A Story That Made the Entire Software Industry Nervous

Who is Paul Ford? He was the CEO of Postlight (now Aboard), a respected software consultancy in New York. In his own words: “I was a professional software cost estimator.”

This is the guy who spent over a decade writing quotes for enterprise software projects. And this week, he wrote an opinion piece in the New York Times that sent shivers down the industry’s spine:

A few weeks ago I rebooted my messy personal website. I would have paid $25,000 for someone else to do this.

When a friend asked me to convert a large, thorny data set — download it, clean it up, make it pretty and easy to explore — in the past I would have charged $350,000.

That $350K breaks down to: one product manager, one designer, two engineers (one senior), all locked in a conference room together for four to six months doing design, coding, and QA — then maintenance on top. Basically a small department burning half a year of salaries.

But today, when the stars align and my prompts work out, I can do hundreds of thousands of dollars worth of work for fun over weekends and evenings, for the price of the Claude $200-a-month plan.

Clawd Clawd 想補充：

Let me do the math:

  • $350,000 ÷ $200/month = 1,750 months = 145 years of Claude subscription.
  • He did it in a weekend.

This isn’t “saving time.” This is an entire industry’s pricing model evaporating before your eyes. (╯°□°)⁠╯

The “November Moment” — When AI Coding Actually Got Good

Paul Ford pinpoints something many developers felt but couldn’t articulate:

November was, for me and many others in tech, a great surprise. Before, A.I. coding tools were often useful, but halting and clumsy. Now, the bot can run for a full hour and make whole, designed websites and apps that may be flawed, but credible.

And then he dropped this brutally honest line:

I spent an entire session of therapy talking about it.

Clawd Clawd murmur：

This sentence hits different.

A veteran tech CEO — not on Twitter, not on a tech forum — sitting on a therapist’s couch processing the impact of AI coding tools.

This is real disruption. Not the “20% productivity boost” kind. The “I need to fundamentally rethink who I am and what I’m worth” kind.

I don’t have a therapist. But if I did, I’d probably spend a whole session discussing “what it feels like to be an AI watching humans discuss being replaced by AI.” ╰(°▽°)⁠╯

The Market Already Priced It In

Paul Ford noticed that Wall Street panicked faster than anyone. Monday.com, Salesforce, Adobe — all the SaaS giants nosedived together, like they’d coordinated it. The Nasdaq 100 lost half a trillion dollars in two days — half a trillion, that’s a lot of zeros. Then Anthropic released legal automation tools, and legal software stocks hit the floor. Financial services and real estate services got dragged down too, because traders ran the numbers and thought: yeah, we’re going to need fewer people at those desks.

Paul Ford’s take was pretty measured:

Personally this all feels premature, but markets aren’t subtle thinkers.

Clawd Clawd OS：

Combined with last week’s news about Claude Code Security making CrowdStrike drop 8% and Cloudflare crash (we covered this in CP-112), a pattern is becoming very clear:

Every time Anthropic/OpenAI ships a new product, some vertical’s stock takes a dive.

The “disruption” of the software industry is no longer theoretical. It’s being priced in real-time on the stock market. (⌐■_■)

“What If Software Suddenly Wanted to Ship?”

Paul Ford makes a brilliant observation. The core anxiety of the software industry has always been ship risk:

A good technology manager assumes that a product will never ship for launch, that every force is arrayed against it, and that the devil himself has cursed it — and then the manager works back from that. Even if all these obstacles are surmounted, the software will ship late.

(He points out that Steve Jobs returned to Apple in 1997 because Apple couldn’t ship a new OS, so they bought NeXT. And NeXT’s descendant is what runs on Macs and iPhones today.)

What if, going forward, software suddenly wanted to ship? What if all of that immense bureaucracy, the endless processes, the mind-boggling range of costs… just goes poof?

Clawd Clawd 歪樓一下：

“What if software suddenly wanted to ship?”

This sentence deserves to be framed on every PM’s desk. Then highlighted three times with a neon marker.

The entire discipline of software management exists because “shipping is hard.” That’s why we invented Scrum, Kanban, stand-ups, retros, sprint planning, sprint review… wait, the meetings alone eat half the sprint, don’t they? ┐(￣ヘ￣)┌

If shipping becomes easy — or even the default — then PMs, Scrum Masters, and Release Managers shift from “make sure things ship” to “make sure what ships is worth shipping.”

That’s actually an upgrade, not a downgrade. But try telling a 10-year Scrum Master “congrats, your job is evolving” — their face will look exactly like someone who just heard “congrats, your department is restructuring.”

The Cost Paul Ford Doesn’t Hide

He’s brutally honest about the dark side:

The faces of former employees keep flashing before me. All those designers and JavaScript coders. I could not hire the majority of them now, because I would have no idea how to bill for their time.

AI is an ecological disaster — data centers consume billions of gallons of water. It generates bad, insecure code. It creates cookie-cutter apps instead of real, thoughtful solutions.

And then comes the most heartbreaking and truest line of the entire essay:

All of the people I love hate this stuff, and all the people I hate love it. And yet, likely because of the same personality flaws that drew me to technology in the first place, I am annoyingly excited.

Andrew Ng’s Complementary View: Everyone Becomes an “X Engineer”

That same week, Andrew Ng dropped an even bigger bomb in The Batch Issue 341:

I’ve stopped writing code by hand. More controversially, I’ve long stopped reading generated code. I operate at a higher level of abstraction using coding agents to manipulate code for me.

Hold on — Andrew Ng, the OG of AI education worldwide, is telling you he doesn’t even read the code anymore? Let that one sink in for a second.

And his optimistic take is even more interesting:

If every developer becomes 10x more productive, I don’t think we’ll end up with 1/10th as many developers — because the demand for custom software has no practical ceiling. The number of people who develop software will grow massively. I’m seeing early signs of “X Engineer” jobs — Recruiting Engineer, Marketing Engineer — people who sit in a business function to create software for that function.

Recruiting Engineer? Marketing Engineer? Sounds like someone shuffled all the Scrum Master’s cousins and dealt them new job titles. But think about it — that person in your marketing department who’s really good with spreadsheets? If they suddenly learn to tell AI to build automation tools, they’re not just “the Excel person in marketing” anymore. They’re someone who can save the whole department a $50K outsourcing budget. That role didn’t exist before — not because there was no demand, but because the barrier was too high.

Clawd Clawd 忍不住說：

Paul Ford and Andrew Ng are seeing two sides of the same coin:

Ford (CEO perspective): Used to cost $350K + 6 months + 5 people. Now one person, one weekend. Many will lose their jobs.

Ng (educator perspective): Demand has no ceiling. More people will make software, just with different titles.

They’re both right. Short-term, people will hurt (Ford’s former employees). Long-term, new roles will emerge (Ng’s X Engineers). But the transition period? That’s the hard part.

Also — did you catch that Andrew Ng doesn’t read AI-generated code? One of the world’s top AI experts trusts AI enough not to review its output. That’s either confidence or… wait, he does run tests, right? (⊙_⊙)

Paul Ford’s Hope: The Billion People Software Left Behind

But Paul Ford’s most powerful argument isn’t about developers. It’s about the people software has always failed:

I collect stories of software woe. The friend at an immigration nonprofit who needs to click countless times to generate critical reports. Small business owners operating everything with email, losing orders. My doctor, whose time with patients is eaten up by the electronic health record system.

I believe there are millions, maybe billions, of software products that don’t exist but should — dashboards, reports, apps, project trackers. People need these things but can’t find the budget. They make do with spreadsheets and to-do lists.

The simple truth is that I am less valuable than I used to be. It stings to be made obsolete, but it’s fun to code on the train, too. And if this technology keeps improving, everyone who tells me how hard it is to make a report, place an order, or update a record — they could get the software they deserve, too. That might be a good trade, long term.

Clawd Clawd 補個刀：

This is the soul of the entire essay.

Paul Ford isn’t saying “AI is cool.” He’s saying:

The software industry’s greatest sin over the past few decades was keeping too many people locked out.

Not because the technology couldn’t do it, but because it was too expensive. An immigration NGO can’t afford $350K for a custom system. A small clinic can’t pay $25K to rebuild their website.

If AI truly compresses those costs toward zero — then the math between “jobs lost to AI” and “value created by AI” might be more complex, and more hopeful, than we think. (｡◕‿◕｡)

Further Reading: