The Tech Trek

Sandesh Patnam, Managing Partner at Premji Invest, breaks down how long duration capital changes the way you evaluate companies, founders, and moats. We talk about what most growth investors miss, why product strength still matters, and how to separate real AI businesses from thin wrappers in a noisy market.Premji Invest is a captive, evergreen fund built to grow an endowment that supports major education work, which gives the team flexibility on time horizon and partnership style. Sandesh shares how that shows up in diligence, how they think about backing contrarian founders, and why the best companies in this AI era may still be ahead of us.Key TakeawaysFocus on the long arc, not quarter by quarter optics, founders make better decisions when they are not trapped in short term metricsIn growth investing, TAM models and KPI spreadsheets can distract from the core question, does the product have real strength and an expanding roadmapEnduring outcomes often come from backing a contrarian view early, then helping it move from contrarian to consensus over timeEvergreen capital changes behavior, you can slow down, build relationships, and partner across private and public markets instead of treating IPO as the finish lineIn AI, separate the stack into data center, foundation models, and applications, then look for defensibility like vertical depth, data moats, and compounding usage valueTimestamped highlights00:38 Premji Invest explained, evergreen structure, one LP, and why public markets can be part of the journey, not the exit04:47 Two common growth investor lenses and what gets missed when product and roadmap do not lead the thesis08:48 Partnership mindset, building trust, and being the first call when things get hard12:48 The contrarian to consensus path, what creates alpha, and how to support founders through the lonely middle19:54 Why rushing decisions is a trap, and how flexibility changes when and how you can partner with a company20:55 AI investing framework, three layers, what looks frothy, what can endure, and where moats still exist26:48 The cost of intelligence is collapsing, why this may still be the early internet moment, and what that implies for the next waveA line that stuck with me“We want to be the first port of call when the seas are turbulent.”Practical moves you can stealPressure test the roadmap, ask when product two ships, what adjacency comes next, and what tradeoffs change at scaleWhen evaluating AI apps, demand a defensibility story beyond the model, look for proprietary data, vertical workflow depth, and value that improves with usageTreat speed as a risk factor, if you cannot complete your churn cycle of doubt and validation, step back rather than force certaintyCall to ActionIf you liked this one, follow the show and share it with a founder, operator, or investor who is building in AI right now. For more conversations at the intersection of tech, business, and execution, subscribe and connect with me on LinkedIn.

Software engineering is changing fast, but not in the way most hot takes claim. Robert Brennan, Co founder and CEO at OpenHands, breaks down what happens when you outsource the typing to the LLM and let software agents handle the repetitive grind, without giving up the judgment that keeps a codebase healthy. This is a practical conversation about agentic development, the real productivity gains teams are seeing, and which skills will matter most as the SDLC keeps evolving. Key TakeawaysAI in the IDE is now table stakes for most engineers, the bigger jump is learning when to delegate work to an agentThe best early wins are the unglamorous tasks, fixing tests, resolving merge conflicts, dependency updates, and other maintenance work that burns time and attentionBigger output creates new bottlenecks, QA and code review can become the limiting factor if your workflow does not adaptSenior engineering judgment becomes more valuable, good architecture and clean abstractions make it easier to delegate safely and avoid turning the codebase into a messThe most durable human edge is empathy, for users, for teammates, and for your future self maintaining the systemTimestamped Highlights00:40 What OpenHands actually is, a development agent that writes code, runs it, debugs, and iterates toward completion02:38 The adoption curve, why most teams start with IDE help, and what “agent engineers” do differently to get outsized gains06:00 If an engineer becomes 10x faster, where does the time go, more creative problem solving, less toil15:01 A real example of the SDLC shifting, a designer shipping working prototypes and even small UI changes directly16:51 The messy middle, why many teams see only moderate gains until they redraw the lines between signal and noise20:42 Skills that last, empathy, critical thinking, and designing systems other people can understand22:35 Why this is still early, even if models stopped improving today, most orgs have not learned how to use them well yetA line worth sharing“The durable competitive advantage that humans have over AI is empathy.”Pro Tips for Tech TeamsStart by delegating low creativity tasks, CI failures, dependency bumps, and coverage improvements are great training wheelsDefine “safe zones” for non engineers contributing, like UI tweaks, while keeping application logic behind clearer guardrailsInvest in abstractions and conventions, you want a codebase an agent can work with, and a human can trustTrack where throughput stalls, if PR review and QA are the bottleneck, productivity gains will not show up where you expectCall to ActionIf you got value from this one, follow the show and share it with an engineer or product leader who is sorting out what “agentic development” actually means in practice.

Deborah Hanus, Co-founder and CEO at Sparrow, joins Amir to unpack the founder journey from academia to building a scaled company. They dig into why leave management is still a messy, high stakes problem, and how Sparrow is turning it into a clean, guided experience for both HR and employees.Sparrow helps companies provide employee leave across the United States and Canada, and Deborah shares what it really takes to scale a compliance driven business without slowing down. From founder resilience and early stage emotional swings to hiring, onboarding, and culture design, this one is packed with lessons for operators and builders.Key takeaways• Academia can be real founder training, especially for building resilience and hearing “no” without losing your edge• Early stage startups feel brutal because you have too few data points, it is easy to overreact to every win or setback• Compliance and leave are fundamentally data problems, the right info to the right person at the right time changes everything• Scaling leadership is mostly communication and alignment, five people and 250 people require totally different systems• Culture does not stay stable by accident, values must drive hiring, training, rewards, and performance managementTimestamped highlights00:37 What Sparrow does, and the 300 million dollars in payroll cost savings milestone01:37 Why academia can prepare you for founding, and how customer pain beats outside skepticism03:40 The leave compliance mess, and why state by state rules made the problem explode08:25 The two real ways startups die, and why morale matters as much as cash12:55 Leading at scale, onboarding, clarity, and the feedback questions that keep teams aligned19:54 “Scale intentionally” as a culture principle for a company that cannot afford to break things25:48 Keeping values stable while everything else evolves as the team growsA line worth sharing“Companies end when you run out of cash or you run out of morale.”Pro tips you can steal• Treat the employee journey like a product journey, from recruiting through promotions and hard moments• Before a big change, collect questions early so the message lands where people actually are• After a meeting, ask “What were the main points?” to see what people heard, then tighten your messaging• Invest in onboarding and goal clarity to prevent teams from drifting into competing prioritiesCall to actionIf you enjoyed this conversation, follow and subscribe so you do not miss what is next.

Max Bruner, Founder and CEO of Anzen, joins Amir Bormand to break down why insurance is quietly one of the biggest data and workflow opportunities in tech right now. They dig into Max’s unconventional path from foreign policy to building an executive liability marketplace, and what it really takes to modernize a slow moving industry with AI.If you care about building in real world markets, scaling with discipline, and using AI for more than content, this one will sharpen your thinking fast. Key Takeaways• Insurance is not flashy, but it is foundational, massive, profitable, and packed with repeatable workflows that software can improve• The best tech opportunities are often in slow moving industries with lots of data and outdated systems• Better decision making comes from predicting outcome impact and pressure testing your thinking with a strong community around you• AI value is clearest when it drives real operations, faster transactions, lower costs, and better service• Fundraising is a pipeline game now, treat it like sales, build the plan, hit the numbers, run a tight processTimestamped Highlights00:42 What Anzen actually does, a one stop marketplace for executive liability quotes across the US02:29 From Arabic studies and foreign policy to discovering insurance through political risk08:12 The curiosity engine, how deep research habits shaped his ability to build in new domains11:23 Decision guardrails, learning from outcomes and using trusted people to keep you efficient13:12 Why choose insurance, building in industries that make the world work, plus the profit reality17:29 The startup advantage, modern infrastructure vs incumbent legacy systems, and why catching up takes time20:36 Raising in today’s market, what changed, what worked, and why the pitch volume mattersA line worth stealing“Sometimes in tech we miss the application, there are massive industries to go change if we apply technology in the right way.” Max BrunerPro Tips for builders• Pick markets with repeatable workflows, you can ship measurable value faster• Spend your time where the outcome impact is high, skip low ROI rabbit holes• Build a real financial plan before fundraising, then operate close to it• Run fundraising like a sales process, pipeline, volume, and discipline winCall to ActionIf you enjoyed this conversation, follow the show and leave a quick review, it helps more builders find it.

Stu Solomon, CEO of HUMAN, joins Amir to unpack a blind spot most teams underestimate: a huge share of online activity is not people at all, it is automated traffic. They break down how verification really works at internet scale, why agentic workflows change the rules, and what it will take to build trust when bots transact with bots.If you have ever wondered how fraud, fake clicks, account abuse, and synthetic behavior get caught in real time, this episode is a clear, practical look behind the curtain.Key takeaways• Most of the internet is machine traffic now, the goal is no longer spotting bots, it is separating good machines from bad ones• Trust is built by combining behavior, infrastructure signals, and identity or credential history into fast decisions at scale• Agentic systems lower the barrier to entry for attackers, less skilled actors can now create outsized impact• The hard part is accountability, when a machine acts with your authority, who owns the outcome• Adoption follows convenience, but visibility matters, if it feels like a black box, people will not trust itTimestamped highlights00:33 HUMAN in plain English, making split second decisions about who is human, and whether they are safe03:59 The trust stack, behavior signals, infrastructure clues, and identity or credential history10:19 The real shift with AI, lower barriers for attackers, plus the rise of agentic autonomy14:37 The cake story, an agent completes the task, then surprises you with a 750 dollar bill17:22 Bots talking to bots, where accountability and liability get messy fast24:18 Security builds trust, trust unlocks adoption, and society is already closer than it thinksA line you will remember“We have always operated on the notion that if you are human, you are good, and if you are a machine, you are bad. That is simply not the case anymore.”Practical ideas you can use• Add guardrails when you delegate to tools, especially budgets, limits, and approval steps• Watch for trust signals, not just identity checks, behavior plus infrastructure plus history beats any single data point• Design for visibility, show users what the system did and why, so trust can compound over timeFollow:If this episode helped you think more clearly about trust, fraud, and agentic systems, follow the show, subscribe for more conversations like this, and share it with a teammate who is building in ads, ecommerce, identity, security, or AI.

Mek Stittri, CTO at Stuut who builds agented AI for finance ops. He breaks down trust-with-verification in technical leadership. Short tactics on staying three-levels-deep, using weekly signals, and when to step in. A look at how AI shifts engineers toward system design and managing agents, not just code.

Michael Topol, Co-founder and Co-CEO at MGT Insurance, explains why insurance is quietly becoming one of the most interesting data and AI problems in tech.We get practical about turning messy legacy data into usable signals, how agentic tools change decision making, and why culture and team design matter as much as the models.MGT Insurance is building a fully verticalized AI and agentic native insurance company for small businesses, pairing experienced insurance operators with top tier technologists. Michael breaks down what changed in the last few years that makes real disruption possible now, and what modern product delivery looks like when prototyping is cheap and iteration is fast.Key takeaways• Insurance is a data business at its core, but most incumbents cannot use their data fast enough because it lives across silos, mainframes, and old systems.• Modern AI lets teams combine internal data with public signals to speed up underwriting and improve consistency, without losing human judgement.• Vibe coding and rapid prototyping collapse the gap between idea and implementation, bringing product, engineering, and the business closer together.• Senior talent gets more leverage in an AI driven workflow, and small teams can ship faster by focusing on problem solving, not just building.• Pod based teams, fixed outcome planning, and strong culture help regulated companies move quickly while staying inside the rules.Timestamped highlights00:44 What MGT Insurance is, and what “AI and agentic native” means in practice02:09 Why small business insurance matters more than most people realize06:06 The real blocker for incumbents, data exists but it is not usable08:55 Vibe coding in a regulated industry, where it helps first12:54 Requirements are shifting, prototypes bring teams closer to the real problem17:26 The pod structure, plus the Basecamp inspired approach to scoping and shipping20:52 Better, faster, cheaper, why AI finally makes all three possible22:11 Where to connect, and who they are hiringA line you will remember“Insurance is really just a big data problem.”Pro tips you can steal• Build cross functional pods early, include a domain expert, a technical product lead, and a senior engineer from day one.• Scope for outcomes, not perfect specs, then let the team decide the depth as they build.• Use AI to automate collection and synthesis, then keep humans focused on the decisions and trade offs.Call to actionIf you enjoyed this one, follow the show and share it with a builder who is trying to ship faster with a smaller team.

Marco DeMeireles, co-founder and managing partner at ANSA, offers insights into how modern VCs navigate the complexities of investing in open source and mission-critical sectors. He discusses the importance of founder DNA and strategic paywalls in success, along with ANSA's concentrated portfolio approach that allows for rigorous diligence. Marco highlights the significance of providing hands-on support to startups while focusing on compounding growth and efficient business models. He also delves into smart category selection in the AI landscape, targeting niches with high-ROI workflows.

AI is everywhere, but most teams are stuck talking about efficiency and headcount. In this episode, Dave Edelman, executive advisor and best selling author, shares a sharper lens, how to use AI to create real customer value and real growth.We get into the high road vs low road of AI, what personalization should look like now, and why data has to become an enterprise asset, not a bunch of disconnected departmental files.Key Takeaways• Efficiency is table stakes, the real win is using AI to build new experiences that customers actually want• Start with customer friction, find the biggest compromises and frustrations in your category, then design around that• Personalization is no longer limited by content scale in the same way, AI changes the economics of tailoring experiences• You do not always need one giant database, modern tools can pull and connect data across systems in real time• Treat data as an enterprise resource, getting cross functional alignment is often the hardest and most important stepTimestamped Highlights• 00:46 Dave’s origin story, from early loyalty programs to Segment of One marketing• 03:33 The high road and low road of AI, growth experiences vs spam at scale• 06:51 Where to start, map the biggest customer frustrations, then build use cases from there• 16:31 The data myth, why you may not need a single mega database to get value from AI• 21:31 Data as a leadership problem, shifting from functional ownership to enterprise ownership• 25:14 Strategy that actually sticks, balancing bottom up automation with top down customer led directionA line worth stealing“Use those efficiencies to invest in growth.”Pro Tips you can apply this week• List the top five customer frustrations in your category, pick one and design an AI powered fix that removes a compromise• Audit your data reality, identify where the same customer facts live in multiple places, then decide what must be unified first• Run a simple test and learn loop, create multiple variations of one experience, measure what works, and keep iterating• Put strategy on the calendar, make room for a recurring discussion that is not just metrics and cost cuttingCall to ActionIf this episode helped you think differently about AI and growth, follow the show, leave a quick rating, and share it with one operator who is building product, data, or customer experience right now.

Amanda Kahlow, CEO and founder of 1Mind, joins Amir to break down what AI changes in modern sales and go to market, and what it does not. If you lead revenue, product, or growth, this is a practical look at where AI creates leverage today, where humans still matter, and how teams actually adopt it without chaos.Amanda shares how “go to market superhumans” can handle everything from early buyer conversations to demos, sales engineering support, and customer success. They also dig into trust, hallucinations, and why the bar for AI feels higher than the bar for people.Key takeaways• Most buyers want answers early, without the pressure that comes with talking to a salesperson• AI can remove friction by turning static content into a two way conversation that helps buyers move faster• The hardest part of adoption is not capability, it is change management and trust inside the team• Humans still shine in relationship and nuance, but AI can outperform on recall, depth, and real time access to the right info• As AI levels the selling experience, product quality matters more, and the best product has a clearer path to winTimestamped highlights00:31 What 1Mind builds, and what “go to market superhumans” actually do across the full buyer journey02:00 The buyer lens, why early conversations matter, and how AI gives control back to the buyer06:14 Why the SDR experience is frustrating for buyers, and where AI can improve both sides09:42 Change management in the real world, why “everyone build an agent” gets messy fast13:04 Why “swivel chair” AI fails, and what real time help should look like in live conversations15:52 Hallucinations and trust, plus the blunt question every leader should ask about human error22:26 Competitive advantage today, and why adoption eventually pushes markets toward “best product wins”A line worth sharing“Do your humans hallucinate, and how often do they do it?”Pro tips you can use this week• Start with low stakes usage, bring AI into calls quietly, then ask it for a summary and what you missed• Build adoption top down, define what good looks like, otherwise you get a pile of similar agents and no clarity• Focus AI on what it does best first, recall, context, and instant answers, then expand into workflow and process laterCall to actionIf this episode sparked ideas for your sales team or your product led funnel, follow the show so you do not miss the next one. Share it with one revenue leader who is trying to modernize their go to market motion, and connect with Amir on LinkedIn for more clips and operator level takes.