Max Seybold

New post: The big stack game of LLM poker https://lnkd.in/d5mfqgwk I’m sure you read David Cahn’s excellent and provocative piece "AI's $600B Question", in which he argues that, given NVDIA’s projected Q4 2024 revenue run rate of $150B, the amount of AI revenue required to payback the enormous investment being made into NVDIA each year plus the electricity to power GPUs to train and run large language models is now $600B, and we are at least $500B in the hole on that payback. The numbers are certainly staggering… and are just going to get bigger. Unless the scaling hypothesis falters, this is a contest now of “not blinking first”. If you’re a big stack player like META, MSFT, GOOG, or any of the foundation model pure plays, you have no choice but to go all in — the prize and power of “winning” is too great. If you blink, you are left empty handed, watching someone else count your chips. It’s likely hundreds of billions will be destroyed, and trillions earned. For all of us in the startup ecosystem, among many things, it’s going to create a rolling thunder of AI opportunities. Read more here: https://lnkd.in/d5mfqgwk

149

10 Comments

This is just incredible. Neal Ghosh's demonstration of how #VentureIQ takes deeply technical and complicated content across a range of industrial use cases and ages/stages through an AI-generated conversational podcast (a la #notebookLM) is really mind-blowing. Check out his post, below, and if you haven't seen VentureIQ, yet, let's get you set up to learn more. You can visit VentureIQ.ai or DM any of us over here to book a demo!

2

🔥 🔥 🙌 🙌 🎆 🎆 I'm *fired up* to share this. #VentureIQ by the 9point8 Collective is officially live. We've heard from investors of every size and type that it takes too long and costs too much to evaluate deal flow; this limits the scale and speed at which these capital allocators can operate. To address this problem, we've created an innovative solution which delivers high fidelity diligence information near instantaneously. From a time and cost perspective, VentureIQ immediately creates leverage to look at more deals - and creates space to focus deeply on the most meaningful opportunities. To learn more, you can reach out to me directly or visit VentureIQ.ai. I'll close with some thoughts on my team, to whom I'm so grateful. Blair Merlino, Neal Ghosh and Evan Allen don't just talk about interesting concepts, point out problems or pontificate on what could be. Instead, they identify a problem, frame and validate the solution, and do the work to bring it to reality. VentureIQ is the best example I've seen of this, to date. #VentureCapital, #FamilyOffice, #PrivateEquity #CorporateDevelopment #Startups

9

2 Comments

Exciting to see startups aiming to use AI to discover chemicals, materials, drugs, etc. Semi companies have taken notice: AMD is investing in and Cadence is acquiring them. Semi/EDA companies have broken out by becoming critical path, whether it’s through critical software, packaging, IP, design kits from foundries, or user communities. Aside from just making it easy to make new materials, these “AI for making X” companies also need to articulate a vision as to how they can establish monopolistic positions in their industries.

144

5 Comments

Key takeaways from VC Goat - Vinod Khosla podcast. - The firm cares more about working on interesting problems with large technical solutions rather than just maximizing IRR. The team is there because they believe in the mission. - Khosla assumes they've lost the money the day they invest, and then maximizes for the upside opportunity. He calls it "option value investing" rather than IRR investing. - Khosla has contrarian bets in AI (neuro-symbolic computing, probabilistic approaches), biotech, robotics, crypto, and more. - He believes aviation fuels, fusion energy, new transit systems, and other contrarian areas ignored by most investors will be huge opportunities. - Most large innovations come from outsiders, not industry insiders. Khosla looks for founders who can learn a business rapidly rather than have deep domain expertise. The Future Impact of AI - AI will be deflationary and increase productivity growth to 4% annually vs the typical 2% forecast. This will cause great abundance but also increasing income inequality. - Bipedal robots will take over most manufacturing and manual labor jobs within 20-25 years. This will free humans to be more creative and pursue their passions. - Education will shift from job training to creativity. Uniquely human elements like taste and curation will be most valued in an AI-enabled world. https://lnkd.in/gqBsmbks

66

5 Comments

Worth reading: Situationsl Awareness by Leopold Aschenbrenner PDF : https://lnkd.in/gkxQ3yMD Web site: https://lnkd.in/gyJUCBhb Interesting perspective, time of course will tell.

2

The success of a Fund boils down to DPI (distributions back to LPs in relation to how much capital was deployed). The industry often gets infatuated with TVPI, which accounts for paper gains. However, these numbers are frequently not worth the paper they are printed on. Knowing when to stick or twist is always a challenge, so I favor a balanced approach: the third-third-third strategy. Sell a third of the investment at the first opportunity. Gradually sell the second third. Hold onto the final third for potential upside. Even this model needs adjustments based on market conditions, requiring a mix of aggressive selling or patience. The real issue for LPs is the industry's recent obsession with maximizing TVPI without enough attention to DPI. This can lead to portfolio imbalances and potentially disappointing (or zero) cash returns. 🎙 Listen to my full episode of Fintech Leaders with good friend Miguel Armaza for more on data as the new gold, secrets behind successful VCs, and the importance of founder-market fit: Spotify: https://spoti.fi/3KitB4b Apple: https://apple.co/4bXmmuC Youtube: https://bit.ly/3Vh4HIz

37

1 Comment

What do you do when a radical new technology puts your main product right in the crosshairs of disruption? Listen to David Okuniev — co-founder of Typeform | Ask awesomely — discuss the challenges of innovation within existing structures. David shared a game-changing insight: Radical innovation is really, really difficult to do inside your own product. He emphasized the need to break free from the constraints of familiarity and embrace change from outside the box. Henrik Werdelin and I have both seen our fair share of this in our respective careers. What struck us most was how David leveraged structure to overcome the innovator’s dilemma. By creating a culture of experimentation and providing space for bold ideas, he propelled Typeform beyond incremental improvements. What other hacks have you seen or employed to help your organization overcome the innovator’s dilemma? Share your stories below! 👇 And if you want to dive deeper into our conversation, click the link in the comments to catch the full podcast episode!

4

2 Comments

I recently had the honor of being featured on The Deal Scout. In this insightful 7-minute interview with Joshua Wilson, I shared early secrets and best practices from 35+ years in the Venture Capital industry that are essential to navigating the VC landscape and establishing a foothold in this dynamic field. You can listen to the interview on Joshua Wilson's post here 👇 #GPN #GuyPaulNohra #VentureCapital #Entrepreneur #Leadership #Philanthropy

10

⚡ Kicking off 🌎 Earth Day with exciting news as we unveil 8090 Industries' latest investment in Exowatt led by Hannan P. along with Sam Altman Andreessen Horowitz, Atomic, Overmatch Ventures, Protagonist, and many more. As the boom in AI heightens the energy demand for data centers, Exowatt is pioneering a fundamental breakthrough in baseload renewable energy that helps makes $0.01 per kilowatt-hour, 24 hours a day a reality for commercial and industrial use cases. 🏗 How does it work? The Exowatt P3, represents a breakthrough 3-in-1 modular system, a heat collector, a heat battery, and a heat engine capable of providing dispatchable power and heat 24 hours a day. Unlike traditional solar panels that convert sunlight into electricity directly, Exowatt uses a unique approach by storing solar energy in a thermal battery, which can retain this energy for up to 24 hours per day. The Exowatt P3 modules are designed to fit the space of a standard 40-foot shipping container. They can be deployed on small and large commercial and industrial projects, linearly scaling with workload size and infrastructure requirements. Best captured by Founder and CEO Hannan P. himself, an industrialist you're going to want to pay close attention to: “You don’t have to go back to fossil fuels to solve the data-center energy problem…That’s counterproductive and unlike traditional renewable solutions that require significant upfront costs and extended setup times, Exowatt's modular system can be deployed rapidly and cost-effectively – and it's available THIS YEAR. Exowatt is built to respond quickly to the escalating energy demands of the modern world, especially those spurred by the rapid growth of AI." 👀 Who Cares? Exowatt has a backlog of demand for OVER 500 megawatts for data centers across the U.S. and plans to begin deployments later this year. With this $20 million seed funding, Exowatt intends to expand its team and deploy the Exowatt P3 with its first set of data center customers. Stay tuned! Lots to come! cc Kerem Ozmen Wes Mendenhall Filip Vurdelja Haluk Sabanci Dincer Grant Brown Ishan Meswani Garuth A. Andrew Irwin Charles Oppenheimer Fatimah Bello #energyabundance #decarbonization #climatetech

80

2 Comments

Differentiation is crucial in capturing an investor's attention. Do not only create unique solutions but also offer a clear competitive edge. It’s about demonstrating why your solution is not only different but better. 🧐 #UniqueSolutions #InvestorInsights

3

From Crypto Ice Age to AI Inferno- Gartner Curves or Extreme Weather Conditions Precipitated by VC? A few thoughts I've been reflecting on recently on learnings from being through many crypto cycles I can apply to this moment in AI investing... https://lnkd.in/d6KSM3EN

29

2 Comments

Friends, Q2 was a quarter of progress and learning for us at HawkTower, where we invest in early-stage startups applying breakthrough computing technologies to maximize California's natural resource markets. Our focus is on technology-enabled interventions (e.g., AI-enabled geospatial radars for lithium-ion batteries in the waste stream) for critical, market-level imperatives (e.g., disaster prevention) that can drive rapid commercial adoption (e.g., placed on every recycling truck in the world). Our current investment priorities are: - Agricultural productivity and efficiency - Biomanufacturing - Labor and mobility automation - Waste stream optimization and decarbonization - Disaster prevention and management Here’s a snapshot of what we’ve been up to: Milestones and Investments: - In late April, we officially launched our fund, filing with the SEC. Yay. - By June, we completed our first fund close ahead of schedule, with backing from family offices, foundations, high-net-worth individuals, and organizations aligned with California's future. - Our portfolio now includes three early-stage investments (Farm-ng, Agriful Software, and Nexstera Tech). Strategic Partnerships: - We believe that startups operating closer to their customers and resources, especially in agriculture and biomanufacturing, have a strategic advantage. For instance, the Salinas Valley, a major agricultural hub, is less than an hour from Silicon Valley and serves as a center of thought leadership in the convergence of agriculture and AI (from Western Growers, Taylor Farms, and others). - Similarly, the North San Joaquin Valley is emerging as a biotech and agtech innovation hub, and we are proud to support BEAM Circular, a startup hub for the circular bioeconomy. Key Relationships: - Karen Warner, CEO of BEAM Circular, joined HawkTower as a strategic advisor. - Jackie Cruz, Ed. D. Cruz, Hartnell College’s VP of Advancement, advises on Salinas Valley and workforce development. - Ben Palone from Western Growers, with his technical background in agriculture, became a trusted advisor. - Matthew Hoffman from Driscoll's Global R&D team provides insights on global agriculture, product positioning, and investment strategy. - We collaborated with Monterey Bay DART for workforce development and automation in agtech. - Dominic Milano and the Milano Technical Group from Merced collaborated on key initiatives and advised us on investments. - HawkTower is participating in Santa Cruz Works's accelerator cohort as a mentor and advisor. Thank you Doug Erickson. - We published an overview of entrepreneurship across the UC system with input from across the University of California system. Moving into the fall, we remain deeply committed to supporting founders shaping California's future natural resource markets. Stay tuned and be in touch! : ) Db

95

5 Comments

The Qwen2 series from Alibaba Group focuses on efficient parameter usage in large language models, offering foundational and instruction-tuned models with parameters ranging from 0.5 billion to 72 billion. This series demonstrates improved performance over its predecessor, Qwen1.5, and competes effectively with proprietary models across a variety of benchmarks, including language understanding, generation, multilingual capabilities, coding, mathematics, and reasoning.   The flagship model, Qwen2-72B, achieves impressive benchmark scores: 84.2 on MMLU, 37.9 on GPQA, 64.6 on HumanEval, 89.5 on GSM8K, and 82.4 on BBH. Its instruction-tuned variant, Qwen2-72B-Instruct, also shows strong performance with scores of 9.1 on MT-Bench, 48.1 on Arena-Hard, and 35.7 on LiveCodeBench. The series supports around 30 languages, including widely spoken languages such as English, Chinese, and Spanish.   To promote accessibility and innovation, the model weights for Qwen2 are available on platforms like Hugging Face and ModelScope, along with additional resources on GitHub. Built on the Transformer architecture, the series features both dense models and a Mixture-of-Experts (MoE) model, employing advanced techniques such as Grouped Query Attention and Dual Chunk Attention to enhance performance and extend context capabilities. The pre-training phase utilized over 7 trillion tokens, significantly boosting its coding and mathematics skills.   Post-training processes included supervised fine-tuning and reinforcement learning from human feedback, aligning the models more closely with human preferences and enhancing their instruction-following capabilities. Evaluations have shown that Qwen2 outperforms various baseline models, including Mixtral-8x22B and Llama-3-70B, in both language capabilities and instruction-tuned functionalities.  Arxiv: https://lnkd.in/gjmyAK5C

4

1 Comment

🚀 Ready to launch your VC fund? VC Lab Cohort 17 has a new AI-powered mission control, LP mixers, new tracks, and more. Learn how we can help... 🛠️ AI-Powered Launch Suite + Decile Base: Your AI co-pilot + Smart LP targeting & network analysis + AI-powered thesis builder + Personalized outreach tools 🤝 Enhanced Support + Regional Demo Days + Global GP-LP mixers + Premium mentoring + Institutional-grade resources 🌌 Specialized Tracks + Space Technology + Later-Stage (Series A/B) + Family Office Funds + Enhanced Emerging Manager Program 👉 Perfect Timing + Exceptional deals at reasonable valuations + Billions in LP capital seeking emerging managers + Unique window as established firms face challenges The Early Admissions Deadline for Cohort 17 is Monday, November 18th. Join the 50+% of new VC firms we've launched in the last two years. 🚀

53

5 Comments

Why First-Time VC Managers Can Outperform Experienced Ones Institutional investors, and most of the Limited Partners (LPs) investors to the VC Funds, often prefer seasoned managers who have a proven track record of success and strong returns. While experience can be valuable, this principle doesn't always hold true in the VC industry. Experienced managers often operate based on mental models constructed from previous facts and information. While these models can provide valuable insights, they can also create flawed perceptions of new products or technologies. For instance, an experienced VC manager might dismiss an innovative idea by thinking, "It won't work because previously it didn't work." This reliance on past experiences can prevent them from recognizing groundbreaking opportunities. Before Elon Musk's success with Tesla, many investors were skeptical about electric cars. Venture capital investing is heavily influenced by mathematics and statistics, particularly the power law distribution. Success often depends on a small percentage of investments that yield significant returns. Therefore, having a well-diversified portfolio is crucial. A VC fund investing in, say, 40 startups can manage risk effectively because even if only 5% of those startups succeed, the returns can cover all fees and generate substantial profits. The returns might more higher if the fund operates in Emerging markets where operation costs for startups are low compared to more developed countries. Well-structured VC fund mitigates this risk through diversification and the statistical probability of success. While experience is valuable, it's not the sole determinant of success in venture capital investing. First-time managers can achieve results comparable to or even surpassing those of experienced managers if they avoid being limited by outdated mental models, employ sound statistical principles, and maintain a diversified portfolio also adding smart VC fund principles. Future Laboratory Ventures

29

2 Comments

🧩 One thing we've noticed with #VentureIQ technology is how well it can distill and communicate complex technical material 🧩 Whether it's a #deeptech venture, a novel #AI application, or even a patent submission, we can quickly distill the business and strategic implications for investors...and sometimes even the founders/PIs themselves! 🤯 Based on our work so far, the tech can handle a wide array of research areas and technological sectors -- anything from blockchain algorithms to vaccine development. It handles the nuance of the field, the specific terminology, and extracts the key concepts which matter for venture-building. 🤓 To demonstrate, I took a patent I helped write with Eric Laber on reinforcement learning. I ran it through #Ventureiq to build a commercial assessment (key differentiators, target markets). Then for fun I had #NotebookLM make a podcast out of it. I've uploaded the result it here (DM me if you'd like the actual report): https://lnkd.in/eiTejsny This is more than just automation or streamlining operations - it's about unlocking clarity and insight when it was previously unattainable, and the ability to do it over and over again no matter the sector or vertical. Interested in more info? DM me so we can chat.

11

2 Comments

Excited to share our portco Noble has completed its $15mn Series-A led by Paradigm with participation from Polychain Capital, Foresight Ventures, Wintermute, Informal Systems and renowned angels from Celestia Labs, @Skip Protocol, Succinct, Conduit and others Noble's mission is to accelerate #stablecoin adoption for everyone with an internet connection. By partnering with leading stablecoin & #RWA issuers like Circle, Ondo Finance, Hashnote, and Monerium, Noble currently manages over $500 million in assets and has facilitated over $5 billion in volume in less than a year! Noble began with a simple idea: to improve #UX of stablecoins for asset issuers navigating the complexities of a multichain environment. Today, Noble connects over 50 applications with a one-click #UX workflow for seamless stablecoin interactions. The platform has partnered with dozens of blockchains, including dYdX, @Osmosis, Celestia Labs, INJECTIVE, Neutron Foundation, MANTRA, THORChain amongst several others, to provide deep and reliable stablecoin liquidity for their applications There are 100+ #appchains in the modular ecosystem, leveraging solutions from Cosmos Network: Internet of Blockchains to Celestia Labs to Ethereum to Movement Labs & others. Noble's goal is to provide reliable stablecoin & RWA liquidity so developers can focus on building exceptional products without worrying about the underlying infrastructure. Noble is committed to developer sovereignty and delightful #UX. They want to see developers building sovereign #appchains succeed. The platform's innovative forwarding model, support for $USDC and other stablecoins for transaction fees, and lightning-fast block times on @Cosmos stack (sub one-second) deliver a seamless and efficient user experience. Noble's focus on seamless integration allows users to focus on what matters most: their applications. It is excited to introduce innovative products like the Noble Dollar, a collaboration with the M^0 Foundation, to bring users closer to the source of #stablecoin issuance. Noble's infrastructure is already making a significant impact on the crypto ecosystem. By providing reliable and accessible #stablecoin liquidity, Noble is helping to prove the viability of the sovereign #appchain model. In the near future, Noble will directly empower end-users with a suite of products, including swaps, forex, payments, and lending, to make it easier than ever to interact with the world of stablecoins More here: Fortune: https://lnkd.in/dQyAH6wX ChainCatcher: https://lnkd.in/daBufnPE

98

7 Comments