building token awareness into everything we do. Because the teams that treat tokens as infrastructure will out-execute the teams that treat them as an afterthought. So here's our question for you today: What is your Claw strategy and workflow — and is tokens a line item in your budget yet? 👇 #OpenClaw #NemoClaw #NVIDIAgtc #JensenHuang #Disney #PhysicalAI #AIAgents #Tokens #LionsRestStudios #FutureOfWork #AgenticAI p.s. My Kling render looks more like Baymax and less like Olaf - of course, the tokens are spent.

2

1 Comment

NVIDIA CEO Jensen Huang once remarked that every profession must now consider how to leverage AI. In the field of visual production, that shift is already tangible. Tasks that once required days of iteration can now produce refined results within seconds. The overwhelming speed of AI can easily be mistaken for a fundamental solution. In an environment where polished results are generated instantly, it becomes increasingly difficult to distinguish whether we are truly providing direction or simply labeling AI’s accidental output as “creative direction” after the fact. I don’t know—perhaps in the near future, the very criteria for judging creativity will change. This is not about what is right or wrong. However, the issue we can easily overlook here is that, regardless of the situation, the speed of the technology introduced to fill current system gaps far outpaces our own cognitive stamina. When a tool begins to overtake human thought, it becomes easy to lose sight of the fundamental question of what to create, captivated instead by the immediate brilliance of the output. Looking around, I can observe an interesting contrast born from this phenomenon. Projects with solid production processes tend to use AI strictly as an assistive tool, remaining cautious about fully integrating it into their pipelines. In contrast, environments where discipline-specific pipelines are less established, or where a few individuals must cover a broad range of tasks in a manpower-driven structure, often appear to embrace AI more rapidly to fill those structural gaps. When speed outpaces thought, selection can become a reaction rather than a judgment. Precision does not explain purpose, and high-quality rendering cannot substitute for clear strategic intent. Now that we have gained a powerful ally—and competitor—in AI, the responsibility to define purpose and context has only grown heavier. In the end, deciding where to steer the wheel still rests with us. And that judgment is not always as steady as we would like to believe. Are we using AI, or are we being driven by it? 🧐 #AI #GenerativeAI #VisualProduction

10

1 Comment

NVIDIA just snagged key tech and talent from Groq, the startup crushing inference speeds. Not a full takeover, more like licensing their killer IP and hiring the brains behind it, including founder Jonathan Ross and President Sunny Madra. Deal's worth about $20 billion in cash, NVIDIA's biggest ever. Groq's LPUs are beasts for AI output, running everything on-chip with SRAM for zero latency and predictable performance. Way more efficient than GPUs juggling memory bottlenecks. Perfect for real-time stuff like chatbots or massive data centers. NVIDIA dominates training, now they're bolting on Groq's inference edge to build end-to-end AI factories. Smart play against rivals like AMD or Cerebras. They structured it non-exclusive to skip antitrust headaches, closing fast without regulators slowing them down. Groq keeps operating independently with new CEO Simon Edwards, running their cloud service. Founded in 2016 by ex-Google folks, this validates their tech hard. But it tightens NVIDIA's grip on the market, could accelerate AI everywhere from voice assistants to autonomous cars. Exciting, though we need more competition to push boundaries. Official word: https://lnkd.in/daDFmWHf Reuters scoop: https://lnkd.in/dpx_b_pf Deep dive: https://lnkd.in/dcQ9cyfQ #NVIDIA #Groq #AI #Chips #FutureTech #Innovation

1

Niantic Spatial SDK v3.15 has been released, bringing significant mixed reality capabilities to Quest 3 and Quest 3S headsets. This update provides developers with advanced features, including a centimeter-accurate outdoor Visual Positioning System

1