The industrialization of software development
David Peterson

Everyone is debating whether AI will kill software companies. But they’re missing the point. The interesting story isn’t whether people will replace their hardened Salesforce instance with a few well-articulated prompts (they won’t). Or whether decades of code complexity and integration know-how will form an impenetrable moat (it won’t). The interesting story is what happens when code becomes free.

Last week, a company shipped production code that no human wrote and no human reviewed. The AI agents that generated the code also wrote the tests, ran those tests in mock environments of Slack and Okta, and deployed the changes. The humans, for their part, got an update when the feature went live.

This isn’t a demo or a research project. This is how the team at StrongDM builds software.

Every wave of technological innovation has been catalyzed by the cost of something expensive trending towards zero. That’s what we’re witnessing right now. Code is (basically) free. The question is: when code is free, what becomes expensive?

“Dark factory” software development

I’m not sure if what StrongDM has invented is the future, but, per Simon Willison, it’s the most vivid version I’ve come across.

A few months ago, the team at StrongDM set two rules for themselves: no code can be written by humans and no code can be reviewed by humans. Those constraints led them to a system they call a software development “dark factory,” inspired by a fully automated manufacturing facility that can operate without lights because there are no humans that need to see.

This was just recently possible, according to the team. The catalyst was Opus 4.5 and GPT 5.2. These models were good enough, finally, to “compound correctness.” In other words, the more the agent "worked" on the system, the more stable it became, rather than more fragile. This is the unlock that allowed the StrongDM team to stop human review entirely and start letting the factory run “dark.”

The hardest problem with letting agents run wild is testing. Having agents write tests only helps if they don’t cheat, but agents, as we all know, really like to cheat. 

The solution that StrongDM came up with is twofold. First, instead of unit testing they do scenario testing. Think of “scenarios” as end-to-end user stories stored outside of the codebase (almost like a “holdout” set in machine learning). Success, measured by user satisfaction, is judged by an LLM evaluation of whether the path would likely satisfy a real person.

Second, they built what they call a “Digital Twin Universe,” which is a set of high-fidelity clones of third-party APIs (Okta, Slack, Jira, etc.). This lets their agents run thousands of integration tests against "fake" versions of the internet without hitting rate limits or incurring real-world costs. More importantly, they can test failure modes that would be impossible to test in production. What happens if Okta goes down for 10 minutes? Now they can find out.

This is a fascinating example of what software development looks like when code is nearly free: high-fidelity mocks of the real world were always desirable but economically impossible, and now they're not. It also nods at the challenge (and expense) of handling validation at scale.

From products to processes

Taking the StrongDM team’s approach one step further, I recently spoke with a team that has no fixed product at all. Instead they have specs.

They interview the company and gather all the tribal knowledge they can. They ingest the context from existing software systems. And they use their “factory” to spit out a product that fits the company like a glove.

It looks like a high-end consulting engagement, but the deliverable is a proprietary product with the deployment speed of a SaaS implementation. If your factory can generate, test and war-game a bespoke product in the time it takes to have a discovery call, what’s the difference between a consulting shop and a software company?

Where value accrues

When code itself stops being a moat, where does value accrue? A few ideas:

The factory. Who has the most rigorous testing? The most realistic digital twins? StrongDM's Okta clone encodes years of "how Okta actually behaves when misconfigured." That operational knowledge compounds over time. The teams running agents longest will probably have the best factories. And this isn't something you can catch up to quickly. It's accumulated knowledge. (Though I do wonder if this just gets open-sourced. Why wouldn’t it? Surely the wisdom of the crowd will be better eventually, even if there is some proprietary edge today.)

State monopolies. You can clone Salesforce's code in a month, but you can't replicate the fact that every sales team is already using it, with years of workflow dependencies and integrations. AI can't hallucinate a user’s historical state or the network of people already using the platform. The software is reproducible, but the adoption, and all the associated history, isn't.

Trust and certification. When no human reviews the code running your payroll system, who's liable when it breaks? Who gets audited? Who carries insurance? Established companies can point to years of compliance work, security certifications, enterprise relationships. In regulated industries, this becomes the primary barrier to entry. Trust is expensive to build. It requires time, capital, and proving yourself in lower-stakes environments before enterprises let you touch production systems.

Everyone assumes the commoditization of code will lead to the democratization of software. But what if it does the exact opposite?

When everyone can generate code for free, competitive advantage shifts to things more expensive than code ever was. Building “digital twins” that simulate the internet and then running thousands of experiments knowing 99% will fail. Earning enterprise trust through years of certification and compliance work. These are not inexpensive endeavors.

Shipping code may be free. But the cost of shipping code that will win is getting way higher.

It may seem that I’m going to conclude that incumbents are the inevitable beneficiaries. But I think startups are actually best positioned to build the first software factories. My contrarian take, though, is that startups will need more capital in this new era, not less. The lean startup era is giving way to capital-intensive competition at factory-scale. And the companies that win won't be the ones writing the most elegant code. They'll be the ones who can afford to run the most sophisticated factories.

This isn’t the end of software. But it is certainly the end of artisanal software, coded meticulously by hand.

The debate about whether AI will "kill" software companies misses the transformation entirely. We're not watching software die. We're watching it industrialize. And the companies that figure out how to build factories, not just products, are the ones that will matter.

Could the future of software be fluid
How do we get the best of AI without losing the soul of software?

The future belongs to young missionary teams
Why it makes more sense betting on youth in the current moment

The AI-native enterprise playbook
Ten real-time observations on a rapidly evolving playing field

No more painting by numbers
It’s the end of the “SaaS playbook.

ENTERPRISE/TECH NEWS

The SaaSpocalypse. Last week, about $300B was wiped off the value of public SaaS stocks. According to most analysts (in this case, Reuters), “software stocks have been under pressure in recent months as AI has gone from a tailwind for many of these companies to a possible disruption. The latest selloff was triggered by a new legal tool from Anthropic's Claude large language model.” As Don Muir wrote in Forbes, “Markets do not erase that much value unless a core assumption breaks.The assumption that just broke was the durability of legacy SaaS growth.” The so-called SaaSpocalypse rapidly became the central (only?) major topic of discussion in tech land for the rest of the week. 

One of the sharpest expressions of the SaaSpocalypse view was from Sam Lessin, in a linkedin post. “Meanwhile, the thing nobody in enterprise wants to admit: the SaaS era is over. I said this in 2024. Software is a business tool, not a business model. The moats companies spent a decade building are dissolving in real time. AI doesn't just compete with your product -- it makes the entire category feel optional.”

Jason Lemkin tweeted a solid breakdown that touched on the subtleties. “SaaS isn’t dead, even with the massive sell off this year (and really since July 2025) But … what is dead is the classic pattern and way to scale.” The tweet itself is short but a must-read.

Steven Sinofsky, board partner at A16Z, wrote that the death of software has been greatly exaggerated. On the contrary, he argues that there will be more and more powerful software than ever before: “AI-enabled or AI-centric software is simply moving up the stack of what a product is. Software did not create online banks. Banking always required software. Software that faced a consumer, banking or traveling or shopping or reading or viewing, just became an essential part of the bank, travel, etc stack. Sometimes this created new from-scratch companies and sometimes it created new companies inside old ones. Industries were restructured as assets moved around. However big and complex you think a legacy business is today, it will be vastly larger and more complex tomorrow, and it will do vastly more. Think I’m crazy? Consider what banking was like in 1995. If you have any experience you know your choices, features, options, etc were one-thousandth of what you have today, even if fundamentally you got a paycheck, paid your bills, and might have had a credit card.”

Stefan Waldhauser made an even more powerful argument a few days earlier: “First, the blanket statement “Enterprise SaaS is dead” is nonsense - if only because you can’t lump all enterprise software systems together. It’s important to first distinguish between a System of Record (SoR), a System of Engagement, and a System of Intelligence/ Automation because these layers of software value creation face very different challenges in the age of AI…Enterprise software will not disappear with the advent of agents, but it will change as more actions are performed by AI than by humans. The system of record will maintain the status quo and become more important. The system of engagement will lose importance as a human work interface and has to change. The system of intelligence/automation will become the valuable layer that performs work across all systems.”

And finally, a humorous take, from Alex Iskold: “Why would enterprises stop at just building their own software? After they finish vibe coding their own CRM, Databases, and re-implement TCP/IP to make it absolutely perfect for them, I think they should also build buildings, make their own security cameras, pens and rugs.

Israel’s AI ecosystem. TLV Partners published a map of the Israeli AI infrastructure landscape - and we were excited to see four Angular companies (Firebolt, Dualbird, Specific, and FalkorDB) on the list!

HOW TO STARTUP

The future of product in the AI era. The incredible Gokul Rajaram, with whom we are fortunate to have co-invested a few times, gave a wonderful and deep interview to Patrick O’Shaughnessy. “As a legendary operator (Google, Facebook, Square, DoorDash) and prolific investor, Gokul breaks down how tools like Claude are merging design and engineering roles and why human "judgment" is the only future-proof skill. They explore the durability of software companies in an agentic world, the mechanics of building massive ad businesses, and lessons learned from working directly with leaders like Larry Page, Mark Zuckerberg, and Jack Dorsey. Finally, Gokul shares a framework for career growth and hiring "builders" in the AI era.” It’s a must-listen. 

Is being a VC-backed founder low-status? Michael Dempsey wrote a really interesting article suggesting that being VC-backed has become a default (and, therefore, low status) option for many young people. “This is about the path itself becoming the default. It is about a world in which starting a venture-backed company has become the thing you do when you’re ambitious and want to be perceived as smart, in the same way that going into banking used to be the thing you did when you were ambitious and wanted to be perceived as successful. When a path becomes the default, it no longer says something meaningful about you, but instead starts being an optimization problem. And optimization problems are increasingly not that interesting…When the institutions are optimized to fund the legible thing and the individuals are optimized to build the legible thing, the identity of the founder itself fades. Being a founder used to carry with it the implication that you had seen something others hadn’t, that you were willing to be wrong in public about an unlikely future you believed in, and take a risk of banging your head against a wall with high career risk for a long time. Now it is almost eye-roll inducing in many non-tech circles as people struggle to distinguish Another Founder with Another Launch Video building Another (insert zeitgeist here) Startup. The noise outpaces the signal as the system is optimized to produce as many founders as possible, as safely as possible. The legible thing becomes the average thing, and thus the average thing becomes the low-status thing.”

HOW TO VENTURE

Checking for loyalty. Wired reported on increasing job turnover among the most talented AI researchers, and how VCs are reacting by emphasizing how committed teams are to each other and to the mission before investing. “Early founders and researchers at the buzziest AI startups are bouncing around to different companies for a range of reasons. A big incentive for many, of course, is money. Last year Meta was reportedly offering top AI researchers compensation packages in the tens or hundreds of millions of dollars, offering them not just access to cutting-edge computing resources but also … generational wealth. But it’s not all about getting rich. Broader cultural shifts that rocked the tech industry in recent years have made some workers worried about committing to one company or institution for too long, says Sayash Kapoor, a computer science researcher at Princeton University and a senior fellow at Mozilla. Employers used to safely assume that workers would stay at least until the four-year mark when their stock options were typically scheduled to vest. In the high-minded era of the 2000s and 2010s, plenty of early cofounders and employees also sincerely believed in the stated missions of their companies and wanted to be there to help achieve them.”

Angular Ventures founding partner Gil Dibner joins CTech for its 2026 VC Survey.

CruxOCM CEO & Co-Founder Vicki Knott talks to Collide about how she built a tech company that blends first-principles engineering with machine learning.

DualBird closes on $25 million in push for cloud-native acceleration of AI workload.

FalkorDB
Developer Relations Advocate - GenAI & Graph Systems (Bay Area)

Paradime
Product Engineer - Backend (Remote)

Reco
Senior Sales Engineer (US, Remote)