Jay Chia

AI memory is fundamental for conversational AI and AI agents. It is critical for building personalized experiences and improving reasoning. Check out Letta’s course to understand how to use AI memory to enhance the customization and quality of your AI app.

15

🚨🚨 🚨 Unlocking Innovation: The Power of the Adjacent Possible Innovation programs can stall when the team tries to ideate way beyond their collective perspective and frame of reference. Like a software engineer brainstorming genetic modification of mycelia, or a senior executive ideating process improvements which they haven't been participated in for decades. It's challenging, exhausting, and ultimately self-defeating. Enter the concept of the "Adjacent Possible" - a game-changer for innovation. What is it? The Adjacent Possible, coined by Stuart Kauffman, represents the realm of new ideas within reach based on our current knowledge, resources, and capabilities. Why it matters: ➡️ Realistic Innovation: it focuses on achievable advancements, not far-fetched concepts. ➡️ Incremental Progress: each innovation opens doors to new possibilities. How to leverage it: • Constantly explore the boundaries of your adjacent possible. What's in your line of sight is what you know best. • Combine existing ideas in novel ways. Even small perturbations from status quo can generate large innovation leaps. • Stay curious about developments in adjacent industries. Cross-domain patterns create fresh sources of ideas, while also expanding the adjacent possible. Most of all, it takes the 'genius' pressure off the team. No one knows everything, but each one of us knows a lot about something, and we can all drive plenty of innovation potential by focusing on the spaces we know best. Are you exploring your Adjacent Possible? Share your thoughts below! 👇 #Innovation #innovationsystems #AdjacentPossible

8

2 Comments

Attached is NVDA's price chart for the last year - up c.175% (up 2500%+ in L5Y period!!) 🚀 I nonetheless have conviction that it currently remains undervalued... Among other things, due to strong competitive positioning as a GPU-maker, AI-driven demand fuelled by investor and consumer interest, and moats it has built. This is mirrored by equity analysts – GS’ latest price target was $800, BofA’s $1,500. I do believe the growth is capped, and driven in equal measure by under-supply and bubble-like demand as it is by fundamentals. Even so, Nvidia remains an attractive medium term investment. ⚖️ Demand – Supply Dynamics: AI is a generational innovation, and agents /apps are already upending SaaS, or at least so think VCs who deployed $64.6B in AI start-ups in 2023 alone. As early movers and now the industry standard (c.80% market share), Nvidia's A100 and H100 chips are not only the most efficient option, but also crucial for large scale AI training and inference (chips are the biggest cost head for LLM companies). With unprecedented interest in AI (ChatGPT acquired 1M users in 5 days of launch, OpenAI reported $3.4B ARR in Apr’24, AI agents are widely recognised as the tech of the future), Nvidia stands to capture a large portion of the $250-$500B data centre conversion market over the next 5 years​. This is proven by the $22.6B in data centre revenue Nvidia brought in during Q2’24, up 23% from the previous quarter and up 427% from a year ago 🥇Competitive positioning: While Intel, AMD, Arm have been active, the only credible competitor, AMD, has a relatively smaller market share in the high-performance AI GPU segment, and Nvidia’s consistent technological advancements keep it further ahead, e.g., Nvidia’s Rubin which will succeed Blackwell in 2026 is orders of magnitude better than anything on-market. 🤝Moats via ecosystem building: Collaborations with hyperscaler cloud providers like AWS and Microsoft Azure have expanded Nvidia’s reach. Moreover, acquisitions like Mellanox have strengthened its position in high-performance networking, critical to AI infrastructure. Coreweave, a prominent cloud services partner / distribution channel for chip capacity is an NVIDIA investee and is built almost entirely on Nvidia chips. Finally, they continue investing in networking, software tools, data centre solutions but also purpose-built AI PC and smartphones to create an ecosystem that not only locks in customers and enhances switching costs, but pivots more of their revenue to more predictable recurring contracts. 📈 Financial underpinnings: Nvidia’s expected FCF for next 2 years stands at $100B, providing huge ammo for R&D, strategic investments and M&A (only $30 to $35B earmarked for buybacks) and in turn, growth. Even so, their shares trade at a P/E ratio of c.20x  NTM earnings, far below historical median of 35x+, indicating the market expects them to grow in the medium term. KEEN TO HEAR THOUGHTS! #Nvidia #AI #Compute #Chips #Tech #Growth

10

1 Comment

⌛ The deadline for YC’s S24 batch is April 22! I get a ton of requests for application feedback, so I’ve consolidated my advice here: 𝐓𝐞𝐚𝐦 > 𝐢𝐝𝐞𝐚 + 𝐩𝐫𝐨𝐠𝐫𝐞𝐬𝐬 Your idea will change; that’s what it means to be pre-product-market fit. A bet on an early-stage company is a bet on the team to iterate rapidly and figure it out; think of your idea and your progress as useful evidence. Don’t be discouraged if you feel like you haven’t made enough progress, because the amount of progress you’ve made is not nearly as important as how fast you made it. In other words: “slope, not intercept”. 𝐁𝐞 𝐬𝐢𝐦𝐩𝐥𝐞. 𝐁𝐞 𝐝𝐢𝐫𝐞𝐜𝐭. Explain your idea clearly in 1-2 sentences: who are your users and what problem do you solve? Avoid jargon, marketing-speak, technical intricacies, or an exhaustive list of features. Tacking on “...with AI” or “...with blockchain” is not any more helpful than “...with computers”. Bad: “We are transforming the global software development industry with a revolutionary AI platform that streamlines the SDLC for tremendous efficiency gains using cutting-edge LLMs.” Better: “We’re a developer tool that automatically reviews code and fixes bugs.” 𝐊𝐧𝐨𝐰 𝐲𝐨𝐮𝐫 𝐮𝐬𝐞𝐫𝐬 You should feel like an expert on your users, and have some unique/contrarian insights about them. You’ve hopefully identified a “hair-on-fire” pain they have, and you’re building a product to fix their problem. It’s hard to spend too much time talking to your users. Interviewing 50-100 people seems like a good place to start. If you’ve talked to less than, say, 10-20, it might be hard to know you’re solving a real problem. 𝐊𝐧𝐨𝐰 𝐡𝐨𝐰 𝐲𝐨𝐮’𝐫𝐞 𝐠𝐨𝐢𝐧𝐠 𝐭𝐨 𝐦𝐚𝐤𝐞 𝐦𝐨𝐧𝐞𝐲 What’s your path to $1B in revenue? Keep it simple. You don’t need a 5-tier pricing strategy and a multi-year product roadmap, you need to know roughly how many potential customers you have and how much they’re likely to pay (bottom-up analysis), and how big the overall market is (top-down). Reasonable estimates are OK. Remember, the VC model is based on a power law of startup returns, where one huge exit makes up for all other investments going to zero. This makes a tiny chance at becoming a $10B company quite valuable, and a high chance at becoming a $50M company surprisingly useless. 𝐒𝐡𝐨𝐮𝐥𝐝 𝐲𝐨𝐮 𝐚𝐩𝐩𝐥𝐲? Emphatic yes! YC was an awesome experience for us and every YC alumnus I’ve asked so far - it changed the trajectory of our startup. Even if you don’t get in, we found the application process itself was valuable and forced us to think deeply. If you’d like some eyes on your application, send me a message - I’m happy to help!

78

14 Comments

Vishal Sachdev highlights the strategic integration of open source and proprietary tech in architecting tech stacks, developer ecosystems and resulting business models. The world class companies effectively balance value commoditization in open source and value capture in proprietary tech. This is the strategic challenge for companies operating in #blockchain and #AI. I believe those who operate at the intersection of blockchain *and* AI stand to win this strategic battle...

3

1 Comment

NVIDIA's Jensen said "ChatGPT democratized computing, Llama2 democratized generative AI" ...and now Llama3 is the next level 🔥 It's fascinating to watch Meta's strategic moves. With Llama3, they prioritized training on a massive 15T token dataset to pack all into a lean 70B param model, instead of building a massive model. This allows Llama 3 to match trillion+ parameter models like GPT-4, but at 1/10th the compute, storage and inference costs! 💰 This technique was published by Google DeepMind a few years ago https://lnkd.in/gx7VU8aA Meta is not a dark horse in AI anymore. They might be the top dog. https://lnkd.in/gRRJYuxt #llama2 #llama3 #llm #chatgpt #gpt4 #ai #ml

16

5 Comments

YC's "request for startups" - Winter 2025 - Government + Public Safety 1. Build LLMs to automate specific government tasks like form filling, application reviews, and document summaries 2. Create computer vision systems for license plate detection and crime prevention 3. Build software that reduces police paperwork time from hours to minutes 4. Create tools for emergency response coordination and dispatch 5. Build platforms that help communities and law enforcement communicate effectively - Manufacturing 1. Build ML-powered robotics systems that reduce labor costs in US manufacturing 2. Create automation tools specifically for American factories competing with overseas production 3. Build specialized industrial robots for factory inspection and maintenance 4. Develop systems that help manufacturers operate efficiently in US industrial hubs - Chips + Engineering 1. Build LLM tools specifically for FPGA design and optimization 2. Create Al systems for ASIC design that reduce development costs 3. Build tools that optimize specialized computation like crypto mining or data compression 4. Develop Al-powered CAD/CAM software that makes engineering tools more accessible - Stablecoins 1. Build platforms for businesses to hold and manage stablecoins 2. Create tools that make it easy for developers to integrate stablecoin payments 3. Build systems for banks to issue their own stablecoins 4. Develop infrastructure for cross-border stablecoin payments and remittances - New Jobs 1. Create tools that help people run local service businesses 2. Build platforms enabling people to earn income providing services online 3. Create systems that help service providers work more efficiently with Al assistance 4. Develop tools that help small businesses operate at the same level as large corporations #YC

4

🧑‍🍳 Work with Cathy Guo (and me!) on Iterative We're hiring a Senior Associate to help evaluate the startups that apply to Iterative. We receive 1K+ applications roughly every 3 months and invest in roughly 1.8% of those applications. Maybe somewhat counter-intuitively, we want to increase that % and take more risk. The trick is figuring out what areas (founder, market, country, etc.) that we should be taking more risk in. You'll be an active participant in how we evaluate companies. See the job description for the typical stuff but here's 2 things we DON'T do as part of our investment process. 🤔 Invest Based on Who Else Is Investing We don't care who else is in the round. Most of the time, we make our decision before everyone else does. That means you'll have to do the work of building your own conviction. 🤓 Invest on School, Previous Work Experience, etc. We don't have any filters around what school a founder went to, what company they used to work at, etc. We evaluate all founders based on our interactions with them. Why settle for a proxy of whether someone is good (that's what all of these are) when you can talk to them directly? If you can't tell whether a founder is good by talking to them.. well.. Instead.. you'll have to think for yourself and argue from first principles. Mostly with me :). If that's how you want to invest or want to learn how we do that, you should apply. For more details about the role, look for the 📌 in the comments.

172

8 Comments

PMF Daily Dose # 12: An iteration to Y-Combinator's Essential Startup Advice👩‍🔬🔄🚀 I’ve edited and re-framed it specifically for B2B software startups w/ some revenue, some customers, and some PMF. I hope the bucketing and edits are helpful, but my one big push is this: The advice from Y Combinator is focused on startups going from 0-1 and my push is that these same practices are absolutely necessary to go from 1-10, it’s just that the context and way in which you execute them changes as your company grows. Don’t lose your startup DNA while going from 1-10 while honoring your current stage as a larger, slower moving, and more complicated organization. I’ve created 3 buckets: 1/ Everything is Handmade the First Time 2/ Simplify 3/ Be Empathetic to Yourself and Build Meaningful Relationships I’ve made edits/additions to many of the YC bullets based on my experience building startups over the last 18 years. ⚡️1️⃣ Everything is Handmade the First Time⚡️ ➡️ Do things that don't scale. When you find value scale it while remaining small and nimble to build the next wave of features in a fast, iterative, and handmade way until you validate your decision to build/scale or trash. Dual-track agile is a necessary tool for startups going from 1-10. You must balance discovery and delivery to win. ➡️ Find the 90 / 10 solution. ⚡️2️⃣ Simplify⚡️ ➡️ Find 10 then 50 then 100 customers who love your product. Stages matter. Don't get too far ahead of yourself. ➡️ Startups can only solve one problem well at any given time. ➡️ Don’t scale until you have built something people want. And don’t be afraid to gracefully exit team members that no longer fit within the constantly evolving strategy of your org. Be kind to them. Give them notice and severance. But it is a disservice to everyone to keep them if it’s not a fit. ➡️ Only enter into long negotiated deals with big customers if you can engage them for their time -> reputation -> money along the course of the negotiation ie get them on your product and giving you feedback while you negotiate. ➡️ Avoid big company corporate development queries - they will only waste time, unless you can engage them for their time -> reputation -> money along the way. See above. ➡️ Ignore your competitors. ⚡️3️⃣ Be Empathetic to Yourself and Build Meaningful Relationships⚡️ ➡️ All startups are badly broken at some point ➡️ Be careful but not avoidant with conferences. In B2B SaaS you may need to go to conferences to build relationships with customers. Go with a giving mindset and a relationship mindset. Engagement and sales will come after that. ➡️ Sometimes you need to fire your customers (they might be killing you) ➡️ Get sleep and exercise. Take care of yourself and your family/community. ➡️ Founder relationships matter more than you think. ➡️ Be nice! Or at least don’t be a jerk. Karma is real. Here's the original post from YC. https://lnkd.in/g6ftdjtZ

5

The field of LLMs is getting more and more interesting. So many real world use cases are emerging, making a different shift in digital transformation. Stay tuned for more details. #LLMs #digitaltransformation

9

(position filled!) 🚀 We're hiring: Founding Software Engineer (Employee #1)! If you don't speak English fluently in America, you're twice as likely to experience serious medical errors due to miscommunication. Behind this statistic are millions of real patients—parents, children, and loved ones—who receive worse care simply because of a language barrier. We’re on a mission to change that. We're a stealth, venture-backed AI startup building a multilingual healthcare communication platform to ensure every patient has a voice and every provider has the tools to deliver equitable, high-quality care. If you’re a visionary engineer ready to help lay the foundation of a product that will improve lives and redefine what’s possible, we want to talk to you. As our Founding Engineer, you'll be working alongside a team of A-players that includes: • MD with 150+ publications in AI healthcare research • Ex-Apple engineer from UC Berkeley EECS and 2x founder • Former Chief Medical Officer of the U.S. VA and Department of Defense • YC and PearX serial entrepreneur • Head of Data Science at Providence Health 🚀 The Role: As employee #1, you'll be: • Architecting core product features (backend/infrastructure) • Developing our AI infrastructure and fine-tuning LLMs • Working directly with customers • Shaping our technical foundation from day one 📩 Excited about fixing this $60B healthcare disaster? Apply here: http://bit.ly/3CAeiDs #HealthTech #AI #Startup #SFBayArea #Hiring

75

9 Comments

*Developers*...JACoB ("Just Another Coding Bot") is officially open-source and ready for you! Kevin Leneway and Chris Pirich have been incubating this project at PSL for months and we're super excited to formally launch and see what the world thinks. JACoB is a coding agent that automates repetitive programming tasks to boost developer productivity. JACoB is especially good at handling tasks like turning designs to code via Figma integration, refining and improving GitHub issues, and conducting code reviews. JACoB stands apart from other solutions by offering the freedom of an open source tool and the quality of closed platforms like GitHub Copilot. (Check out JACoB's Elo score relative to other tools in the Product Hunt link below!) Please give it a try, share it with friends, and vote or drop us a comment in Product Hunt! https://lnkd.in/gXAEewh3 Pioneer Square Labs

41

1 Comment

LLMs are terrible at basic math. Why? Here's why: Unlike people, LLMs don’t conceptualize mathematics. They simply recognize patterns in the training data (text) they’ve learned from. When asked what 2 + 2 is, an LLM recalls instances of seeing ‘2’, ‘+’, ‘2’ and so on, allowing it to approximate solutions but not match the precision of a calculator. That’s how we end up with LLMs asserting that 2 + 2 = 5. But for analytical teams, like finance, that makes LLMs really hard to adopt in day-to-day workflows. The productivity gains of having AI do calculations are negated by needing to fact-check and audit every calculation and number. LLMs are also not deterministic, so no amount of prompting is going to guarantee you’ll get the correct answer every time. And that’s just for basic math – the outputs I’ve seen from LLMs on more complex calculations and analyses can get very wonky. Add thousands of rows of data and you’re squarely in hallucination land. So what’s the solution? Calculate things manually? No. The good news is that this is changing quickly and there are options: 1 - AI is getting better at math. Researchers continue to train LLMs with more specialized datasets that include a much wider array of mathematical problems and solutions. The newest models, like OpenAI’s latest family of models, are incorporating things like symbolic reasoning and “chain of thought” methods to break complex problems into smaller, logical steps the LLM can more easily process. Results have been really promising and we’ve seen each new iteration of models continue to beat previous benchmarks. 2 - Leverage AI agents For intensive workflows, and to really make sure your outputs are correct, working with companies that are building high-performing AI agents is likely your best bet. Agentic workflows allow LLMs to be used for what they are great at (inferring user intent, making decisions, etc.) while outsourcing the things they struggle with (complex calculations) to other parts of a broader system. Text-to-SQL, “tool-use”, few-shot prompting, and other strategies can be combined with real-time data integrations to dramatically improve performance on analytical operations. At Concourse we’ve spent the better half of 2024 getting our AI Agents to be really good at math, and I’m so excited to continue growing our capabilities to tackle even more complex mathematical problems for our customers in 2025. Reach out if you’re interested in learning more about how we’ve built Concourse and the impact we’re already delivering for finance teams.

42

5 Comments

Today we’re releasing Chai-1r, an updated version of Chai-1 with full support for restraints. Prompting Chai-1r with prior structural knowledge can boost accuracy by >2x. https://lnkd.in/euP7frVQ Custom restraints can refine predictions with knowledge obtained from scientific literature or lab assays. For example, a “contact” restraint specifies that two tokens are within a certain distance. A “pocket restraint” indicates a token should be in contact with another chain. Chai-1r excels in antibody-antigen interaction prediction with mean DockQ % of - Chai-1r (+4 epitope residues): 43.7 - Chai-1: 35.6 - Boltz (an AlphaFold3 reproduction from MIT & Genesis): 26.4 - AlphaFold2: 20.6 Restraints are available immediately both in our public codebase (https://lnkd.in/e3SGx_dj) and on our web server (https://lnkd.in/e9YfR6GV).

578

13 Comments

Big news from StackAI! We closed a $3M Seed Round led by Google's AI fund Gradient Ventures. We now power the GenAI agenda of public companies and fast-growing startups, have over 40k registered users, and support more than 200 organizations. LLMs are the steam engines of our century. But it is only when we connect this engine to a set of wheels and a chassis that we can build truly powerful machines that take us beyond where we could go before. This is not "the bicycle for the mind," as Steve Jobs put it for the Macintosh; in comparison, this feels more like "the car for the mind"... Thanks to Stack AI, you don't need to reinvent the wheel or direct your best engineers to infuse AI throughout your organization. Our platform allows you to deploy custom AI assistants and AI-powered workflows with the click of a button. Both technical and non-technical users can leverage our end-to-end drag-and-drop visual interface and APIs. The next step in our journey is to accelerate the adoption of Artificial General Intelligence and become the Middle-Layer of AI. We are hiring the best engineers, makers, marketers, and operators to bring GenAI's power to as many people in the workforce as possible; reach out if you'd like to be part of this once-in-a-lifetime opportunity! A heartfelt thank you to all our existing investors and lead investors, Y Combinator, Gradient Ventures, Soma Capital, Epakon Capital, Beat Ventures; your support has been nothing short of exceptional. Thank you as well to our mentors Nicolas Dessaigne, Tom Blomfield, Richard Aberman, Michael Seibel, Garry Tan, Collin Wallace, Ashutosh Joshi, Colin Rogister, Dave Anderson, and Stephen Balaban. You can find our official announcement and blog post in the first comment.

1,072

208 Comments

The newest issue of The Pragmatic Engineer found a surprise uptick in eng hiring in the typically slow summer season - I'm curious to see how the macro environment impacts this uptick as we head into the fall, but I remain hopeful. I asked Gemini to give me the tldr of the article to share with my team (sorry, Gergely Orosz) and found it missed one of the most interesting pieces of the article. Gemini's highlights - Unexpected surge: Tech recruitment has seen an unusual spike in June and July, contrary to typical trends. Increased demand: Startups, especially AI-focused ones, are driving the hiring uptick. Global phenomenon: This trend isn't just limited to the US; Europe is also experiencing increased recruitment activity. Reasons for hiring: Companies are backfilling positions, investing in growth due to improved financials, and responding to increased VC funding. Challenges remain: While the job market is improving, finding qualified candidates and offering competitive compensation remains difficult for many companies. What I found most interesting - The nature of the backfills: the article finds that a lot of new headcount are actually backfills of folks that were "over-slashed" during rounds of layoffs. Not entirely surprising with the "trim the fat" and "rebuild lean" mentality, but interesting nonetheless. Either way, the uptick is encouraging (and don't always rely on Gemini's tldrs). #hiring #startups #VCbacked #softwareengineering

11

1 Comment

Old school and probably not popular but I still believe products should be customer-satisfaction focused/led rather than technology-led. Use cases for AI should be byproducts of the product improvement process primarily. The thing though is that founders need to fundraise so they're likely to sprinkle some AI on their decks to even look serious. Not sure that's how the most enduring companies have been built. Of course, if sources of capital become AI-biased, founders have little choice. Happy to keep an open mind and dive in when a use-case that can 10x customer experience for Good Tenants customers and users shows up.

5

Insights from a deep dive into the world of Fine-Tuning LLMs! 🚀 Yesterday I had the incredible opportunity to host The GenAI Collective X OctoAI's Builders Roundtable on Fine-Tuning Large Language Models (LLMs). Here are some valuable insights we gained across three topics we discussed: 1️⃣ Fine-Tuning Foundations: Fine-tuning is particularly useful for specialized tasks such as legal or medical jargon processing, tone customization, and generating specific formats. It allows models to achieve high performance on niche datasets not covered by pre-trained models and at a much lower cost. 2️⃣ The Process of Starting with Fine-Tunes: Getting started with fine-tuning involves selecting the right data, ensuring high-quality, task-specific inputs, and carefully monitoring the model's performance. Evaluation is key to ensure your fine-tuned model meets your specific requirements. 3️⃣ Key Considerations: Effective fine-tuning requires balancing learning new information with retaining pre-trained knowledge. Watch out for overfitting and catastrophic forgetting, and always define clear evaluation criteria to measure success accurately. A huge thank you to our brilliant panelists: Aditya Jain, Benjamin Hamm, Kyle Corbitt, Sophia Yang, Ph.D., and Wing Lian. Your expertise and passion made this discussion truly enlightening! If you're interested in learning more about fine-tuning LLMs, make sure to watch the recording: https://lnkd.in/gJtSCiJu Excited for more engaging AI conversations like this in the future. 🌟 #LLM #AI #finetuning

29

3 Comments