On The Frontier

Intercom was shipping and experimenting with AI enabled products since 2017. This episode is for those interested in hearing about the evolution of AI models, the customer results, and how these models can be applied in production.

Fergal Reid is Vice President of AI at Intercom. He started at Intercom nearly 7 years ago to help the company begin delivering products leveraging AI. In this episode, we cover how ML & AI has evolved since 2009 and why the reasoning engines that exist in models today are a substantial improvement. We dive into Intercom’s AI-enabled products starting in 2017 to now Fin (Intercom’s AI chatbot) in 2023. Fergal dives into the technical considerations, how the team delivers usable AI enabled software, dealing with hallucinations and the tangible outcomes that Fin has enabled for Intercom’s customers. We discuss latency tradeoffs, using multiple models, building MVPs, and much more. Teams at any stage of their AI journey will benefit from this episode.

Where to Find Fergal:

* LinkedIn: https://www.linkedin.com/in/fergalreid/

* Twitter: https://twitter.com/fergal_reid

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Newsletter: https://www.shomik.substack.com

* Podcast: Apple Podcasts, Spotify, YouTube.

In this episode, we cover:

(00:49) - Fergal’s Background

(03:23) - Why This Wave of AI is Capturing So Much Attention (pun intended)

(05:51) - Capabilities of Any One Model Across Multiple Domains

(08:04) - Evolution of Intercom’s AI Enabled ChatBots From 2017 On

(15:07) - Managing Latency & Performance Tradeoffs Amongst Models

(17:11) - RAG & Other Considerations Accounted for when Shipping Products

(19:43) - How Much Prompting is Needed for What Use Cases

(20:41) - Routing Context Between Multiple Models in an Application

(23:14) - Updating Context in Vector Embeddings

(24:58) - Best Practices for LLM Data Access

(27:04) - Outcomes from Customers Using Fin

(29:08) - Customer Response to Hallucinations

(32:09) - Pricing Based on Successful Resolution

(35:28) - Advice for Technical Teams Experimenting with LLMs

(38:20) - LLMs as Reasoning Engines vs Fuzzy Databases

How to Subscribe:

Available on Apple Podcasts, Spotify, YouTube.

One of the most exciting private companies today is Databricks, growing over 50% at $1.5B ARR with a recent valuation of $43B announced in September 2023. This is the deep dive you’ve been waiting for to understand the product, growth, and competitive landscape.

Chris Huskey is a Partner at Octahedron Capital, a hedge fund that invests in public and private markets. Databricks is a large position in Octahedron’s portfolio. In this episode, we cover every aspect of Databricks. The origin story, culture & ethos of the company, TAM expansion, platform components, competitive dynamics, financial metrics, and the future ahead for the company.

Where to Find Chris:

* LinkedIn: https://www.linkedin.com/in/chuskey/

* Twitter: https://twitter.com/chrishuskey

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Newsletter: https://www.shomik.substack.com

* Podcast: Apple Podcasts, Spotify, YouTube

In this episode, we cover:

(00:36) - Databricks Founding Story

(03:47) - What is Databricks Product

(07:13) - Financial Metrics

(08:33) - What is Apache Spark

(11:38) - Who are the Personas & Users of the Product

(14:11) - Enterprise Use Cases

(15:48) - The Rise of the Lakehouse Architecture

(17:27) - Breaking Down the Platform’s Components

(20:23) - Phase Shifts in Market Demand for Databricks Products

(22:15) - TAM Analysis

(26:04) - Competitive Landscape

(38:32) - GTM & Pricing Model

(41:48) - LLMs as a Tailwind

(48:49) - Cash Flow Dynamics

(52:20) - Risks & Software Velocity as the Mitigator

(57:24) - LLM Infrastructure

Show Notes:

Databricks Series A Pitch Deck: https://bestpitchdeck.com/databricks-series-a

Databricks Series D Pitch Deck: https://bestpitchdeck.com/databricks-series-d

Octahedron Capital’s: A Few Things We Learned This Quarter (highly recommend subscribing!)

How to Subscribe:

Available on Apple Podcasts, Spotify, or YouTube.

Curious about how to scale Open Source companies. Want to hear anecdotes of existential company killing moments that turned into defining assets? This episode has everything from leadership tactics, hard earned advice, managing in competitive industries, new category creation and much more.

Marten Mickos is CEO of HackerOne and the OG of Open Source. He has 36+ years of enterprise software experience. His resume would take too long to fully cover but he previously was the CEO of MySQL (one of the first open source companies) and CEO of Eucalyptus Software. He’s also served as a board member and advisor to companies like Nokia and Verizon Media. Broadly this episode can be categorized into three buckets: 1) building Open Source software with limited resources and overcoming numerous near death company experiences, 2) creating a new category with bug bounties as a service at HackerOne, and 3) specific startup advice on leadership, culture, and embodying decentralized decision making. There are countless gems in this episode that are too hard to cover in a summary so just listen and take in the anecdotes and lessons that Marten kindly shares in detail.

Where to Find Marten:

* LinkedIn: https://www.linkedin.com/in/martenmickos/

* Twitter: https://twitter.com/martenmickos

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Newsletter: https://www.shomik.substack.com

* Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(01:04) - “Technology was not my dream nor is not a Purpose of its Own but Rather a Tool for Human Beings” - Marten’s Early Career

(04:27) - Joining the MySQL Team

(06:53) - Turning an Existential Moment with MySQL Getting Sued in 2001 at Seed Stage Into an Asset and Marketing Engine for Customers

(10:34) - Open Source Business Model at MySQL Before Playbooks Were Written

(12:45) - Closing the First Few Customers

(14:22) - “If Nobody Hates What You’re Doing in OSS Than You Don’t Have Enough Users”

(18:55) - The IKEA of Relational Databases - AKA Tips on Branding

(21:12) - How Business Model is Brand Dependent - RedHat & Android as Examples

(24:19) - Existential Moment #2 - Turning that into an Asset Again

(29:20) - Establishing the Brand for Eucalyptus Software & HackerOne - “When Building a Brand You Must Polarize”

(33:32) - How to Approach New Market Creation

(36:57) - Getting Early Customers at HackerOne

(39:22) - Balancing a Two Sided Marketplace at HackerOne

(43:20) - Parallels Between Open Source Companies and HackerOne

(45:34) - AI Impact on Security

(52:06) - Joining An Existing Companies with Founders Still There as CEO

(54:51) - Establishing a Culture of Decentralized Decision Making

(57:05) - Knowing When to Centralize Key Decisions vs Maintaining Decentralization

(59:55) - “I’ve Decided to Be Happy Because It’s Good for my Health”

(1:00:02) - Failure over a 36 Year Leadership Career and the Contrast Between Rational Thinking and Instincts

(1:10:00) - The Intrinsic Motivation to Keep Climbing New Learning Curves

(1:12:25) - How to Operate in a Constrained Capital Environment

(1:16:13) - Why Consumerization of Everything (yes including IT) is an Inevitable Force

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

This episode is for whomever is curious about Secure Service Edge (SSE) or Secure Access Service Edge (SASE) Security and why seemingly every public security company seems to be trumpeting their solutions in this space which Gartner predicts will be one of the fastest-growing categories of spend in the coming years.

Kunal Agarwal is the Founder & CEO of Dope Security which is building a modern Secure Service Edge (SSE) platform. He was previously Sr Director of Product at Forcepoint and GM at Symantec. Kunal started his journey into security as a teenage hacker. I’m also joined by Emilio Escobar, CISO of Datadog as a co-host. In this episode, we dive into what is SSE & SASE Security including the different components that make up these solutions, why networking & security are converging now, and how design and implementation for modern systems can change how enterprises adopt security products.

Where to Find Kunal:

* Twitter: https://twitter.com/iamkunala

* LinkedIn: https://www.linkedin.com/in/kunala/

* Company Website: https://dope.security/

* CISO in Cars Podcast: https://dope.security/podcast

Where to Find Emilio:

* Twitter: https://twitter.com/eaescob

* LinkedIn: https://www.linkedin.com/in/emilioesc/

* Emilio’s Newsletter: https://emilioe.substack.com

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Software Snack Bites Newsletter: https://www.shomik.substack.com

* Software Snack Bites Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(00:42) - Kunal and Dope Security’s Background

(04:47) - SSE & SASE Security Definition

(06:31) - Zero Trust & SSE/SASE - Same or Different?

(08:40) - Why SSE & SASE Growth is Scaling Rapidly

(12:04) - What is a Secure Web Gateway (SWG)?

(13:38) - Are Firewalls Still Needed?

(15:43) - What are Cloud Access Security Brokers (CASB)?

(18:42) - What is Remote Browser Isolation and an Enterprise Browser?

(23:17) - Where Enterprise Adoption of SSE/SASE Is Currently

(27:33) - Where Existing SSE/SASE Solutions Fall Short

(28:57) - UI/UX Treated as First Class Need for Security Products

(35:20) - Who Has the Pain in Enterprises related to SSE/SASE

(37:07) - Why End User Design is Challenging at Legacy Security Companies

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

Many of us in the private markets are curious about how public markets view AI’s impact and the approach that public market investors take to investing in companies. This episode uncovers that from one of the top public software investors in my opinion.

Glen Kacher is Chief Investment Officer at Light Street Capital. He founded Light Street 13 years ago after stints at Integral Capital Partners and Tiger Global. In this episode, we cover a long term successful public market investor’s view on how AI will impact the software industry and other industries as well. We talk through where value seems likely to accrue in the interim and over the long term. We also cover how Glen and his team deep dive into new areas and form theses through deep research. Glen has nearly 30 years of public market software investing experience so needless to say this episode captures many of those earned insights.

Where to Find Glen:

* LinkedIn: https://www.linkedin.com/in/kacher/

* Twitter: https://twitter.com/GlenKacher

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Newsletter: https://www.shomik.substack.com

* Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(00:49) - Learnings from Julian Robertson, Founder of Tiger Global

(03:59) - Founding Light Street Capital

(05:51) - First Investment in the Fund & Adjusting Theses Through Business Model Changes

(10:06) - Learnings from the Dot Com Bubble

(12:06) - Technical Debt and Iterating with Customers Over Long Term

(15:35) - Research Process Behind Diving into AI

(18:13) - Is AI a Platform Shift, Enablement Shift, or Something Else

(21:06) - AI’s Impact on Software Margins and Efficiency

(26:09) - Software Moats in the Age of AI

(28:42) - Platform Companies Enduring Advantages

(30:32) - How AI Will Change Public Investing Workflows

(32:47) - Balancing the Micro & Macro and Optimism for the Future

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

This episode is packed with tactical advice for all things sales and go-to-market. It’s relevant across stages whether your company is at $0M ARR or at $100M+ ARR.

Ethan Schechter is VP Sales at Snyk. He started at Snyk 6 years ago as the first sales leader. In this episode, we cover tactical insights from Ethan’s experience leading early sales to scale at Snyk and ZeroTurnaround. He describes how to position a product when “shift left” was a new and scary term, how to make first sales, tradeoffs to consider with early vs late stage deals. He shares stories, anecdotes and advice for other startups about the earliest days of selling, hiring, and balancing various gtm strategies. I’ve been personally wanting to do this for awhile and Ethan absolutely delivered with a packed episode of learnings.

Where to Find Ethan:

* LinkedIn: https://www.linkedin.com/in/ethanschechter/

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Newsletter: https://www.shomik.substack.com

* Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(01:21) - Selling to Developers Before the Market Fully Formed

(03:58) - Effective Champion Enablement

(05:06) - Joining Snyk as 1st Sales Leader After Being CRO of ZeroTurnaround

(07:40) - Hiring an Effective Early Sales Leader

(12:02) - Closing Snyk’s Early Customers

(17:34) - Balancing Inbound and Outbound Sales in Early Days

(20:48) - Orienting the Whole Org Around the Customer

(23:59) - Positioning and Messaging the Product for the Right Personas

(26:27) - “Trading Dollars for Speed”

(29:23) - Creative Packaging and Pricing to Win Early Customers

(32:00) - Contract and Risk Acceptance

(33:57) - Common Mistakes with Early Sales Hires

(36:05) - Building the Early Sales Playbook

(37:52) - Advice for Early Sales Hires

(40:09) - Going from $0-$1M ARR

(41:19) - Scaling from $1-$10M ARR

(45:03) - Selling Multi-Product from Single-Product

(46:49) - Reinventing and Tweaking GTM Through Scaling

(49:02) - How to Scale Yourself

(52:47) - “The Minute You Stop They Win”

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

This episode is useful for all founders, startup employees, and investors who are interested in early stage startup advice, enterprise product tips, and learning how employees can scale from Day 0 to public markets.

Kevin Wang is Chief Product Officer at Braze. He started at Braze ~12 years ago as employee #5 going from software engineer, to PM, and eventually to CPO. In this episode, we cover learnings from Kevin’s incredible journey and how he as well as the company scaled from Day 0 to a public company. He shares stories, anecdotes and advice for other startups about the earliest days of building product for enterprises, acquiring customers, hiring, and much more. Kevin also talks about how Braze’s products delivers an end to end platform for enterprise customers across various orgs and use cases. This episode is packed with tactical insights and stories.

Where to Find Kevin:

* LinkedIn: https://www.linkedin.com/in/kevin-wang-96131916/

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Software Snack Bites Newsletter: https://www.shomik.substack.com

* Software Snack Bites Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(00:54) - Joining Braze as Employee #5 and Transitioning from Eng to Product

(04:23) - Braze’s Starting Strategy from Day 0 to Build a Successful Company

(08:56) - Stories of Losses that Felt Big at the Time but had Minimal Impact

(11:30) - Doing Things That Don’t Scale

(16:43) - Acquiring Earliest Customers

(18:08) - Product Market Fit Moment

(20:40) - Building for Enterprise Customers from the Beginning

(23:40) - Advice for Navigating $10 to $10M ARR Valley

(27:46) - How Kevin Scaled from Employee 5 to CPO

(33:26) - When to Dive Deep Directly vs Stay Higher Level

(36:42) - Enabling the Product to Spread Across the Org and Use Cases

(38:55) - Broadening the Ecosystem Around Braze Through Data Integrations

(40:22) - How Products Can “Land & Expand”

(44:07) - Focusing on Depth vs Breadth in Product

(45:51) - How to Remove Features

(47:43) - What Kevin Loves About the Braze Platform

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

This episode is useful for all founders, startup employees, and investors who are interested in product marketing, self serve, and PLG for developer focused, enterprise software products.

David Dorman is VP, Self Serve at Grafana and formerly Director of Product Marketing at DigitalOcean. In this episode, we cover learnings from DigitalOcean’s incredible content strategy, what Self Serve means and specific advice that David has for teams in this area, how developer marketing differs from Self Serve workflows, and advice for teams trying to improve their strategies in these areas. This is a tactical episode where David shares specific insights into how teams are structured, metrics to track, and where and how to interact with developers for the hosted product.

Where to Find David:

* Twitter: https://twitter.com/ddorman

* LinkedIn: https://www.linkedin.com/in/davedorman/

* Personal Blog: https://dorman.io/

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Software Snack Bites Newsletter: https://www.shomik.substack.com

* Software Snack Bites Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(01:02) - Learnings from DigitalOcean’s Content Strategy

(05:53) - Product Marketing’s Roles and Tactics

(10:36) - How Self Serve Differs Between DigitalOcean and Grafana

(16:58) - Early Stage Advice for Approaching Self Serve

(22:28) - Structuring Self Serve Teams

(30:08) - How to Manage Multi-Product User Engagement

(31:53) - Anti-Patterns in Self Serve

(35:04) - Interplay of Dev Rel /Marketing and Self Serve

(39:37) - How to Position Products in Competitive Markets

(43:45) - Most Common Advice for Startups

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

This podcast episode is a new format where I introduce everyone to my go-to brain trust to chat new technologies with. This episode can be consumed end to end or you can jump to specific categories that particularly interest you (timestamped below).

Alex Mackenzie is a Partner at Tapestry VC and Ian Livingstone is a Team Lead at Snyk. In this day and age, there’s a new technology that seems to capture the interest of the ecosystem (particular tech Twitter) every month. When I need to dive into these technologies or have them explained, I go to Alex & Ian for help understanding why they are interesting. In this episode, we cover a variety of areas and purposefully go deep enough to understand why they’re interesting but not as deep as we could’ve gone to cover the most surface area.

We cover the following areas: Javascript Runtimes & Frameworks (Next.js, Bun, Deno), Edge Computing, Emerging Programming Languages (Zig, Typescript), Observability (Open Telemetry, eBPF), Encryption (Secure Enclaves, Multi-Party Computation), & AI (LLMs, Agents, Embeddings, WebGPU).

Where to Find Alex:

* Twitter: https://twitter.com/alex__mackenzie

* LinkedIn: https://www.linkedin.com/in/alex-mackenzie-6aa80ab4/

* Blog: Why Now

Where to Find Ian:

* Twitter: https://twitter.com/ianlivingstone

* LinkedIn: https://www.linkedin.com/in/irlivingstone/

* Podcast: Yet Another Infra Deep Dive

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Software Snack Bites Newsletter: https://www.shomik.substack.com

* Software Snack Bites Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(01:18) - Javascript Runtimes (Bun, Deno)

(09:31) - Javascript Frameworks (Next.js, Gatsby.js)

(12:45) - Edge Computing

(21:03) - Emerging Programming Languages (Zig, Typescript, Dart, Mojo)

(28:53) - Observability (eBPF, Open Telemetry, Open Policy Agent)

(36:20) - Encryption (Secure Enclaves, Multi Party Computation, Federated Learning, Homomorphic Encryption)

(45:22) - AI (LLMs & Agents, Observability for LLMs)

(52:59) - AI (Fine Tuning & Inference)

(56:39) - WebGPU

(58:42) - How Higher Fidelity in Data Maps to Security Trends

(60:00) - Godot

Show Notes:

* Going Beyond Transformer Architectures for LLMs - Alex’s write up explaining the Hyena Hierarchy Architecture

* Zig - Why is Zig an exciting programming language

* How Will the World Run on GPUs - Ian’s podcast about GPU hosted clouds

* Is Front End Eating the Whole Stack - Ian’s podcast deep diving into front end developments that were mentioned in the beginning of this podcast

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.

This podcast episode is for all of those who are interested in trends like cloud data warehouses, data & AI, edge networks, SASE security, zero trust security and companies like Snowflake, Databricks, Cloudflare, Z-Scaler, Palo Alto Networks, and of course the big cloud providers.

Muji is the Founder of Hhhypergrowth, which is my favorite research publication for all things software infrastructure and security. I HIGHLY RECOMMEND subscribing. Can’t say enough about how worth it Muji’s research service is even for folks who are deep in infra and security as I am. I first came across Muji when he published an incredibly detailed piece on Snowflake at the time of the IPO (linked in Show Notes) that of course talked about the financials and business model, but more so talked about the technical architecture and why that was helping Snowflake win customers so rapidly. Furthermore, he dove into why that architecture enabled massive product adjacencies to move towards in the future (most of which have emerged).

In this podcast, we cover all the cloud data infrastructure, security, and edge networking trends we all have been wondering about. As well as how various products/companies in these spaces may be converging over certain workloads/use cases and why.

Where to Find Muji:

* Twitter: https://twitter.com/hhhypergrowth

* Website: https://hhhypergrowth.com/

Where to Find Shomik:

* Twitter: https://twitter.com/shomikghosh21

* LinkedIn: https://www.linkedin.com/in/shomik-ghosh-a5a71319/

* Software Snack Bites Newsletter: https://www.shomik.substack.com

* Software Snack Bites Podcast: Apple Podcasts, Spotify, Google.

In this episode, we cover:

(01:15) - Overview of Cloud Data Infra - Trends and Architectures (including differences between Snowflake, Redshift, BigQuery, Databricks)

(11:12) - Operational vs Analytical Databases

(15:13) - Databricks & Snowflake Differentiation (especially as it relates to AI)

(19:35) - Overview of Edge Networks - Trends & Architectures (Cloudflare, Fastly, AWS)

(25:04) - Edge Network Providers vs Big 3 Cloud Providers

(31:48) - Overview of Security & Networking Convergence (SASE Security, Zero Trust Networking, SSE, CNAPP, etc)

(37:15) - Public Company SASE Security Landscape (Cloudflare, Z-Scaler, Palo Alto Networks)

(46:03) - XDR/EDR & Zero Trust Converging (Crowdstrike, SentinelOne, Microsoft, Palo Alto Networks)

(48:06) - How Muji Conducts In-Depth Technical and Qualitative Research

(55:23) - Audience Question: Big Cloud Providers vs Nvidia in ASICs and GPUs

(58:39) - Audience Question: Outside of AI, What Area of Software Infra Still Has the Most Innovation and Growth Ahead?

Show Notes:

* Snowflake Deep Dive - Must read for anyone wanting to learn more about Snowflake

* Flavors of Security - Breaking down the various components of security and why areas like the convergence of Networking & Security are so interesting

* Modern Data Stack Overview - Deep historical overview of how data infrastructure has evolved and what is occurring now

* Hhhypergrowth Subscription Page - No brainer to subscribe for paid posts

How to Subscribe:

Available on Apple Podcasts, Spotify, or Google.