Ben Mathews

The food supply chain remains one of the largest, most complex, and least digitized systems in the global economy. Companies like anchr are embedding AI directly into distributor workflows unlocking massive efficiency gains across an industry that quite literally feeds the world. Excited to see Smayan Mehra, Tzar Taraporvala, and the team hiring their next founding teammates. If you want to help rebuild the backbone of food distribution, this is a rare opportunity.

12

2 Comments

Don't Build a Better Wheat Farm" - Why Defensibility Stakes Are Higher in Deep Tech Just published a new piece on my "Ignore the Confusion" blog, building on thoughtful insights from Eric Ver Ploeg at Tunitas Ventures about startup defensibility. Eric's core thesis: Too many startups pitch like wheat farmers - "huge TAM, slow incumbents, growing market, domain expertise" - but fail to think through long-term defensibility until it's too late. From a deep tech perspective, the stakes are even higher: ** Unlike software, deep tech founders must commit to defensibility strategies from day one - their funding depends on it ** Patent vs. trade secret decisions are often difficult to reverse and shape your entire competitive strategy ** Even "picks and shovels" providers (the tools that make industries more efficient) become commodities without proper moats The key insight that resonates: Defensibility can't be retrofitted. Whether you're building software or deep tech, your moat must be architected into the business model from the start. Thanks to Eric Ver Ploeg for sharing these insights on startup strategy and letting me build on his framework from a deep tech lens. Read the full post: https://lnkd.in/dEj_iF-Q #DeepTech #StartupStrategy #Defensibility #VentureCapital #Innovation

21

2 Comments

Today Isembard announced a $50m Series A fundraise led by Union Square Ventures, less than a year after Notion Capital led their Seed round. 💥This is a company right at the intersection of AI & supply chain sovereignty & they’re exceeding even our most optimistic expectations. By the end of the year they will have 25 factories across the UK, US, Germany, France & Ukraine. 💥Isembard is rapidly scaling its network of AI-first factories for precision manufacturing at the same time as market demand is accelerating. Macro forces (reshoring, rising defence spend, concentration risk in Asia, and the growth of neo-primes) are structurally increasing demand for fast, local, flexible manufacturing that the current industry is unable to deliver on. 💥The innovation is in the way the company builds and franchises factories as a product with tightly integrated units of machines, software, robotics, & process intelligence that turn design files into certified components with speed & reliability. 💥Their AI software platform, MasonOS, connects all sites into a single operating system, replacing fragmented shops with standardised, high-performance industrial nodes & enabling real-time quoting, predictable delivery, & low defect rates. By owning the full production stack, Isembard delivers premium manufacturing performance with the scalability & flexibility of a software platform. 💥We are at the top of the AI hype cycle and there are understandable concerns about the ROI on the vast sums of money being invested into this new super-cycle. But, the ROI for Isembard is both clear and compelling. They will deliver components 10x faster and at 50% of the cost of current suppliers. They will also build a de-centralised, global network of factories to meet the growing demand for national or regional sovereignty. 💥This is a company with a clear vision to disrupt a massive, fragmented $1.8tn component manufacturing market. Today marks another huge stride towards that vision. And they are only just getting started. We’re thrilled to be on this journey with Alexander Fitzgerald & the team & we’re also excited to welcome Rebecca Kaden & USV into the investor base. Notion Capital Union Square Ventures Alexander Fitzgerald Rebecca Kaden Maximilian Eichler Stephen Millard Britt Mulder

56

6 Comments

With RL techniques like GRPO, it’s become easy to fine-tune small models to outperform foundation models on vertical tasks. We’re open sourcing Osmosis-Apply-1.7B: a small model that merges code (similar to Cursor’s instant apply) better than foundation models. Links in comments! Using foundation models for high specificity, low complexity tasks like code merging can be overkill - instead, it’s better to use a specialized model instead for higher performance and latency. We fine-tuned the model on 100K real-world code commits where the model needs to merge a code edit snippet into an original code snippet. On the test dataset (subset of CommitPackFT), foundation model performance ranged between 0.77-0.93 reward score. In comparison, Osmosis-Apply-1.7B achieved a 0.98 reward score - while also being 3X-5X more affordable than the next cheaper model AND ~10X faster! Why this matters: the best AI agents use multiple models to ensure task specialization and reliability at scale - on top of cost and latency benefits as well. Links to the model and full write-up (with reward function, hyperparameters, etc.) are in the comments - reach out if you’re interested in using reinforcement learning!

240

18 Comments

Thrilled to announce that Together Fund is investing in Sentra, alongside a16z speedrun! You track results in Jira. Decisions in Notion. Conversations in Slack. But the reasoning, the debates, trade-offs, and context behind why you chose A over B, disappears into what we call "Dark Matter." A decision made in March looks insane by July because no one remembers the constraints that made it smart. I lived this firsthand at Twitter scaling from 800 to 8,000 employees, and at Google while launching AI Overviews to billions at planet scale. The problem isn't process. Process is compensation for something deeper: organizational amnesia. An organization’s "Systems of Record" doesn’t solve this, they encode it. They store what happened, never why. That's why we are investing in Sentra. Sentra is the always-on collective memory that eliminates organizational amnesia by maintaining accurate context for all members and agents, functioning as an operational nervous system. It connects to every channel where work happens, meetings, Slack, email, code commits, docs, calendars, and treats them not as artifacts to search, but as living signals to synthesize. The fleeting and the permanent, unified into a memory that understands. The founding team is built for this: - Jae Gwan Park (CEO): Product-first founder, memory systems research at UofT and MIT - Ashwin Gopinath (CSO): Former MIT professor, created "Reflexion" (NeurIPS 2023), agents that learn from mistakes, 2x founder - Andrey Starenky (CTO): Early Vapi engineer, ex-IBM, built to process enterprise-scale data firehose Together is an operator-led fund. We invest in problems we've lived. This is one of them. Many congrats Jae, Ashwin and Andrey, we are so excited to partner with you! Read the full thesis: https://lnkd.in/gixj9cE4 Book a demo: https://www.sentra.app/ #OrganizationalMemory #AI #Sentra #TogetherFund #a16z #ContextGraphs

71

3 Comments

What do you do with a fresh $100m and dozens of the largest companies in the world as customers (The Coca-Cola Company, Novartis, Pfizer and all those other fancy logos, you know who you are!)? I'll tell you what ORO Labs does -- build AI agents on supply chain data ontologies which find problems fast, report real time spend, save literal 100s of millions of dollars, and super power procurement. Therefore your marketing leader in Brazil partners faster, your supply chain of chemicals components is more agile, your bankers buy the tech same day. The world moves faster. Simply said, ORO Labs is a marque examples of a system of record which is AI powering the largest companies in the world. The business is also run by the nicest and best of people. I'm so excited for Sudhir Bhojwani Lalitha Rajagopalan Yuan Tung and all of the team. So excited for this milestone and to be moving at speed.

30

1 Comment