Arthur van Hoff

Mind-reading without surgery? It's happening. UCLA engineers just published a breakthrough in Nature Machine Intelligence that sounds like sci-fi but is very real. They created a wearable headset that reads brain signals (EEG) and uses AI vision to understand what you're trying to do. No surgery. No implants. Just a cap on your head. The results? Nearly 4x faster performance. The system helped people: - Control cursors on screens - Move robotic arms - Complete tasks way faster than without AI help And here's the most important part: it worked for both able-bodied people AND a paralyzed individual. Most brain-computer interfaces today require risky surgery to implant electrodes. This one doesn't. That's huge for accessibility and safety. The AI acts like a co-pilot, watching what you're doing and interpreting your brain signals in real-time to help you accomplish your goal. We're getting closer to a world where people with paralysis can interact with technology as naturally as we do—and maybe even better. The future of assistive technology is looking pretty incredible. Let me know if you'd like a link to the paper!

14

6 Comments

Persona is joining Luma AI! When Andrew Gavin and I started Persona, we set out to build something ambitious: state-of-the-art real-time avatars that felt alive, expressive, and autonomous in real-world conversations. We believed early that the future would be fundamentally multimodal, and that getting there wouldn’t come from narrow optimization. It would require tight integration across machine learning, engineering, product, and artistry — and the ability to build state-of-the-art technology with a small, deeply cross-disciplinary team. That approach allowed us to push the frontier of what was possible with real-time avatars. It’s work I’m deeply proud of. As we got to know Luma AI, it became clear that we’ve been working toward many of the same long-term goals — even if we started from different places. Luma has assembled an exceptional team that spans research, engineering, and artistry, and is doing some of the most impressive work in the field: building models that are competitive with the very best out there, and pursuing an audacious vision for what creative AI can become. We’ve been incredibly impressed not just by what Luma has already built, but by the ambition of that vision — and by the fact that they now have the resources, talent, and momentum to fully realize it. This feels less like an ending and more like a reunion — realizing there’s another community of people who share the same beliefs about how to build the future. It turns out we weren’t alone in believing this future was possible. More soon.

182

30 Comments

🚀 From Vector Search to Database Agents - Weaviate’s Journey to Powering Production AI 💡 Don't miss this deep dive with Weaviate CEO & Co-Founder, Bob van Luijt, on the evolution of vector databases into AI-Native Databases! 🧠 We discuss Weaviate's journey to powering production AI, covering: - Vector Search and RAG (Retrieval-Augmented Generation) use cases. - The exciting new wave of Database Agents (query & transformation). - Practical engineering topics like multi-tenancy and storage optimization. If you're building AI in production, this is a must-watch! Bob offers clear explanations, a short demo walkthrough, and key lessons from scaling an open-source database. 👉 https://lnkd.in/d5hvWfYD #AI #VectorDatabase #Weaviate #RAG #OpenSource #LLMs #Engineering #Database

4

1 Comment

Simaril (YC Spring 2026) is SOTA prompt injection defense for LLMs. This is the missing link for OpenClaw for Enterprise and all agents working on mission-critical data and workflows. The cofounders were the team that stopped billions of dollars worth of damages at Amazon.com and AWS. https://lnkd.in/gVKp7qbg

133

22 Comments

Most conversations these days focus on software-based *cognitive* automation—especially agentic workflows. It was refreshing to speak with Konstantinos Kandylas, CEO of Terra Robotics, about his experience building *physical* automation for agriculture. In today’s episode, Kostas shares the story behind the Laser Weeder: an AI-powered weeding solution that identifies and eliminates weeds with lasers, rather than herbicides. We cover how his team decides what to automate, AI hype, and common misunderstandings about automation in agriculture. Three takeaways: 🌾 Automate the bottleneck, not the process 💡 Focus on value creation, not just cost-cutting ⚖️ Prioritize by effort × customer impact 👉 Read the full interview here: https://lnkd.in/grSFGi9N 🔎 This episode is an early peek into my book, Designing Automated Futures, coming Spring 2026 with Rosenfeld Media. 🚀 Building agentic workflows or robots? Interested in being interviewed for an upcoming Substack interview series for my book? Reach out! #TerraRobotics #LaserWeeding #AgTech #automation #LaserWeeder

12

1 Comment

Jensen Huang just gave the clearest framework I've seen for understanding where AI actually lives. He calls it the Five-Layer Cake. 🍰 Energy → Chips → Infrastructure → Models → Applications Everyone's fighting over the top layer. The applications. The interfaces. The demos. But Jensen's point cuts deeper. The real bottleneck isn't the model. It's everything underneath it. China can build a hospital in a weekend. It takes the US 3 years to stand up a data center. That gap isn't a technology problem. It's an infrastructure problem. At gisax.io, this is exactly how we think about building AI systems for our clients. Anyone can wrap an API and call it an AI product. We're more interested in the data pipelines, the infrastructure decisions, and the architectural foundations that make those products actually work at scale six months after the demo. The frosting gets the attention. The layers underneath determine whether the whole thing holds. Which layer are you building in? (ai infrastructure artificial intelligence ai systems data infrastructure machine learning systems ai architecture scalable ai systems) Source : https://lnkd.in/gRdXR2Rz #Gisax #NVIDIA #AIInfrastructure #JensenHuang #FounderPerspective #AIStrategy #DeepTech #CustomSoftware

7

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

Ah, no, unless this model can be updated every week or so, this is a bad idea. It’s a cycle. Every time something new in the software arena comes out and gets a ton of interest, somebody wants to bake it into a chip and crow about how much faster it goes. Remember the Lisp chip? The Java chip? Transmeta? Don’t bet on hardware unless what’s being baked in there is done evolving.

6

5 Comments

Without standardization, web3 will remain like a Tower of Babel, where communication between chains is more akin to translation than transaction, argues Temujin Louie. https://lnkd.in/g8tuvUNS

7

Welcome Natan Linder, Co-Founder and CEO of Tulip Interfaces on this episode of 🔥The Lantern 🔥 Tulip Interfaces helps manufacturers build production systems that put people first with no code, low code, and other AI tools to turn operational knowledge into apps, workflows, and continuous improvement. 🔍 What happened in 2025 as the team heads into 2026? 🔎 Tulip wrapped their largest-ever “Operations Calling” conference with 700+ attendees to end 2025. Their platform is processing north of 620 million automations, 70 million application-based processes, and 35 million station hours. They're about to cross 30 billion contextualized data points! They also acquired the talented Akooda (acquired by Tulip) team, led by Yuval Gonczarowski, bringing enterprise AI search and deep learning expertise into their platform to help customers sift through massive operational datasets. In January, they announced a $120M fundraise led by Mitsubishi, that values Tulip at $1.3B. 🤔 Why does manufacturing need specialized AI? 🤔 Unlike software engineers who can freely experiment with their AI agents, real world operations require more rigor. Downtime and/or destroyed raw materials costs millions. Tulip is building AI copilots, composer products, and full agentic frameworks that are safe for the factory floor in regulated industries like pharma, medical devices, and aerospace. 🧠 What has Natan learned from building Tulip? 🧠 Work very close to customers. Natan believes a startup's focus should be split into 80% on customers, 10% on markets, and 10% for hallucinations. That customer obsession drove their Akooda (acquired by Tulip) acquisition. Operations leaders are knowledge workers too and need AI verticalized for their specific needs to democratize access to manufacturing expertise and elevate operations teams. 🤝 How can we help Tulip scale? 🤝 Tulip is in full sgrowth mode heading into 2026. If you're an engineer passionate about manufacturing, AI, or hard tech, or know world-class operators who want to transform how physical products get made, reach out! Boston is building the future of intelligent manufacturing and Tulip is leading the charge. The Lantern is brought to you by MGMT Boston, Rho, & Invest Northern Ireland – Americas this month

39

2 Comments

🚀 The Opus Post-Mortem: Decoding Claude 4.7 🧠 Is your AI feeling a bit sluggish lately? 🐢 If you noticed Claude 4.6 acting differently, you weren't imagining things. Our latest deep dive into the "Opus Post-Mortem" reveals the mechanics behind the phantom degradation of 4.6 and the massive leap forward with the 4.7 release. 📉➡️📈 The telemetry data for Opus 4.6 showed a significant 73% decrease in output character length, indicating a thinking collapse. This led to a critical 33.7% file reading failure rate and a 12x increase in required user interruptions. 📉❌ But here is the good news: Opus 4.7 is here to fix it all. 🛠️✨ Key highlights of the 4.7 upgrade: • Adaptive Default Thinking: Smarter allocation of reasoning power. 🧠✅ • X-High Effort Level: Exclusive high-performance mode for complex tasks. 💪🚀 • Biomolecular Reasoning: A staggering >2x jump in performance. 🧬🔥 • Production Tasks: A 3x multiplier metric for real-world reliability. 🏭🔝 While 4.7 introduces a 1.3x tokenizer tax due to upgraded text processing, the results in live environment tests are undeniable. From finance chart analysis to complex SaaS financial modeling, 4.7 consistently catches errors its predecessor missed and provides actionable, specific insights. 📊🏦 The 4.7 release isn't just a fix; it's a structural upgrade that restores functionality and returns deep thinking to your workflow. 🌟🙌 Full Video here: https://lnkd.in/dNVdHjUd Have you made the switch to 4.7 yet? Share your experience with the new reasoning capabilities below! 👇 ♻️ Repost to help your network understand why their AI just got a major brain upgrade. ➕ Follow [Deven Goratela](https://lnkd.in/dVt7VtDu) for the latest insights into staying ahead in the rapidly evolving world of AI and automation. #ClaudeAI #Anthropic #ClaudeOpus #GenerativeAI #PracticalAI #AIGrowth #TechInsights #Automation

1

#BlastYourCompanyRecord #ROI into #Space with🚀 Next AI CEO Brian Plain & www.NextAICompany.com's Space Highlights: Scientists Develop End‑to‑End Encryption for Git Services Story inspired by #ScienceX #staff, 10/07/25 🌌 Why It Matters From #Fortune500giants to #scrappystartups, from computer science students to #indiedevelopers, #gitservices are the beating heart of #modern software. They #storesourcecode, manage collaboration, and often safeguard the crown jewels of innovation: sensitive algorithms, #proprietaryarchitectures, & emerging #AImodels. Until now, the Achilles’ heel has been #security. Traditional git hosting relies on server‑side #protections, leaving #repositoriesvulnerabletobreaches, #insiderthreats, or #statelevelactors. 🔐 The Breakthrough #Researchershavenowdeveloped end‑to‑end encryption (#E2EE) for git services, a paradigm shift that ensures code is encrypted on the developer’s machine and remains encrypted in transit and at rest. #Zerotrustarchitecture: Even the hosting provider cannot read the code. Seamless developer experience: Encryption layers integrate directly into git workflows. #AIprotection: Proprietary machine learning models and training data can now be version‑controlled without fear of exposure. 🛰️ Brian Plain’s Space Highlight As the Next AI CEO, Brian Plain frames this as more than a #securityupgrade—it’s a #strategicinflectionpoint for the #AI #economy: “End‑to‑end encryption for git is not just about protecting code. It’s about protecting the future of intelligence itself. When your repositories hold the DNA of next‑generation AI, you need guarantees that no one—not even your host—can compromise it. This is sovereignty for developers.” 🌍 Strategic Implications #ForEnterprises: Safeguards trade secrets across global teams. #ForStartups: Levels the playing field by giving small teams enterprise‑grade security. #ForAcademia: Enables open collaboration without sacrificing intellectual property. #ForAIGovernance: Sets a precedent for how sensitive AI models should be stored, shared, and audited. 🌠 The Vision Ahead Plain's Next AI-Agentic Programming Team of #MasterAgents are position this as part of a broader “AI Trust Infrastructure”—a constellation of technologies (E2EE git, federated learning, secure multiparty computation) that will define #thenexteraofsafe, #scalable, and #sovereignartificialintelligence, with Next AI Company LLC, online at www.NextAICompany.com. ✨ Space Highlight Takeaway: End-to-end encrypted git isn’t just a technical achievement—it’s a vital safeguard for civilization. According to Brian Plain, it represents the kind of innovation that guarantees the #AIrevolution is #builtontrust, #resilience, and #unassailablesecurity. FULL ARTICLE: https://lnkd.in/ek9_jesj

2

Nvidia just released nemoclaw Link in comments. Their toolkit of enterprise safe claw bots. I will take a look and have a review, isolated sandboxes, NVIDIA compute, security policies. You can do most of this in docker and docker compose, but the benefit here is one click install. Enterprise level AI employees are not ready for prime time, but soon they might be. And when they are ready? 24/7 always on AI Employees.

8

3 Comments

Here's a lovely detailed walkthrough from Sathvik Divili showing how to build a complex voice AI workflow. The video covers the use case, theory of how to design multi-node conversations, the implementation, and a live demo. Sathvik digs into Pipecat Flows and why it's a useful library. But the general approach here is not specific to Pipecat/Pipecat Flows. The thoughts here will be useful to anyone working on "context engineering" for voice agents.

65

2 Comments

Building a Universal #AI protocol layer that solves - context / memory / cost and #LLM output quality. 1) WHY? Better results at lower costs, means quicker turnaround at scale 2) HOW? We have developed a set of Algorithms and patents that drastically improve how LLMs work at virtually every injection point of the process flow. 3) FOR WHOM? Business and startups that use LLM API's to process data, execute functions and who struggle with cost, quality, context and memory across large datasets and complex data processing. Stealth tech startup AI #claudecode #openai #grok

3

Rick Doten’s latest article explores how agents can subtly deviate from their mission—not by mistake, but by intent—a phenomenon called drift. His next paper will dive into how to detect and prevent this using an Ethical Framework, Sentinel-Council Architecture, and Knowledge Graphs. A must-read for anyone working with Agentic AI.

8

1 Comment

New: I spoke with the head of Claude Code, Boris Cherny, about how the viral AI coding tool is reshaping Anthropic. The buzz around Claude Code has reached a fever pitch recently, but some developers say AI coding products in general are reaching an inflection point. A few years ago, AI coding tools basically amounted to autocomplete, suggesting lines of code after developers started typing. Now, the space looks completely different. Developers can spin up AI agents from their phone that will spend hours coding up a feature. Just how good are these tools? Well, engineers inside of Anthropic seems to be using Claude Code for almost everything, and its business is growing like wildfire. - Claude Code made up ~12% of Anthropic's ARR by the end of last year, WIRED has learned. The product soared past $1 billion in ARR, by at least $100 million, in December, and could play a significant role in the company's revenue growth moving forward. - Cherny says nearly everyone at Anthropic, even its sales team, is using the agent for almost everything. That said, competition in AI coding is fierce. Cursor is at least as large as Claude Code, in terms of revenue. OpenAI and Google are hot on their tails. Definitely a space to watch in 2026. Read more in my newsletter, Model Behavior. https://lnkd.in/gCUzcjd2

213

9 Comments