Navigators

Jay Hack co-founded CodeGen, one of the first companies building autonomous coding agents, before ClickUp acquired it and made him head of AI. In this episode of Navigators, Paul sits down with him to ask why, with models this capable, so few problems have actually been solved.Jay's answer centers on one word: siloing. He argues the biggest thing preventing companies from operationalizing agents isn't model quality, it's that sales data lives in Salesforce, chats live in Slack, and docs live somewhere else, each platform incentivized to keep your data locked in. ClickUp's bet is the opposite: put tasks, docs, and chat in one place, build a dense context graph on top of it, and let subagents pull from it cheaply and fast. He connects that same philosophy to why he's bullish on skills as memory, why curing cancer hits the same closed-loop problem as shipping software, and why an agent recently ordered him a pizza using nothing but a browser.Topics covered:07:45 What refactors taught CodeGen about narrow use cases10:39 The biggest problem stopping companies from operationalizing AI11:19 Why a dense context graph beats siloed data12:09 Evals first: building the AI stack as a context engine15:21 Skills as memory: ClickUp's answer to compaction24:13 What's actually stopping AI from curing cancer26:30 Teaching an agent to order a pizza with just a browser

Shensi Ding (Co-Founder & CEO of Merge) started building integrations six years ago, before AI was the market. Now OpenAI, Perplexity, and Factory run on Merge’s connectivity layer. In this episode of Navigators, Paul sits down with her to talk about what it actually takes to connect agents to the systems they need.When Merge benchmarked its tools against public MCP servers, it found a 99.999% success rate compared with roughly 60% for public servers. Shensi says most of those public servers are side projects and marketing that nobody maintains. That is also why AI security hasn’t had its big incident yet: the tools fail before any PII can leave. She explains why that is about to change, why it’s irresponsible to run AI without a DLP layer, and why you can’t vibe code critical infrastructure.Chapters:0:00 Intro1:19 What Merge builds: the connectivity layer for AI4:02 The agentic-tooling gap MCP opened5:14 Why thousands of tools is a security risk6:13 "They only see what's on the menu"6:55 Why it's irresponsible to run AI without a DLP layer7:39 Merge for Workforce: governance for IT12:40 99.999% vs 60%: benchmarking tool reliability12:59 Is MCP the right protocol?18:30 Why the founder sets the AI ceiling21:05 996, barrels and ammo, and hiring for AI37:57 Why humans become the blocker to the singularityBrowserbase gives AI agents a browser to operate in. Navigators is our series of conversations with the people building the agentic future.

Fireflies gives away unlimited meeting recordings and transcriptions. Krish Ramineni (Co-Founder & CEO of Fireflies) says that was the plan from day one: commoditize yourself before someone else does it for you.In this episode of Navigators, Paul sits down with Krish to talk about what happens to software when models drive most of the interaction. Krish thinks most SaaS will be given away for free and that companies will charge for the agentic work instead. If you don't become agentic, you become a data source or a connector inside someone else's model, and all the value goes to them. He explains why you can't vibe code a CRM, why seat-based pricing breaks when companies run with fewer people, and why Salesforce had to decide whether to go headless.They also cover how Fireflies hit unicorn status on $19M raised, how open weight models made the free tier possible, and why Fireflies routes across seven to 10 model providers.00:00 Intro 01:48 Years of failed products before the tech caught up 04:25 Why meetings were the data nobody could capture 07:21 Running Fireflies on Fireflies with 130 people 08:41 The support agent that handles more tickets than any human 10:37 Why Fireflies won't train on customer data 11:55 Open weight models and an unlimited free tier 13:19 Why app companies need a model router 16:05 You can't vibe code a CRM 17:32 SaaS goes free, agentic work gets paid 19:47 Become infrastructure or become a connector 23:20 A unicorn on $19M raised 25:17 The second act: email, Slack, and a context graphBrowserbase gives AI agents a browser to operate in. Navigators is our series of conversations with the people building the agentic future.

Harrison Chase (Co-Founder & CEO of LangChain) has rebuilt his product stack four times since 2022, chasing models that kept changing underneath him. In this episode of Navigators, Paul sits down with him to talk about what LangChain got right, and the one piece of infrastructure it built years too late.Harrison admits the team underestimated how massive the demand for raw inference would become, and how big general-purpose coding agents would get. He walks through the shift from LangChain to LangGraph to DeepAgents, why sub-agents rarely work the way people expect, and why LangChain now runs its own hiring, content, and coding work through internal agents exposed in Slack.Topics covered:The gateway LangChain should have built a year earlierWhy agent capability converging on human capability changes the tools debateCode mode: why agents writing code beats agents using human-style toolsRunning GTM, content, and coding agents entirely inside SlackWhy sub-agents need to own whole tasks, not small piecesThe agent identity problem: delegated access vs. service accountsWhy LangChain's own AI slop problem got bad enough to need a Slack emojiNavigators is a series by Browserbase where we sit down with AI leaders navigating the future of what's happening in AI today.Host: Paul Klein (@pk_iv), founder, BrowserbaseGuest: Harrison Chase (@hwchase17), co-founder and CEO, LangChainBrowserbase: https://browserbase.com LangChain: https://langchain.com

Immad Akhund (Co-Founder & CEO of Mercury) uses what he calls the cosmic cube to decide what Mercury builds next: one axis for customer type, one for product, one for compliance and infrastructure. In this episode of Navigators, Paul sits down with him to talk about why that space is functionally infinite, and why trying to expand into all of it at once is how companies lose.Immad also explains the two-year mistake of waiting to launch a corporate credit card, why he treats every new product like a fresh seed-stage startup inside Mercury, and why talent, not capital, is the real limit on how many bets the company can run at once.CHAPTERS:00:00 Why 1-on-1s stopped working once he was managing execs07:34 Agentic cards: letting AI spend on your behalf11:09 Why fintech is three times harder than software13:50 Chewing glass: getting bank-charter ready16:08 Reaching the edge of knowledge in six months28:26 Killing 1-on-1s and what Jensen Huang has to do with it36:39 The "cosmic cube": how Mercury picks its next bet---Navigators is a series by Browserbase where we sit down with AI leaders navigating the future of what's happening in AI today.Host: Paul Klein (@pk_iv), founder, BrowserbaseGuest: Immad Akhund (@immad), co-founder and CEO, MercuryBrowserbase: https://browserbase.com Mercury: https://mercury.com