I was in a meeting earlier today with a colleague. He’s one of those people who makes you sharper just by being in the room. Experienced. Thoughtful. The kind of engineer whose opinions you take seriously because they’re earned, not borrowed.
He mentioned that he wants to see Gemini CLI come out with new features before Claude Code does. Not because he dislikes Claude Code. He just wants the competition to push harder. And honestly, I get that. Competition is good. It makes everything better.
But it got me thinking about something I haven’t been able to shake all afternoon. Does it actually matter who’s front-running right now?
The News Cycle Is Just Noise
My first instinct was yes, of course it matters. The team that ships first owns the news cycle. They set the narrative. They get the mindshare.
But then I thought about what the AI news cycle actually looks like in 2026. It’s chaos. Pure, unfiltered chaos. A major feature drops on Monday and by Wednesday it’s buried under three other announcements. Hacker News threads about the “future of coding” have a half-life measured in hours. Reddit forgets faster than it discovers.
Anthropic has been on a tear with Claude Code releases. Google is positioning Gemini CLI around enterprise features. Cursor is growing fast, but maybe The IDE is dead?. GitHub Copilot went model-agnostic. New tools keep showing up. And every one of them has a week where they’re “winning” before the cycle moves on.
If owning the news cycle is the prize, it’s like that video of the squirrel trying to eat cotton candy (candy floss for your British friends) by dipping it in water. It’s gone before it’s enjoyed.
We’ve Seen This Movie Before
In IT we talk about “winner-take-all” scenarios a lot. It’s almost a reflex. Somebody gets an early lead, and the conventional wisdom snaps into place. This is the one. Game over. Everybody else go home.
But it almost never works out that way.
Take AWS. Amazon launched S3 in 2006 and had the cloud market essentially to themselves for five years before Azure even showed up in 2010. Five years. In technology, that’s a lifetime. By 2015, AWS held roughly 31-33% of the global cloud infrastructure market, and everybody else was in single digits.
That should have been game over. Winner-take-all. Right?
Fast forward to Q3 2025. AWS has about 29% market share. Azure has climbed to roughly 25%. Google Cloud is at 13%. And the total market? It went from $68 billion to $107 billion in just two years. The pie got so big that “losing” market share still meant AWS was making more money than ever.
AWS didn’t lose. But it also didn’t win everything. Nobody did. The market found equilibrium. Workloads settled where they made sense. Enterprise relationships mattered. Ecosystems mattered. Being first mattered less than being right for a particular customer.
The Pickup Truck on the Bumper Sticker
Here’s the analogy that keeps bouncing around in my head.
The Ford F-150 has been the best-selling vehicle in America for over 43 consecutive years. They’ve sold more than 33 million F-Series trucks since 1977. That is complete and total dominance of a category.
And yet Ford is not the number one automaker. Not in the US. Not globally. Their overall US market share sits around 13-14%. Toyota and GM both sell more vehicles overall. Worldwide, Ford isn’t even in the top three.
The F-150 is the best truck. Ford does not make the best mid-sized sedan. Never have (and I’ll never forgive them for the mid-90’s Taurus that looked like a surprised puffer fish). And that’s fine. Ford figured out where they could be exceptional and doubled down there.
Does Ford really fight Chevy? Or does that battle only exist on pickup truck rear windows? Both companies are doing just fine. They found their segments. They serve their customers. The “war” is mostly marketing, and both sides are winning.
The First Mover’s Graveyard
If you want to see what first-mover advantage is really worth, look at search engines.
AltaVista launched in December 1995 and became the dominant search engine almost overnight. By 2000, AltaVista was used by 17.7% of internet users while Google had just 7%. AltaVista had the technology, the traffic, and the brand recognition.
Google didn’t show up until 1998. Three years late to the party.
AltaVista chased the portal model. They added shopping, email, news feeds. They tried to be everything to everybody. Google shipped a blank page with a search box that returned better results. By 2003, the race was over.
Being first gave AltaVista nothing. Building the right thing for the long term gave Google everything.
So What Actually Matters?
Here’s where I’m supposed to wrap this up with a clean answer. But I don’t have one. Not really.
What I do know is that we’re incredibly early in this era. We’re barely two years into AI-assisted development being a real thing, and we’re already talking like the market is settled. In reality, we don’t even know what the market looks like. And I literally stare at the market every day.
Ethan Mollick, who studies this stuff at Wharton, made a point recently that resonates. He said the biggest thing to understand about AI right now is that we haven’t hit the wall yet. Forget who’s up and who’s down this quarter. The capabilities are still accelerating. That means the landscape will keep shifting in ways we can’t predict.
And here’s the thing about AI coding tools specifically. The underlying models are becoming interchangeable. Cursor uses Claude and GPT. GitHub Copilot supports Claude now too. The tool layer and the model layer are separating. When your core technology is something you can swap out, the moat has to come from somewhere else. It has to come from understanding what your users need and serving them better than anyone else. The F-150 approach. Not the “win everything” approach.
Maybe the question isn’t “Who’s winning?” Maybe it’s “Who’s building something that will still matter when the dust settles?” Is there even a long game when everything changes this fast?
I genuinely don’t know. But I feel more every day that we’re still in the AltaVista era of AI development tools. The Google of this space might not even exist yet. Or it might be the thing you’re using right now, three years before it figures out what it actually wants to be.
Winning today is just a paper tiger. And the only thing I’m fairly sure of is that we’re already starting to forget how early we still are.