Damien Tanner sips coffee in a dim London flat, terminal flickering, as Opus 4.5 spits out code faster than any dev team.
That’s the scene. Seventeen years post-Pusher sponsorship on The Changelog—back when GitHub was a toddler—he’s declaring the era of the Small Giant. SaaS? Dying. Code reviews? Obsolete. Small teams—or solo acts—now forge empires. Bold claim from a guy who’s done it twice.
But here’s the thing: he’s not wrong. AI agents are the great equalizer. No more bloated orgs churning mediocrity. One dev, armed with Sonnet 4.5 (Opus who?), builds what took hundreds before.
SaaS is Dead? Pump the Brakes
Tanner doesn’t mince words in that epic 13-minute chapter. SaaS is toast because it’s slow, pricey, and can’t pivot. Everyone’s copying the same Notion clones, Stripe wrappers—pick your poison.
“SaaS is dead!?”
Wait, he backpedals later. Not dead-dead. Just the fat SaaS model. The one where VCs fund 50-engineer herds to ship CRUD apps. AI flips it: generate the boilerplate, tweak with humans. Boom. Giant output, tiny input.
Dry humor aside, it’s half-right. SaaS ain’t vanishing—look at the sponsors: Depot, Tiger Data, Notion, Fly.io. They’re thriving. But Tanner’s nailing the shift. Custom beats commodity. Why rent when you code your own in hours?
Punchy truth: incumbents are scared. Their moats? Mud puddles now.
Today’s dev days? Wildly different from Pusher era. Back then, real-time meant WebSockets hacks. Now? AI Engineer summits in SF, where coders mine automation gold. Tanner felt it: big halls, electric buzz, but missing that intimate “together” vibe. Like gold rush speculators, not diggers.
No Code Review? Chaos or Freedom?
Code review’s the ultimate bottleneck. Or non-existent for scrappy teams. Tanner says ditch it. AI’s got your back—Sonnet 4.5 changed everything. Generates solid code, catches bugs better than juniors.
Imagine: non-tech folks at the terminal. “Hey AI, build this.” No merge hell. JIT everything—just-in-time compilation for apps, deploys, lives.
But—hold up. Reliability? Latency kills API models sometimes. Local beats cloud for speed, but APIs win on power. Tanner’s Layercode picks Cloudflare Workers. TypeScript. Not Go—too verbose for this sprint.
It’s messy. Perfectly human messy.
Layercode time hits at 51:24. Tanner’s new baby: direct AI interrupts, no fluff. Set on Workers for global edge. Why? Low latency, scales free-ish. Directing the AI like a conductor—interrupt mid-thought, steer output.
API vs local: latency wars. Reliability flips daily. One outage, you’re toast. But small giants thrive on agility. Big SaaS? Locked in data centers, praying to uptime gods.
Why Can Small Teams Build Giants Now?
The era of the Small Giant. Chapter 20, pure fire. Tiny crews punch above weight. AI agents handle grunt work—code, tests, deploys. Humans: vision, polish.
Historical parallel no one mentions: 1990s Netscape. Marc Andreessen’s 10-man team crushed Microsoft… briefly. Browsers from garages. Now? AI’s the garage multiplier. Solo dev = 1995 team of 50.
My bold prediction: by 2028, 70% of new apps from <5 person teams. VCs pivot to acqui-hire these small giants. PR spin from Salesforce? “AI enhances productivity!” Nah. It obsoletes their model.
Tanner’s not hype-peddling. He’s building. Pusher built community—London events, global reach. Now Layercode invites everyone to code. Non-devs too. Terminal as playground.
SF summit? Energy shot, but venue flop. Windsurf chats—wait, Claude?—mind blown. Miners automating mining. Meta.
What’s next? Horizon hazy. More JIT. Agent swarms. But bottlenecks shift: data, not compute.
Skeptical? Fair. AI hallucinations persist. Sonnet 4.5’s leap feels real, though. Opus follows. Small giants rise because they fail fast, iterate dumb.
Big corps? Sunk cost fallacy. They’ll buy the winners—or die trying.
Look, Tanner’s journey—from charity sponsor in 2009 (GitHub year 1!) to AI frontline—grounds this. Changelog’s 18 years? Nov 9 birthday. World flipped.
Is SaaS Really Dying in 2025?
Short answer: nah, evolving. But Tanner’s right on symptoms. Bloat kills. Custom AI tools win. Notion’s great—for teams who pay. Solo? Code your own doc empire.
Sponsor shout: Fly.io powers Changelog. Edge deploys. Small giant enablers.
Tanner critiques gently: SaaS not dead, just irrelevant for innovators.
One insight: this mirrors open source 2000s. Linux kernel, small contribs built giants. AI democratizes that firepower.
🧬 Related Insights
- Read more: Alibaba’s $53 Billion AI Blitz: Rescuing Cloud Growth or Chasing Shadows?
- Read more: React Hooks Demystified: The Cursor That Makes Rules Inevitable
Frequently Asked Questions
What is the Small Giant era?
Tiny teams using AI agents to output massive software—think solo dev matching 100-engineer SaaS firms.
Is SaaS dying?
Not outright, but the VC-fueled bloat model is. Custom AI-built apps are killing copycats.
What is Layercode?
Damien Tanner’s tool on Cloudflare Workers: AI-directed coding with interrupts, JIT deploys, for small giant builders.
