Brad Smith

What I wish someone would say at the AI event in Paris: The political conditions for flourishing AI might not be just more data center investments or promises to be safer. Instead, they might be around creating conditions for rapid advancement and positive sum competition. Fortunately, Stanford statistician David Donoho laid out a framework for how to create those conditions: - data sharing - code sharing - competitive challenges implemented with frictionless open services I bet that the country that focuses on creating these conditions will be far more successful in AI with much smaller public investments. There is an opportunity for civic society to take Donoho’s paper Data Science at the Singularity as a starting point...

90

7 Comments

General Catalyst is leading The Interaction Company of California's $15M seed round, backing Marvin and Felix as they build Poke, a system for AI-native personal intelligence that carries memory across platforms, retains tone and nuance in comms, and anticipates next steps. This is a manifestation of what I mean when I talk about AI enabling a higher human purpose. Information and communications overload has come to define modern life. And most of our tools today actually reduce the value of communication, automating replies or filtering in a way that reduces our contextual understanding and creates more waste in our back and forth. The “this could’ve been an email” mindset only works when you can glean intent and meaning from your digital communication. Tools like The Interaction Co’s Poke are getting at the crux of the issue - reducing cognitive overhead to focus on impactful higher-level tasks - while maintaining the nuance, tone, and context essential for effective communication. This company is also a great example of the strength of our global seed platform, built to support founders no matter where they start or where they choose to build. Marvin and Felix first met with our team in Europe and, as their vision grew and they moved to the US, have now continued to build with Yuri, Neeraj, and Sho Sho. Welcome to the GC Famiglia, Marvin, Felix, and the entire Interaction Co team! Launch video below showing the more human side of Poke :) Read more here: https://lnkd.in/g-ySiwdu cc Jeannette Samuel

560

40 Comments

If All We Build Are LLMs, AI Progress Will Stall, Hit a Wall, and Fall Away (Credit: Martin Ciupa) Updates on LLMs from Gary Marcus… 1/ https://lnkd.in/gBv_t8wZ... 2/ https://lnkd.in/gS-_JN9V My take… Yann LeCun’s recent comments reflect what many top researchers are quietly acknowledging: large language models (LLMs) like ChatGPT or DeepSeek are powerful tools — but they capture only a thin slice of what true intelligence is. LLMs are masters of language, not meaning. They weave words with astonishing fluency, yet remain disconnected from the real world — from perception, intention, and consequence. They don’t understand what they’re saying, only how language statistically tends to unfold. In that sense, they’re mirrors of human expression rather than minds that grasp reality. As LeCun, Yoshua Bengio, Gary Marcus, and others have argued, building real, reliable artificial general intelligence (AGI) will require something very different — architectures designed not just to imitate thought, but to generate understanding. That means integrating: 1. World models that can predict and reason about real situations, not just text (LeCun’s Joint Embedding Predictive Architecture, 2022). 2. Autonomous learning that discovers causal structure instead of memorizing patterns. 3. Energy-based or modular systems that reason, plan, and act coherently within physical and ethical boundaries. 4. Embodied sentience and salience — systems grounded in sensory experience, capable of focusing on what truly matters and aligning ethically with human values. 5. Cognitive world models and evolutionary learning modules — hybrid systems that combine: • Common-sense reasoning about space, time, and agency, • Evolutionary and meta-learning algorithms that improve over generations of experience, and • Analog–digital integration layers that bridge symbolic reasoning with continuous perception. Together, these advances would give AI something LLMs entirely lack: a living, adaptive model of the world — capable of understanding, learning, and moral reasoning. Without these deeper ingredients, LLMs will remain syntactic but not semantic — brilliant compressors of language, not genuine thinkers. And when it comes to Artificial Superintelligence (ASI), the gap grows even wider. Safety and reliability demand systems that can explain themselves, stay grounded in context, and correct their own mistakes. LLMs hallucinate, lack self-awareness, and can’t verify truth — traits that make them fundamentally unsuitable as the core of safe, autonomous intelligence. In short, LLMs are the spreadsheets of cognition: quite useful, but not the essence of mind. The next breakthrough in AI won’t come from training on more tokens — it will come from discovering the deeper principles of how analog and digital processes integrate, how reasoning connects to the world, and how coherence emerges from interaction.

1,119

92 Comments

Can Startup Societies Rebuild Community IRL? Balaji Srinivasan argues that universities like Harvard, Texas A&M, and Stanford were early “nodes” of American society—training talent and shaping regions. Today, he’s pioneering “Network Schools” as 21st-century nodes: physical communities where like-minded people gather to build companies and lives together. Network Schools attract remote workers, visa-challenged builders, and ideological pioneers. • The motivation? Discontent with existing societies and a desire to start fresh. • Hundreds are already moving in; thousands have applied, showing a massive appetite for new models of community. #StartupSocieties #Crypto #RemoteWork #CommunityBuilding #Innovation #Founders #ThisWeekInStartups

40

6 Comments

AI Could Consume 99% of Power Resources by 2030— a warning by ex-Google CEO Eric Schmidt that we should not ignore. The 99% may be hyperbolic. But the message is clear: "AI’s energy demands are accelerating faster than our infrastructure can realistically support." Training frontier models consumes millions of kilowatt-hours. Inference, the part we use every day, could dwarf that over time. If we don’t invest in: → Chips optimized not only for performance, but efficiency → Cooling and packaging systems that rethink thermal limits → Power sources that are clean, distributed, and scalable → Models designed with power budgets as real constraints ...the compute race will outpace our capacity to power it, sustainably or otherwise. The future of AI must be compute-aware and energy-conscious. Because the most advanced models won’t be the ones that just run. They’ll be the ones that run efficiently, ethically, and at scale. #Venturecapital #AI #Deeptech #Startups Follow us at APEX Ventures and subscribe to our newsletter for exclusive content on groundbreaking Deep Tech startups: 🔗 https://t2m.io/EV2qHQuo

78

10 Comments

The power of distribution wins again 👑 Microsoft Copilot dominates in big enterprises. Smaller teams adopt emerging tools like Cursor, Claude AI & Zed. Cursor and Claude AI usage drops sharply as company size grows. Surprisingly, Google Gemini shows low, flat usage across all company sizes 🤔 The real innovation in AI dev tooling is happening outside big enterprises. If you're building the next-gen stack, your early adopters are in startups (and maybe medium) where there's less friction and more room to experiment.

44

13 Comments

🚀 Companies that innovate in physical industries move through three phases: research, development, and industrial scale. Research and development are critical, but at the end of the day, what truly matters is our ability to scale-to take innovation out of the lab and into the real world. That’s where impact happens. In defense, energy, manufacturing, construction, mining, and other foundational industries, scaling determines whether a breakthrough becomes a prototype or transforms entire sectors. At Eclipse, we’re focused on that last mile: industrial scale. It’s where our experience, time, and energy go, because scaling innovation is what drives real progress for the country and the world.

134

4 Comments

Yesterday, the open internet got its first toll booth. Not from courts or Congress, but from Cloudflare. The company, whose infrastructure touches roughly 20% of all global internet traffic, announced a significant shift: All new customer domains will now block AI bots by default. That means crawlers like OpenAI's GPTBot, Anthropic's ClaudeBot, and Google's Extended bot can no longer freely help themselves to your content. Unless you, the website owner, explicitly allow it, they’re cut off. Cloudflare also announced a Pay‑Per‑Crawl marketplace, where publishers can set terms for access - turning content into a licensed input, not a free good. With one product update, Cloudflare rewrote the default terms of engagement between AI and the web. From open to closed. From assumed permission to enforced consent. From passive scraping to active negotiation. Cloudflare has forced the question that’s been dodged for years: If AI systems are built on the backs of human expression, who owns the value they create? ▪️ For 20+ years, the internet operated on a basic exchange: Publish freely → Get discovered via search → Monetize attention. But AI broke that contract. Users now ask Claude or ChatGPT for answers. The models reply using human-created content - with no credit, no clickthrough, and no compensation. When publishing stops being a path to discovery - and starts being a donation to model training - incentives collapse. What Cloudflare did was reintroduce friction. Not to break the web. But to rebalance it. ▪️ What makes this moment so interesting is that the solution isn’t coming from regulators. It’s not the result of a moral awakening or a philosophical reckoning. It’s a default setting in a SaaS dashboard. But because of Cloudflare’s scale, it might as well be policy. This is capitalism at its best. You don’t always need top-down regulation to enforce guardrails. Sometimes market incentives create their own form of governance. In a world where governments often lag behind technology, infrastructure becomes policy. ▪️ Cloudflare may not have intended to reshape digital economics, but they’ve done exactly that. By forcing AI firms to ask, license, or walk away, Cloudflare has built the bones of a content licensing infrastructure, without ever saying the word “copyright”. ▪️ Let’s be clear: Cloudflare didn’t do this because it’s noble. They did it because it’s profitable. Their customers - TIME, Condé Nast, Reddit, News Corp - are tired of being strip-mined for training data without so much as a backlink. And Cloudflare is in the perfect position to intervene. I won’t be surprised if Fastly, Akamai, maybe even AWS follow. For 30 years, we assumed the internet had no gatekeepers. But it did. We just didn’t notice them until they started saying no. The principle is planted: Scraping without consent is not neutral. It’s extraction. We could see a more cohesive consent layer for the web emerge - driven not by altruism, but by incentive alignment.

6,187

409 Comments

The largest M&A in the history of #Quantum Computing?? There are many ways to build a quantum computer but the industry is mostly organizing itself around a few major approaches. For years, it seemed like Superconducting Qubits enjoyed most of the funding (and most of the drama), with vendors like IBM, Google and Rigetti Computing dropping announcements left and right. Yesterday marks a major paradigm shift - IonQ announced its pending acquisition of Oxford Ionics for a staggering amount of $1Bn. This is the second major consolidation event within the trapped ion modality, following on the 2021 merger of Honeywell Quantum Solutions and Cambridge Quantum into Quantinuum. Among global giants, Israel's Quantum Art also stands as a major player in this field (led by Tal David, Roee Ozeri, Amit Ben Kish & Elkana Ben-Sinai). How will the industry evolve from here?

46

2 Comments

🚀 Excited to continue our partnership with Quantum Machines , joining their $𝟭𝟳𝟬𝐌 𝐒𝐞𝐫𝐢𝐞𝐬 𝐂 𝐫𝐨𝐮𝐧𝐝 as they build the critical infrastructure for quantum computing. 🚀 At Red Dot Capital Partners, we invest in companies that provide the foundational technology for emerging industries. Quantum computing is reaching a point where scaling requires not just more powerful processors, but the right orchestration and control systems—this is exactly where Quantum Machines is leading the market. Congratulations to the amazing QM team - Itamar Sivan, Yonatan Cohen, Nissim Ofek, and a huge welcome to Rotem Shacham (PSG), and Intel Capital to the rocketship! Read more on TechCrunch: https://lnkd.in/ebkyB6RT #quantumcomputing #deeptech #venturecapital

224

9 Comments

What will it take to make foundation models 100x better - without spending billions on compute or turning the planet into a giant data center? This is one of the questions I explore with Rohan Taori, a researcher at Anthropic and co-creator of Alpaca - the project that demonstrated how fine-tuning @Meta’s LLaMA could achieve ChatGPT-level performance for under $600. Rohan shares his journey from early work in computer vision at UC Berkeley to his Ph.D. at Stanford, where he focused on making AI more accessible. He breaks down the key innovations behind Alpaca, including self-instruct, a method that uses a stronger model (OpenAI’s text-davinci-003) to generate synthetic data for fine-tuning a weaker one (LLaMA). We also discuss: The potential and limitations of synthetic data - how far can it take us, and where does it fall short? What’s needed to make foundation models 100x better - beyond just scaling compute The future of multimodal AI and the research directions he's most excited about Our full conversation here: https://lnkd.in/ghYaWYnn

54

2 Comments