Manu Thapar

Very interesting discussion about the future of human computer interaction powered by AI. I can’t wait to get my hands on one of these cool shades. #PoweredByAI

6

Task-specific models and routing architectures are the future. Robert Nishihara, co-founder and CEO of Anyscale, is seeing hybrid strategies emerge, where businesses are evaluating different types of models for different use cases. With the rapid advancement of AI tech, it won't be long before open-source models are "good enough" for the majority of business tasks. The natural progression from here is to develop smaller, task-focused models using proprietary company data. As Robert puts it: "You're going to need small, fast models. And task-specific models have a huge advantage when it comes to cost and speed because you can achieve the same quality with a smaller model if it's specialized." Another interesting development: the growing use of routing architectures. Imagine a classifier as a kind of traffic controller for your AI queries. It assesses the complexity of each query and directs it to the most suitable model: simpler requests are handled by smaller models, while complex ones are passed on to larger, more sophisticated models. “If you do it right,” Robert says, “you can achieve something like the cost of the smaller model with the quality of the more advanced model.” Dive deeper into AI architecture trends in my full discussion with Robert on the B2BaCEO podcast here: https://lnkd.in/ghigvERn

46

3 Comments

** This is not Financial Advice, so do your own due diligence ** $50B TAM for 1M GPU Clusters? legendary CEO Hock Tan of Broadcom believes so. I had the opportunity to listen to Hock Tan at Arista Networks 10 year IPO anniversary. Hock believes that for generative AI to truly develop AGI or ability to reason capability, it will need 1M GPU clusters and only 2 or 3 Hyperscalers or one or two sovereign customers have the capability to build such massive clusters. Hock Tan estimates the $50B TAM per 1M GPU clusters includes $30B for GPUs (assuming GPU ASP of $30K), $5B for Networking (interesting 6:1 ratio between GPU/compute to Networking), $15B for infrastructure like Power, including power generation, cooling, data center space etc. What plausible business case supports such as massive investment? remains to be seen. Hock admits its more of a moon shot type project but believes we are getting there. The biggest inhibitor to scaling GPU or compute capacity remains power. I think this has profound implications for investors. While power generation capacity takes long time to come online (7 - 8 years) and new data center buildouts take (2 - 3 years), the only way for Hyperscalers (like Amazon Web Services (AWS), Google and Microsoft) to bring more compute functionality they will have to upgrade their existing brownfield infrastructure. This should be positive for Intel Corporation, AMD and of course NVIDIA, as not only do we have the possibility of more accelerated compute deployments, but also the upgrade of non-AI compute infrastructure. Prior to the AI momentum, the upgrade cycle of traditional compute was stretching to 5 - 6 years, with most hyperscale CFOs pushing for extending the amortization schedule of the compute assets. Power limitations and growing demand for AI capabilities will force these hyperscalers to rethink their amortization schedule and thus I believe the pendulum swings in favor of infrastructure providers. Shorter upgrade cycles, more emphasis on higher, more power efficient technologies is great for networking such as Arista Networks and optics suppliers such as Coherent Corp., Lumentum etc. Interestingly, Hock Tan believes that as AI LLM models achieve AGI capability they will begin to generate their own data(thus the ability to reason) and will not require massive external data to be trained on, thus representing significant implications for wide-area bandwidth requirements. Would love to hear Bill Gartner's thoughts on this. Hock Tan also believes that power limitations will shift the balance in favor of customer ASICs vs. GPUs. While ASICs in general are more power efficient than GPUs, I think this outcome entirely depends on the maturity of the LLM models. In the past the hyperscalers optimized for $1/Gigabit of bandwidth, but now they need to optimize for power/Gigabit. This has profound implications for networking and optics supply chain. Would love to get your thoughts.

40

19 Comments

5️⃣ 🇪🇺 🚀 early stage rounds to pay attention to this week 👇 1️⃣ Dash0 🇩🇪 (Tech / Observability Platform) – Dash0 is redefining observability with its OpenTelemetry-native platform, providing transparent, predictable pricing and eliminating vendor lock-in for businesses aiming to monitor distributed systems efficiently. 💰 Raised $9.5 million in seed funding from Accel, Dig Ventures , and angel investors like Guy Podjarny and Guillermo Rauch. 2️⃣ Tranched 🇬🇧 (Fintech / Asset-Based Financing) – Tranched automates the securitization process using a blockchain-based platform, making asset-based financing more transparent and reducing traditional costs by up to 90%. 💰 Raised $1.96 million in seed funding from Blockwall, Kima Ventures, Speedinvest, and Andreessen Horowitz. 3️⃣ Pemo 🇦🇪 (Fintech / Spend Management for SMEs) – Pemo’s platform streamlines financial operations for SMEs across the MENAP region, integrating invoice payments, expense management, and corporate cards for real-time cash flow insights. 💰 Raised $7 million in a pre-Series A led by Augmentum Fintech and Shorooq. 4️⃣ Prelude 🇫🇷 (Tech / SMS Verification) – Prelude enhances user authentication by offering fraud-resistant SMS verification, reducing costs and improving security for businesses worldwide. 💰 Raised €7.6 million in seed funding led by Singular and Seedcamp. 5️⃣ In Parallel 🇫🇮 (Tech / AI Strategy Execution) – In Parallel’s AI-powered Intelligent Operating Model allows organizations to test, monitor, and adapt strategies in real-time, bridging the gap between strategy formulation and execution. 💰 Raised €2.8 million in seed funding led by Lifeline Ventures with participation from Notion Capital.

9

"A New Age of Intelligence: India’s Path to AI Leadership" The recent discussion between Mukesh Ambani and Jensen Huang on India’s AI future marks a transformative step for the country’s tech landscape. Their conversation outlined a vision for India as a global AI leader, with Reliance and NVIDIA collaborating to build robust AI infrastructure that democratizes access to intelligence. A key highlight was the role of open-source AI in creating an AI flywheel—a self-sustaining ecosystem fueled by domestic data, AI models, and a vast population of users. As Ambani noted, open-source frameworks like Meta’s LLaMA 3 empower developers to build on foundational models, reducing entry barriers and making AI accessible for all. This approach aligns with Ambani’s mission to replicate Jio’s success by delivering affordable, impactful AI solutions to the 1.4 billion people of India. Reflecting on Ambani’s vision, I see both opportunity and challenge for enterprises. As companies look to manage run costs, open-source AI appears promising. Yet the definition of “open source” in AI is still evolving and, unlike traditional software, requires much more than just source code. To truly “tinker” with AI models, developers need access to a complex ecosystem: trained models, preprocessing code, training data, model architecture, and more. This lack of standardization in open-source AI is both a strength and a hurdle, as enterprises must weigh flexibility against the practical costs of managing these resources. Despite these challenges, open-source AI and enterprise-grade AI together create powerful potential. With NVIDIA’s infrastructure and Reliance’s commitment to accessibility, India’s AI flywheel can drive sustainable growth across the economy, creating an ecosystem that not only supports enterprise innovation but elevates everyday lives. The ambition Ambani and Huang outlined is more than a blueprint for India; it’s a roadmap for making AI scalable, cost-effective, and inclusive on a global scale. Exciting times ahead for India's tech ecosystem! https://lnkd.in/gBBHjq6A #GenerativeAI #OpenSourceAI #AIFlywheel #DigitalTransformation #NVIDIA #Reliance #AIIndia

32

7 Comments

AI Reading for Saturday June 8! https://lnkd.in/epSAY9hn The California AI bill a lot of Big Tech and VC is so unhappy about. - FT More, including the ‘kill switch’. - Ars Technica The 'liberty for the powerful, screw liberty and fairness for the serfs' caucus is big mad, unsurprisingly. - Reason Google’s and Microsoft’s AI Chatbots Refuse to Say Who Won the 2020 US Election. - WIRED AI answers to election questions are wrong a lot - NBC Forbes says Perplexity summary pages are ripping off their articles. Forbes editor: it's not "rough", it's theft. - X (formerly Twitter) Is AI the greatest IP theft and enshittification in history? - BGR Is Sam Altman the ultimate personality hire/hype man, compared to Demis Hassabis who is sort of the ultimate scientist with stuff like AlphaFold and AlphaGo and DQNs? - Business Insider Microsoft is recalling Recall, making it an opt-in feature. - CNBC McKinsey talks to Reid Hoffman about AI as the new 'steam engine of the mind' - McKinsey & Company Are human-like voice AI agents the future of college recruiting and sales? - Fast Company The Dystopian Future of AI Warfare - YouTube Paleontologists and other scientists note inaccuracies in gen-AI art depictions of animals they study. Neil deGrasse Tyson: the stars in the sky are also wrong. - Gizmodo Iran uses AI to crack down on women. Don’t give Fox viewers any more bright ideas. - Fox News An open-source framework for reinforcement learning in trading - HackerNoon I'm here for the robot vs. human power slap league (parody) - YouTube Chinese company releases Sora competitor that looks pretty good. - Tom's Guide AGI accurately predicted (2016) - YouTube

4

🔍 Is NVIDIA's CEO Jensen Huang signaling something bigger by selling his shares? 💭 He recently sold over 4 million shares worth $500M, yet still holds 3.5% of the company. 🤔 We’ve seen other tech giants like Jeff Bezos and Elon Musk take similar actions before. Could this be a smart diversification move or something to watch closely? What’s your take? Let’s discuss! 🌍📈 #Leadership #Investing #NVIDIA #MiddleEast #TechStocks #Fintech #InvestmentStrategy #StockMarket #AI #BusinessLeadership #UAE #GCC #wealthface

9

Ahead of COMPUTEX 2024... NVIDIA founder and CEO Jensen Huang delivered a keynote address tonight with a focus on what's next for the AI ecosystem and what to expect from the NVIDIA pipeline. His two-hour long speech, including demo, was presented in front of a packed audience at the National Taiwan University auditorium. Jensen's address was live-streamed on YouTube and broadcast on TVBS HD, the leading local cable TV channel. The event upstages tomorrow's COMPUTEX opening keynote by Lisa Su, the CEO of AMD, and speeches by Pat Gelsinger, Intel CEO, Cristiano Amon, CEO of Qualcomm, among other tech luminaries. #nvidia #ai #products #demos #hardware #software

21

3 Comments

On April 18th, just last week, Meta unveiled Llama 3, lauded as "the most capable openly available LLM to date". This remarkable achievement by the Meta team comes hot on the heels of Llama 2's release last July. Notably, the new models surpass their predecessors and competing offerings from other providers in terms of performance. Meta's commitment to open sourcing these powerful models while prioritizing model safety and responsible usage is commendable. (https://lnkd.in/ghv3RJRz) Yesterday, Microsoft introduced Phi-3 Mini, a member of the Phi-3 family of models, promoted as "the most capable and cost-effective small language models available". According to Microsoft, the key to achieve high performance in such a small package lies in the quality of the training data. The model has been open sourced and is now available on Ollama and Hugging Face. (https://lnkd.in/gKgSmrBn)   These compact yet powerful models serve as invaluable tools for creators eager to bring their visions to life or simply experiment with cutting-edge technology, particularly for those with limited access to resources.   Shortly after the release of Llama 3, reports began surfacing of enthusiasts running the 8B model on a Raspberry Pi 5 equipped with just 8GB of RAM—a modest $80 single-board computer. (https://lnkd.in/gbeQPhQ2) While the performance understandably reflects the limitations of such a platform, the fact that the model runs and generates an output is nothing short of magical. Similarly, the Phi-3 Mini model can run on ubiquitous devices such as smartphones.   Reflecting on my own journey, I recall the pivotal role that early access to a low-cost computer played in shaping my passion for coding. The gift of an 8-bit Commodore 64 home computer ignited my curiosity at a young age. Later, with the guidance of an outstanding Computer Science teaching staff at my high school IISS Marconi-Hack Bari, I embarked on ambitious projects like a networked multiplayer version of Battleship written in Turbo Pascal and 8086 Assembly, complete with custom sprites and lo-res graphics. These formative experiences paved the way for my career, eventually leading me to various engineering and leadership roles.   It all began with a simple, low-cost, modestly capable computer used for experimentation—a gateway to discovering my life's passion.   I encourage educators and parents alike to consider setting up similar low-cost experimentation environments. By introducing younger generations to open-source AI technologies, we can nurture their creativity and help them uncover their true passions. It's not just a weekend project—it's an investment in their future and the future of innovation and discovery. #education #youth #AI #LLM #future #innovation

24

AI Boom Under Scrutiny: What’s Next for Tech Giants Like Nvidia and Broadcom? 🚨 The tech world is buzzing as Nvidia faces its worst single-day loss in market value, shedding $279 billion amid concerns about AI profitability and a DOJ antitrust probe. Meanwhile, Broadcom’s stock also took a hit ahead of its earnings report, despite its strong position in AI chip manufacturing and its $69 billion VMware acquisition. As businesses increasingly move towards private cloud solutions, could Broadcom’s strategy to diversify through AI and cloud infrastructure be the right move? 🌐 As we look at the growing skepticism around public cloud security from major players like Google and Meta, some important questions arise: 💡 Could this regulatory and market pressure signal a broader shift in how AI companies operate and compete? 💡 Will we see more businesses moving away from public cloud giants like AWS and Microsoft to private infrastructures? 💡 How might Broadcom’s diversified AI chip and cloud portfolio position it to outperform rivals in the long term? As we wait for the next big move in this space, it’s clear that tech companies need to rethink their strategies in the face of heightened regulatory scrutiny and evolving market demands. #AI #TechTrends #CloudComputing #Nvidia #Broadcom #VMware #Innovation #PrivateCloud #DataSecurity #Leadership #TechNews #BusinessGrowth Let’s discuss in the comments! What are your thoughts on where the AI and cloud industries are headed? 👇

8

Thoughts on #Nvidia Into Earnings: – Nvidia is 2% from its all-time high  – We know that Google, Meta, Amazon, Oracle, and CoreWeave are spending billions of dollars to buy GPUs - The return on capex spend is higher than the market thinks, so ‘hyperscalars’ will continue to spend (look at Cloud revenue growth) - countries like Saudi Arabia are starting to FOMO into data center spend  - The new meme is “intelligence cannot be outsourced’. On-prem is back: that helps Nvidia and hurts Microsoft and IT Consultants - Read through **Taiwan semiconductors revenues show 60% YOY growth** - #TSMC’s largest customers are #AAPL and then #NVDA - We know Apple isn’t selling many vision pros. (They have that hospice-type music on their earnings calls now…) - The inference is that Nvidia is selling every GPU that comes off the belt - If Jensen touches a CEO, that CEO drops a couple of billion. He is a master salesman selling to an easily identifiable customer set - Jensen has a posse that includes $DELL   - The steaks for winning the AI ‘space race’ are so high that the demand for GPU computing remains insatiable - I have no idea what the stock price will do. - If stock price drops like $META, which has largely closed that gap (did you get in there anon?), we are a buyer - If the price gaps up, we stand pat - Growth stocks that report have seen stock price declines.  Nvidia is the ‘pace-setter’ and is, therefore, an exception to the rule. This means that comps to Nvidia are causing these dips.  - Mr Market will buy dips on Nvidia. 2026 analyst estimates look low to me - We have titrated down our exposure to a number of semiconductors…except Nvidia and a few other dominant players - The relative value of Nvidia was the strongest in Mag 7 on Jan 1.  - Nvidia had gone largely flat the 2H of ‘23, and was a smoking buy on New Years - After nearly doubling the stock, that is no longer the case.  - However, it looks to us that Nvidia’s constraint is not demand.  They are a price setter.  - We remain long Nvidia into earnings - I expect Nvidia to deliver a beat/ beat / raise

1

Reasonable defense by Gov. Newsom of his veto of the controversial California AI bill “By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology. Smaller, specialized models may emerge as equally or even more dangerous than the models targeted by SB 1047 at the potential expense of curtailing the very innovation that fuels advancement in favor of the public good. Adaptability is critical as we race to regulate a technology still in its infancy. This will require a delicate balance. While well-intentioned, SB 1047 does not take into account whether an Al system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data. Instead, the bill applies stringent standards to even the most basic functions so long as a large system deploys it. I do not believe this is the best approach to protecting the public from real threats posed by the technology.”

2

This is a very intriguing idea: that a computer is no longer a tool, but an entire "factory". That it is able to create stuff all the time whether we are using it or not. But I will add something to that thought: as a "factory", computers today only generate data. It's what we call generative AI, such as images, audio, text, etc. But, if we apply that technology to a machine we create "generative action": physical movement with a specific purpose. For example, a robot that decides to clean up the kitchen regardless of whether we request it or not.

1

This is a very exciting effort and I think it has the potential to be a real "game changer" #foundationalmodels #AI #ML #pioneeringtechnology #metaverse

11

Some interesting research from Amdocs and NVIDIA. One thing to note is that the way a prompt is coded can impact operating costs. Cost optimization of generative AI applications is going to be a real issue. I expect a number of companies to jump on this. The methods for measuring and improving quality are also important. https://lnkd.in/g8DdTwSz performance-and-lowers-costs-with-nvidia-nim/ #generativeAI #costmanagement #innovation #BSS

1

This episode of Sequoia Capital's "Crucible Moments" series offers a fascinating look into MongoDB's journey from a startup to a cloud platform powerhouse. It takes us through the company's journey, from when our founder, wight Merriman, created a product in 2007 that developers adored and then overcame the challenges of scaling it, to the shift from on-premise software to Atlas, our developer data platform. Today, as we approach $2 billion in annual revenue, I'm proud to say that MongoDB has evolved into much more than a niche database. We're now a comprehensive data platform that puts its developers first and is constantly learning and evolving to meet the vision of our customers. Dive into the insights from Dwight Merriman, Dev Ittycheria, and Tom Killalea, taking you through the company's most critical moments and decisions. #MongoDB #cloudcomputing #developers

39

In the fast-paced world of Henerative AI, deploying Large Language Models (LLMs) into production presents a unique mix of challenges and opportunities. Irena Grabovitch-Zuyev and Suchita Venugopal from PagerDuty will share their journey of integrating LLMs into the product line within an ambitious two-month timeframe. 💡 We’ll Cover: ✅ Systematic Approach: Identifying LLM use cases, validating feasibility, engineering effective prompts, and developing a comprehensive testing strategy. ✅ LLM Service: A custom solution for secure and efficient LLM access, emphasizing robust security protocols, customer data protection, flexible LLM model switching, and redundancy for provider outages. ✅ Key Takeaways: Effective processes and strategies for rapid LLM deployment, highlighting non-negotiable aspects such as security measures and customer data protection. ✅ Lessons Learned: Honest discussion of the mistakes we encountered and the valuable lessons we learned to help you avoid similar pitfalls. This is a candid look at the challenges and triumphs of integrating generative AI at speed. #GenerativeAI #LLMs #ProductionDeployment #GenerativeAI #LLMDeployment

9

L1 industry record broken (again) by @Aptos on May 25 for the most daily transactions (154.4M/day) as reported by https://www.artemis.xyz/. This beats the previous record set (also by Aptos) on the day before (May 24 - see ⬇ ) by >59%. Same low fees, high success rate, no changes in failure rate, etc. Web3 is ready for large-scale Internet use cases. "Aptos made headlines in May, with Tapos achieving over 323 million transactions over the course of their three-day campaign. This campaign drove Aptos to the top of the Artemis L1 dashboard in terms of transactions per day and set a record for transactions in a single day by chains we track." - Artemis https://lnkd.in/gByu55B6

112

4 Comments

How Startups Grow 👇 1. How Linktree got from 0 to $37M ARR and 12M users with the simplest product ever: https://lnkd.in/dET-g-Gd 2. How Rubrik's CEO hacked growth to achieve +$500M ARR in 10 years: https://lnkd.in/djKajRDj 3. How Perplexity AI got to +$1B valuation with $20M ARR under Google's Monopoly: https://lnkd.in/dzRXWYw3 4. $100M ARR in 18 months: how Wiz became the fastest-growing software company ever: https://lnkd.in/dyHk2Pda

32

6 Comments