Northlight

@pk_iv curious where this breaks down -- is it the data, the prompts, or the judgment call right before send?

Your agents suck when using the web because 85% of it doesn't have an API. Browserbase gives them everything they need to do work online. Leading AI companies like Ramp, Lovable, and Clay trust us to power agents that do real work on behalf of real people. With a single API key, your agent gets everything it needs to navigate the wild web: browsers, search, fetch, identity, a sandbox runtime, and model gateway. Stop waiting on integrations, build agents that can browse and interact with the web just like humans.

@katiekirsch @a16z @andrewchen @kirbyman01 the person who frames the problem controls the outcome. always has.

Stoked to announce the @a16z Growth Engineer Fellowship! A new 8-week cohort of growth engineers, AI-native GTM leaders, and agent builders shaping the future of company growth. Join us! Apps went live today. Fellowship starts in May 2026. Featuring @andrewchen, @kirbyman01, @luclevesque, @laurynmotamedi, @ramanrmalik, @kylecnorton, @gauravvohra, @lukeharries, @egavrilenko11, and more. Apply here: a16z.fillout.com/t/9RAnDmtvdkus Details: a16z.news/p/introducing-… CC @david__booth, @eriktorenberg, @tdoggyholhol

@AlfieJCarter spent a month pitching mid-market SaaS, nothing moved. switched to 30-person companies where a founder was doing their own outbound -- the list was 10x smaller and 10x better.

Claude can replace your entire GTM toolstack in 30 days. While everyone's obsessed with building custom AI wrappers from scratch, we built something different. Claude as a GTM-specific execution engine. Inside a skills library that reads your brand, voice, and ICP automatically. We're not prompting Claude from scratch every session. We're installing context once and letting every skill pull from it automatically. Here's what it does: → Runs a single /bootstrap command that onboards Claude to your brand, voice, and ICP once so every skill reads that context without re-explaining it → Activates skills automatically when Claude detects a matching task from plain-language descriptions with no manual triggering → Cross-references skills so a cold email task pulls from brand positioning, personal voice, and ICP simultaneously → Runs across Claude Code, Claude Cowork, and any agent on the Agent Skills spec without a custom build Humans still set the strategy, define the ICP, and approve the output. Claude just handles every repeated GTM execution task without starting from a blank prompt. Result: No more re-explaining context every session. Every workflow runs with your GTM knowledge already loaded. I just put together a breakdown covering: - What the plugin is and how the /bootstrap command works - All 4 installation methods - 131 skills across 11 GTM domains including outbound, LinkedIn, content, growth, analytics, SEO, and AI search - How skills cross-reference each other so output reflects your brand not a generic template This is the same skills library we use to run GTM execution at Conigma. No fluff. Just the actual skills and setup. Want the "Claude 131 Skills" playbook? 1. Connect with me 2. Comment "SKILLS" (must be connected for priority access)

You can now tell @NorthlightHQ to export every call recording from Otter, Fathom, or Granola in one prompt. It navigates to each app, finds your calls, and maps them to the right contact in HubSpot, Salesforce, or downloads them as a CSV. We're currently MacOS only, but coming to Windows soon.