Tanya Dua’s Post

"What made us different was perspective; we saw the world differently.” 👀 This is how Jensen Huang characterized NVIDIA's roaring success during this morning's keynote at #TiEcon2024 in Santa Clara with Mayfield Fund investor Navin Chaddha. Rather than merely building graphics chips like its competitors, Nvidia sought to build an ecosystem and engage its community — game developers, scientists and other stakeholders actually building applications using its technology. This collaborative "co-design" approach has ensured that its products can be seamlessly integrated into the real-world, priming it to dominate the ongoing generative AI boom. [Nvidia's GPUs, which help power compute-intensive AI tasks, can be integrated across the board, from Azure to GCP to AWS.] Other key takeaways from Huang's talk: 📌 On the importance of leading from first principles and pivoting when necessary: "I test my assumptions every single day. If facts change, your mind should change." 📌 On AI being the latest industrial revolution: "We're now in the fourth industrial revolution, and we're going to produce intelligence at scale. This way of doing computing is radically different than the past." 📌 On sovereign AI, or every country needing to develop its own AI strategy: "Historically, we've always looked at the natural resources — what's underneath the earth — that belongs to a country. But in this new world, your society's data also belongs to your country." 📌 On what keeps him going: "There's a superpower in ignorance. There's a superpower in not knowing how hard something is. There's a superpower in not believing that something is impossible. The more we know, the more we realize how hard something is. And so I love that feeling of thinking, how hard can it be? That's what inspires you, right? All these hard problems we're looking at, my first inclination is, well, 'how hard can it be?'" Stay tuned for the full download from TiE Silicon Valley in my newsletter! 🖋

🚨Norwest Venture Partners just closed its latest fund at $3 billion. Partner Scott Beechuk joins us for VC Wednesdays to discuss what he’s looking for, and more.🚨 🖋️ What’s the top thing you look for in an early-stage startup? How does that differ from the later stages? With early-stage startups, especially pre-revenue, my lens tends to be more focused on the product, because if you don't have a great product, it's hard to prop that up with good sales and marketing. In the later stages, I look for the ability of the company to continue to scale horizontally. For example, we have a company called Gong in our portfolio that started with a revenue intelligence product, but has now expanded into a whole platform of related products, and that's enabling them to scale in both vertically and horizontally. 🖋️ Norwest just raised a new $3 billion fund. What’s the biggest focus area for that? Where we are today with AI reminds me a lot of the early days of the internet, where there’s a whole new world out there that we can build on. AI has been around for decades, but never at this level of scale, sophistication and quality. So this new fund is an opportunity for us to invest in a new generation of companies born in the AI era, like Amazon and eBay back then. I’m really excited that we have fuel in the tank to be able to capitalize on that. 🖋️ Where do you think the biggest opportunity in AI lies? I don't think there's anything better than AI to transform a lot of front office things, whether it's sales or marketing or customer engagement. As a firm, we’re also investing in areas like security, data privacy, compliance and governance — which have taken on a whole new meaning with AI. How companies handle these will be equally, if not even more, important than AI applications themselves. We look for startups with deep technology stacks that are defensible, and would be difficult to replicate for any company of any size. 🖋️ What are some recent examples from your portfolio? One of my most recent investments is in a company called Lavender 💜,  which uses over 300 psychology heuristics coded into its product to coach salespeople on emails and improve successful outreaches. Another is Connect The Dots, which has used AI to build professional relationship graphs between people to enable warm introductions. The magic is all in the back-end, because something like this wouldn’t have been possible five years ago, since we didn’t have the compute, storage or cost structure to be able to process it. 🖋️ What areas within AI are overhyped? As much as I love the idea of AI agents, they are quickly becoming overhyped. Most companies are just characterizing AI assistants as AI agents, and that's wrong. AI agents are self-aware, they have memory and context, and they interact autonomously with other agents. We’re probably a year away from the first set of AI agents to really hit the market. #VCWednesdays #vc #venturecapital #startups #TechonLinkedIn

I'll be in the Bay Area for TiE Silicon Valley's conference this week, moderating a panel with the inimitable Navin Chaddha on Friday. Come and say 👋 , and check out the full agenda here (https://tiecon.org/agenda/) I'd also love to meet any founders and investors in/around SF when I'm not on stage on Wednesday. DM me and we'll make it happen!

This week, we released LinkedIn’s 2024 Top Companies list, a data-backed ranking of the 50 best employers for growing a career in the U.S. Like in the years past, these companies are leading the way in both attracting and retaining workers. But what's different about this year? That many of them, including Amazon, EY, KPMG, Mastercard and Verizon are prioritizing hiring for #AI roles over most others. 🤖 What are the top AI roles that these companies are on the lookout for? Read this week's edition of LinkedIn News Tech Stack for more. Thank you to Christina Rozsa, Ginnie Carlier, Stephen Chase, Shankar Arumugavelu, Rohit Chauhan and Steven Armato for sharing your thoughts with me! What AI roles is your company hiring for? Share more in the comments below! (And subscribe to Tech Stack for future editions: lnkd.in/TechStack

🚨 A founder-turned-investor, Daniel Kimerling invests at the intersection of finance and technology. He joins us for VC Wednesdays. 🚨 ✒️ Fintech funding has dipped to a four-year low. Are you still bullish on it? I'm bullish on it for the simple reason that financial services represent something on the order of 20% of global GDP in some countries. Yet consumers hate their banks, insurance companies and wealth managers. There are very few industries that have that combination of scale and derision, so a very ripe environment for entrepreneurs to come in and solve real problems for consumers, small businesses and enterprises. ✒️ What are your specific focus areas for 2024? The population of the U.S. is rapidly aging, and that creates a big opportunity for products and services that help the elderly, their caregivers and family members. For example, we’ve invested in True Link Financial, Inc. that offers investment accounts and prepaid cards for seniors. Another is financial services and climate change, where the opportunity is in places like mortgages and insurance. We’ve invested in a company called GlacierGrid, which is focused on helping reduce the cost of energy for small and medium businesses. ✒️ Is AI becoming a focus too? We're less enthusiastic about it than most, because when you think through where the value will be accrued, we're not optimistic that the returns will flow to investors. There are a couple of reasons why. First, there will be rapid commodification of many of the LLMs. Second, many of the models will ultimately be open-source. And third, much of the value is actually being accrued in hardware, either through the purchasing of GPUs from NVIDIA and their peers, or the renting of timeshare infrastructure like Azure, AWS and Google Cloud. So we're very skeptical about investing in the software layer. ✒️ What about crypto, now that it’s seeing a bit of a resurgence? We've been closely following the space since 2017, and ultimately we’ve not been able to satisfactorily answer some basic questions: Who does this help? Why does it help them? And how do you make money investing in it? As an investor, you can make money in crypto if you operate like a hedge fund, in the sense you're buying and selling tokens and speculating on various tokens, chains and protocols. The problem with it is that we are builders and investors, not speculators. ✒️ What’s the No. 1 mistake that early-stage startups make? The most common mistake is to hold financing existential, rather than a source of optionality and value creation. The orthodox playbook is to raise money and spend it like hell, so in 12 to 24 months you have to sing for your supper again. So if for some reason you can't raise that next chunk of cash, you die. My advice is to get profitable on the cash you have, so that you can hang out in safe harbor for as long as you want and don't need to go back and raise more money. #VCWednesdays #vc #venturecapital #startups #TechonLinkedIn

🚨 The youngest partner in the history of the firm, Bessemer Venture Partners' Talia Goldberg invests across consumer internet, software and AI. She joins us for VC Wednesdays. 🚨 ✒️ Have consumer internet investments taken a backseat with AI? AI, if anything, has made it more of a focus. AI enables radically different consumer experiences that are fun, different and are changing the way that we interface with technology, so we're seeing a lot of activity that's either consumer or prosumer in nature. For example, we've invested in Perplexity, which is rethinking search and information discovery and knowledge. Or Character.AI, which is naturally viral. ✒️ What do you look for in an early-stage startup? I place a real premium on speed — teams and founders that have demonstrated they can move fast and ship products quickly. In the early stages, compounding speed with feedback loops from customers and product development is one of the biggest advantages. The more evidence I see of a company iterating and shipping really quickly, and relentlessly focused on execution, the more excited I get. ✒️ What are your biggest focus areas right now? In 2022, I wrote a blog post about how there was a new opportunity in search, and we were looking for new models in it. So, of course, when we see a company like Perplexity, that's both shipping fast and aligned with that, our conviction is higher. We also have a history of investing in vertical software, and we’re excited about how AI will create new opportunities there. The focus is on what we call "bilingual teams" that are both deeply technical and have a strong commercial instinct. They're demonstrating real ROI or creating meaningfully different, 10x better, cheaper, or faster experiences. ✒️ What would you like to see more AI startups tackle? I doubt that a few years from now, prompting chat interfaces will be the dominant form of interacting with AI. I'm excited to see companies innovating there, particularly a voice-first model or an ambient presence where AI is much more proactive in our lives, versus reactive. There’s potential as agent architectures and technologies improve. Apple is the dark horse. They haven't done that much publicly yet, but they're so well-positioned, in part because they sit on an incredible data set. But they also own the device and have their own chips, which is quite powerful, and can allow for much lower latency and higher performance. ✒️ You are the youngest elected partner in firm history. What’s your advice for those starting out? Think independently. A lot of the best investments aren't obvious early on. Come to your own conclusions, and don't be afraid to bet on them. One of the best things about Silicon Valley is that failure isn't looked down upon. That's what venture is all about, and risk is where the great returns come from. We tend to overestimate risks and underestimate the rewards. #VCWednesdays #vc #venturecapital #startups #TechonLinkedIn

When Jensen Huang talks, everyone listens. About a year ago, the NVIDIA CEO declared that generative AI was ushering in a “new computing era" where "everyone can be a programmer." He may have been onto something. GitHub’s Copilot may have been the first AI coding assistant on the block, but AI-powered coding assistants have mushroomed in recent months. A prominent one is Khosla Ventures and Andreessen Horowitz-backed Replit, which just unveiled a product that allows developers to collaborate on projects in real-time (https://lnkd.in/ge8aTtk4), accompanied by an AI agent that automatically fixes any errors (https://lnkd.in/gXE6E6AC). These coding assistants may have made coding easier, but have likely also contributed to choice paralysis — with a plethora of options available, which options are worth trying out? Further, which tools are actually useful? 📣 Computer programmers and coders, I'd love to hear your thoughts.

🚨 With 16 years of experience under his belt, Madrona’s Karan Mehandru invests in SaaS, data infrastructure, security, cloud computing and AI startups across stages. He joins us for VC Wednesdays. 🚨 ✒️ What’s the biggest lesson you learned as an operator that you carry with you as an investor? I sold software for a while. It was deliberate, because the more I looked around, the more I realized how much selling is part of the job — both for investors and entrepreneurs. At the end of the day, you're either selling yourself, your product or your services. The best salespeople are empathetic and understand customer needs. I’d recommend everybody do some selling at some point in life, as it teaches you empathy — one of the traits that I hold sacred. ✒️ What’s the #1 thing you look for in a startup? For our core fund (Seed and Series A), it's founder-market fit, and figuring out the applicability of the founder’s skills in the market they're going after. In our acceleration fund (Series A to C), it’s go-to-market fit. Most startups are product-led, not sales- or marketing-driven. That's the lens I bring: Let's assume what you’re trying to build is built. How do you package it? How do you price it? How do you market it? How do you position it? Who's the buyer? ✒️ AI has become a bigger focus area across the board. What’s your thesis there? If the printing press was content replication at scale, the internet was content distribution at scale, and mobile was content consumption at scale, AI is content creation and synthesis at scale. It’s the biggest shift, and touches every layer of the stack. Last year, I was interested in the middleware layer with investments in unstructured.io, Deepgram and Clerk.io, because no applications could go to production until these things were solved. This year, we’ve hit a tipping point where the underlying infrastructure is conducive for applications that have real durability. That will unlock a massive tsunami of applications. So we’ve made investments like Typeface, Read AI and Rhythms. ✒️ How are you thinking about the implications of AI’s rise on security? It raises some really important questions, like how do we protect the data that's going in and out of these systems? What is the system pinging to give you the results back? The way I think about it is that if the bar for content creation is lower than ever, then the need for compliance and adherence is higher than ever. There's a net-new market being created around LLMs and security. ✒️ What areas within AI are overhyped? Conventional wisdom is that you're going to have vertical agents for travel, finance and so on. But my view is that there's going to be a few amazing horizontal AI agents that become your copilots for all tasks, or allow others to build interesting applications leveraging those agents. So it's a little bit of a middleware play, but it's also an end-user horizontal agent platform. #VCWednesdays #vc #venturecapital #startups #TechonLinkedIn

🚨 In the latest edition of my weekly newsletter Tech Stack on LinkedIn News, I explored the concept of 'sovereign AI.' 🚨 While not entirely new, sovereign AI has gained steam in recent months as countries across the globe race to harness AI’s potential to boost their national security, socioeconomic and cultural interests. 🤖 It was absolutely wild imagining what geopolitics could look like in the AI era. TL;DR: Experts like Pablo Chavez and Prof. Ahmed Banafa think countries are on the path of an AI arms race, not unlike the nuclear arms race of the past. 🚀 🤯 This story really took me back to International Relations and History classes in my undergrad days at National University of Singapore. 👋 Terence Lee and Dr. Malcolm Murfett Curious to hear your thoughts. Read more below and drop what you think in the comments!

🚨A serial entrepreneur-turned-investor, Mayfield Fund’s Navin Chaddha has backed 18 companies that have gone public, including Lyft and developer-tools company HashiCorp. He joins us for VC Wednesdays.🚨 ✒️ What were your biggest lessons from being an entrepreneur that you still apply as an investor? It's all about the people. You need to make sure that everyone you hire is aligned on your mission, vision and values. It's a team sport. Secondly, company-building is a marathon, not a sprint. There's no overnight success. It takes time, and you have to pace yourself accordingly. Third: It’s very important to be customer-first, and your solution needs to be a must-have, not a nice-to-have. Sell painkillers, not vitamins. Fourth — and this was my issue as a first-time founder — focus. You can’t be the jack of all trades. Finally, you need to be agile, because dinosaurs don't survive. ✒️ What’s your biggest focus area right now? I’m focused on how AI, or what we’re calling the ‘cognition era,’ can elevate us. You had IaaS for infrastructure, PaaS for developers and SaaS for business. Now one layer up, the AI revolution is ‘cognition as a service.’ We’re interested in companies solving for the 5 As: how AI helps humans automate tasks; how it accelerates our productivity; how it amplifies creativity; how it augments our capabilities; and finally, how it helps us become superhumans. ✒️ What’s a recent example that fits this thesis? We invested in a company called Sema4.ai, which provides a platform — middleware and tooling that sits on the data layer — on which businesses can build solutions to create ‘digital twins.’ It can be applied to industries from travel to banking. For example, a bank can use it to create a fraud twin for a fraud manager, helping them detect, troubleshoot and remediate the issue. ✒️ Won’t this impact jobs? Some jobs may be impacted, but it'll create net-new jobs. The areas we’re backing actually have shortages. DevOps, SecOps and nursing all have shortages of tens of thousands of people — where digital twins can step in. You need humans in the loop to manage them. From a product management perspective, humans will think about how the product needs to be priced and positioned. We'll get to do things we’re good at, rather than mundane day-to-day tasks. We’re smart, and we must remember who’s the jockey and who’s the horse — the horse is AI. ✒️ What areas within AI are overhyped? The most overhyped area is going after semiconductors and competing with NVIDIA. It’s hard because of their market share and scale — the lead they have on the technology, control and distribution. They have a lock on the market with the CUDA software platform to build AI apps, they’ve pre-bought manufacturing capabilities at TSMC and they have a backlog of demand from the cloud providers. The same is true with LLMs. For me, the opportunity is in the data layer, middleware and tooling. #VCWednesdays #vc #venturecapital #startups #TechonLinkedIn