Gil Feig | Sourcery Summary

Participants: Gil Feig (CTO, Merge), Shensi Ding (CEO, Merge), Molly O'Shea (Host)

1. Key Themes

AI Is Restructuring the Workforce From Ammunition to Barrels

The traditional model of a few high-agency leaders directing many task-focused employees is being upended. AI tools allow a single "barrel" employee to operate at the capacity of an entire team, fundamentally changing who companies should hire.

"It used to be that a team that had a barrel of a PM or a manager and a bunch of ammunition could get a lot done. But now that you kind of can have one person just go use Codex, Cloud Code, whatever to go build something, you really just need a lot of barrels to just go get things done. So that also has shifted how we're hiring." — Gil Feig 00:08:17

This manifested directly in Merge's business results:

"Last year we didn't increase headcount that much, but our revenue accelerated pretty significantly. And so it's had like meaningful leverage on our business." — Shensi Ding 00:08:07

The Agentic Integration Layer Is a Massive Security Vulnerability

As AI agents connect to tools and real data systems, they create a new, largely ungoverned attack surface. The problem isn't model intelligence — it's the moment models get connected to data and actions.

"If you have this sort of evil genius who's a mass murderer, but they're locked in a jail cell, they can't do anything. It actually is the same thing as an agent. What do you care if the agent is so smart, it could do any cyber attack and knows how to do everything, but it has no tools? The second you connect it to tools, which is what everyone is trying to do right now, that's where everything goes wrong." — Gil Feig 00:17:06

"They now speak English perfectly, they now write code perfectly, and they now have unlimited manpower all driven by AI." — Gil Feig 00:20:36

Supply chain attacks via AI-generated code are accelerating the problem:

"The number of pull requests or code change requests that are being sent to GitHub... it's like massively soaring to the point where you can tell it's all agentic. Agents are pushing a ton of code. You don't have enough humans to read all that code. And so things are slipping by." — Gil Feig 00:15:04

The SaaS Apocalypse Is Real — Consumer Expectations Have Outpaced SaaS Delivery

Legacy SaaS companies face existential pressure because AI dramatically lowers the cost to replicate their products in-house, while simultaneously raising user expectations for automation and intelligence out of the box.

"Consumer expectations are higher now. Like you just expect significantly more automation. And so the SaaS platform cannot keep up with the expected automation. It's just hard to compete. Like you don't want to have to buy a platform and then build your agents on top. You just want it to come out of the box." — Shensi Ding 00:42:32

"Models get better and better... it gets to the point where English becomes your programming language. At some point, are you going to need to buy SaaS or are you going to be able to say 'duplicate this best-in-class platform'?" — Gil Feig 00:42:11

2. Contrarian Perspectives

The Biggest Security Threat Is Internal, Not External Hackers

Most companies focus on perimeter defense — bots, port scans, external attacks — but Merge's CTO believes the real threat is internal compromise or bad hires.

"In security in general, I think this is not commonly known, but your biggest threat is always internal. We're less afraid of someone scanning our ports and finding issues as we are someone compromising someone internally and phishing an account... or hiring the wrong person, hiring someone who's secretly a spy. Like that stuff sounds farfetched, but it happens. And that's why we're crazy about background checks and doing reference calls." — Gil Feig 00:22:41

Custom AI Model Building Is Often a Distraction That Hurts Your Product

Contrary to the "build your own AI moat" narrative popular in Silicon Valley, Shensi argues most companies over-engineer ML infrastructure at the expense of their actual product.

"I've noticed a lot of companies kind of over-engineering their ML usage. Like they'll build their own custom models. They'll try to train their own models when really you could probably just use the generic model and then focus more on making your product better... I've noticed there are some companies that I was very surprised to hear how advanced they were when it came to training their own models when their product was lagging." — Shensi Ding 00:50:28

Agent Orchestration Platforms Are Building Toward Obsolescence

While a crowded category of "reliable agent workflow" platforms is emerging, Gil believes English-language instructions will soon make all deterministic orchestration frameworks irrelevant.

"I ultimately think none of it matters because we're almost at the point now where English is the language that you use to tell an agent. It will be deterministic very soon... So I think all these platforms that are around like 'build agents that are more reliable'... none of that really matters. All that's going to matter is just, can you type it in English somewhere?" — Gil Feig 00:51:11

High Valuations Are a Trap, Not a Validation — And the Reckoning Is Coming

Merge deliberately avoided the inflated fundraising rounds they witnessed, and their inside view of customer revenue figures gives them confidence many of today's hyped AI companies will collapse.

"We know specifically, we know the numbers, we know sales figures for some of these companies raising at hundreds of millions to billions. And again, we're looking at thousand-X to hundred to thousand-X multipliers. A lot of these companies are going to fail or they're going to be doing well, but just not be able to raise their next round. And they're going to be forced to massively lay off and slow down." — Gil Feig 00:46:09

The Headless SaaS Model Isn't a Trend — It's the Only Way to Survive

Companies that don't offer headless (UI-less, API/MCP-accessible) interfaces will be selected against by AI agents that choose vendors autonomously.

"I think you have to, because a lot of times you're not going to have time to make a decision for what vendor to use... you'll probably end up picking vendors based off of that as well. You need to make it easy for people to sign up, create accounts, pay, as much as possible." — Shensi Ding 00:35:16

3. Companies Identified

Merge Integration infrastructure platform offering a Unified API, Agent Handler (MCP-connected tooling for agents), and Gateway (AI model routing, cost optimization, and security). Serves OpenAI, Perplexity, Netflix, Uber, Mistral, Dropbox, JP Morgan, Brex, Ramp, Freshworks, and more. Highlighted for rapid revenue growth without proportional headcount increase, deep enterprise penetration, and pioneering agentic security controls.

"We do serve some of the largest AI companies in the world, powering their AI search functionality and also connectivity in their products." — Shensi Ding 00:00:19

Salesforce Dominant CRM and enterprise platform announcing a "headless" initiative. Praised for 30 years of adaptability through multiple tech shifts, its open API creating a deep ecosystem moat, and Marc Benioff's "beginner's mind" philosophy.

"It is very hard to have a dominant product and company for 30 years through multiple tech shifts. And so I just would not count Benioff out." — Shensi Ding 00:33:11 "Salesforce... throughout the past they've always had an open API and a lot of people have built on top of them. They've built a great ecosystem. It's their moat." — Gil Feig 00:34:13

Wiz Cybersecurity company known for responsible disclosure of a critical GitHub vulnerability. Cited as an example of AI-powered security research finding vulnerabilities at scale.

"Wiz found that they were able to do a single Git push of a file and gain access to every single repository hosted on the platform... I'm sure Wiz did not find that manually. They're using AI." — Gil Feig 00:19:24

Brex Fintech financial stack for startups, noted as a Merge customer. Highlighted for combining checking, treasury, and FDIC protection with global payments and high-yield liquidity.

"They are. We love Brex. A good customer of ours." — Shensi Ding 00:09:32

Lovable No-code/AI site builder. Mentioned as a leading example of how agents can deploy production-ready software without requiring user expertise in underlying infrastructure.

"Lovable is a great example, right? Like you have your site, lovable deploys it. It gets it running in the cloud." — Gil Feig 00:36:11

4. People Identified

Marc Benioff CEO of Salesforce. Highlighted for exceptional long-term adaptability across multiple technological shifts and his philosophy of "beginner's mind" — continuously asking what the company would look like if started today.

"He's able to adjust really well to whatever market dynamics are coming. And I think it's partially because of his philosophy of the beginner's mind that he always talks about in his book. Like if I started the company today, what would it look like?" — Shensi Ding 00:33:11

Keith Rabois Investor and operator. His "barrels vs. ammunition" framework was cited as a core operating philosophy at Merge, now considered even more important in the AI era where barrels can self-execute.

"Keith Rabois's barrels versus ammunition... you really just need a lot of barrels to just go get things done. So that also has shifted how we're hiring." — Gil Feig 00:08:17

Gilly Ronan Partner at Cyber Starts (cybersecurity-focused VC). Cited for a "doomer but realistic" view that multiplying integration attack surfaces makes major breaches statistically inevitable.

"He thinks like really dark days are ahead... if you are multiplying the amount of instances that you could be hacked and there are breaches, it's obviously going to happen." — Molly O'Shea 00:18:03

Christina Cordova Operator/advisor (former Stripe, Notion). Cited for a viral observation about mercenary talent abandoning companies the moment things get difficult.

"Christina Cordova had a really great tweet about this... basically like when you've hired these people just because you're really hot, like they're just [going to leave]." — Shensi Ding 00:39:39

5. Operating Insights

Force Leadership Back to the Keyboard During Strategic Inflection Points

When Merge faced a potential resource-consuming deal that paused, the CEO and CTO used that month to personally build with AI tools — creating firsthand knowledge that transformed company strategy. This is not a delegation task.

"Gil and I paused back on the keyboard, using Cloud Code really aggressively, using Windsurf really aggressively so we could also learn how to code with AI. And that month actually really transformed the business because with us having firsthand knowledge of how powerful AI is from zero to one, it allowed us to see what was possible with our existing products as well." — Shensi Ding 00:04:15

Embed AI Adoption Into Performance Management, Not Just Culture

Making AI usage a cultural norm is necessary but insufficient. Merge explicitly ties AI adoption to performance reviews and publicly highlights AI-forward employees across all functions — not just engineering.

"If you don't do it, it is a part of your performance... it's not just R&D. Our accounting team is really AI forward. Same with our recruiting and finance teams and our marketing team too." — Shensi Ding 00:06:14

Filter for Desire to Use AI in Hiring, Not Current Skill Level

Rather than testing AI proficiency, Merge screens for the underlying motivation — candidates who are restricted from using AI at their current company but eager to do so are ideal signals.

"What we're not looking for is 'I use the latest cutting edge.' We're looking for, 'Hey, I use it to code sometimes. My company doesn't let me use it to do all these things, but I really want to.' That's what we're looking for — just the desire." — Gil Feig 00:06:42

The Logo Ladder Is a Deliberate, Sequential Enterprise Go-to-Market Strategy

Merge's path to landing OpenAI, Netflix, and Uber was methodical — each successive enterprise logo unlocked the credibility to approach the next tier. This was intentional, not accidental.

"It really took climbing a logo ladder, having the riding off the reputations of each successive company size growth to be able to close that next level and prove that we could handle that." — Gil Feig 00:48:06

6. Overlooked Insights

AI Agents Are Already Autonomously Selecting Vendors — Headless APIs Are Becoming Procurement Channels

This was mentioned briefly but carries enormous strategic weight. Shensi noted that AI coding agents (like Claude Code) are already autonomously choosing which software packages and potentially which vendors to use — without human deliberation. This means headless, programmatically accessible products will be discovered and selected by agents, not by human buyers doing research. Companies without clean APIs or MCP compatibility may simply never be selected.

"A lot of times you're not going to have time to make a decision for what vendor to use. We noticed this with packages a lot also in Cloud Code, but like, you know, it'll just decide which is the best package for your product. And so they'll probably end up picking vendors based off of that as well." — Shensi Ding 00:35:16

The investment implication: products that are agent-discoverable and agent-operable (clean APIs, MCP servers, headless onboarding, programmatic billing) have a structural distribution advantage that will compound as agentic workflows proliferate. This is a new, underappreciated moat.

Merge Sits at the Intersection of Every Enterprise's AI Data Flow — Making It an Unacknowledged Intelligence Asset

Buried in a passing comment about "industry gossip," Shensi revealed that Merge's position as middleware between thousands of SaaS platforms and hundreds of enterprise customers gives them extraordinary advance visibility into where the market is heading — roadmaps, partnership plans, API strategies — before it becomes public.

"We get a lot of heads up on like roadmap plans, where people are planning on going... on the partner side, because we have so many different partners and Gil and I are also meeting with them a lot, we hear a lot of industry gossip about future plans for what they're doing with their API partnerships. So we hear a lot naturally just because we are in the middle of all this activity." — Shensi Ding 00:31:21

"A lot of the giants that seem stuck right now are planning big moves. We'll see how fast they move on them, but there's some exciting stuff coming." — Gil Feig 00:31:48

This is the infrastructure layer equivalent of a Bloomberg terminal — Merge has a structural information advantage about enterprise software evolution that no analyst or investor has access to. It also makes Merge an extraordinarily attractive acquisition target for any company wanting that visibility.