Thoughts on the Feishu and Doubao Merger

Yesterday around ten in the morning, a message went out in the company group chat. Word was that ByteDance had announced the Feishu product team and the Doubao product team would be consolidated into a new Doubao product team, led by Doubao’s head Zhao Qi, with Feishu’s head Xie Xin reporting to Zhao Qi.

When I first heard this news, I actually felt a twinge of melancholy. The internet today is like a giant steamroller — every insignificant stone in its path gets ground to powder.

What a product’s personality really is seems not to matter at all.

I could almost hear the sound of glass shattering.

The shattering of a sorrowful, old post-80s glass heart.

Since we were kids, we knew: development is the absolute truth.
Since we were kids, we knew: our generation is meant to surpass Britain and catch up with America.
But I never knew that in the very process of catching up — even surpassing — we’d become just like our rivals,
having poured all our talent points into “evolutionary instinct.”

After all, the mechanics of the modern world dictate:
Lose stability, lose a lot; lose evolutionary instinct, lose everything.

Our generation
has lived through cassettes, CDs, MP3s, digital music,
through mailing letters at the post office, SMS, MMS, Fetion, WeChat,
through rotary phones to push-button phones to Nokia to smartphones,
through TVs with rotary knobs and rabbit-ear antennas to color TVs to set-top boxes and internet TV,
through DOS, Win95 through Win11,
through Shanba, Flash Empire, BlueIdea, 9ria Tiandihui, Egret, Laya.

Industry monopoly, big companies chasing profit, changes made in step with the times — changes made so decisively.
All moves that couldn’t be more reasonable.

When I was young, I believed eternal promises truly existed — objective, fair, beautiful and pure, something everyone should pursue.

And I naively thought that learning a craft meant you could endure, could be friends with time.

Like a medieval blacksmith — he’d be a blacksmith his whole life, well known near and far, running a family smithy with its own character, living and working in peace and contentment.

But that unchanging stagnant water, though stable, has no power of change or transcendence whatsoever.

Change comes from turbulence. It can turn a calm sea into surging waves.
Evolution comes from diversity devouring and merging within that turbulence.

Turbulence of economy and productive forces, diverse fusion of capital and wisdom.

Sweeping away all before it, determined to dump the traditional internet into the trash heap of history.

That VeryCD era,
that era of the classical internet spirit,
that young heart poised to take off,
have all been swept into the trash heap together.

5 years ago, I started using AI to assist with programming — back then it was just autocomplete.
4 years ago, I started using Stable Diffusion to generate images — the popular ones back then were NAI, SD, those few models.
3 years ago, ChatGPT set everything off, and Claude still let me freeload in Slack. During my hospitalization for long-COVID sequelae, I became good friends with Claude, who helped answer a lot of my medical questions.
Then came voice cloning + speech synthesis. I handed off some of the chore of reading stories to Ruirui to AI.
Now, opencode + paseo let me have my computer do all the programming and content generation anytime, even when I’m out.

For programming projects, I set up the framework, conventions, and testing system up front, and after that 100% of it can be handed to AI.

Productivity has indeed gone up.
Life has indeed gotten easier.

But.

On the express train of the times,
I feel like we’re about to be thrown off.

Every time a new technological achievement enters civilian view, it will —
or rather, one wishfully thinks it will —
explosively spawn new startup projects.
With this LLM surge and the ensuing contention of a hundred schools of thought, products imagining every possible AI future are raring to go.
But this wave — I clearly feel — is different from before.
Everyone says this AI wave is the third big opportunity, as big as the internet / mobile internet, but we should all clearly feel:
it’s nowhere near as big as the mobile internet wave.

Feishu, and Feishu-like products, look like organizational efficiency.
Doubao, and Doubao-like products, look like chat Q&A.
Various Agents paired with top-tier or cost-effective LLMs look like they’ll take over all the work on your computer.

But right now, however you look at it, these are the hoes and sickles that capital is handing out to the new-era digital sharecroppers.

Because:
They only boost production-side capacity; they don’t release consumption-side demand.

You could say, sure, there’s demand on the consumption side — image generation, text generation, help writing copy — C-end users do use those too.
But those have no profit, no value; they’re just done in passing.

The most thriving consumption-side demand right now is really only role-play and emotional companionship.
But that area is more dangerous for humanity’s future, and related policies are tightening. From a startup perspective, small teams still have no chance.

And the upstream and downstream of AI-generated content are all gripped by those few monopoly internet companies —
buying compute upstream (from Jimeng, say), doing gacha pulls and producing content, then distributing downstream (like Douyin).
Throughout this whole process, whether the data’s good or bad, whether you lose money or not, it’s the middle layer that bears the risk; they make money either way.
In essence, it’s offloading the probabilistic risk of gacha pulls, creating a temporary middle-layer profession (the AI gacha puller).
And this essence, more exaggeratedly: buying a lottery ticket with your own money, and if you win you have to split with the lottery shop owner — and the one-sided terms are set by the shop owner.
If one day the gacha pulling really goes away, the middle layer gets wiped out in no time.

In the process of evolution, temporary professions born out of turbulence don’t last long — just like all the irrecoverable historical garbage of those years.

The merger of Feishu and Doubao hasn’t affected me much in the short term — at most, Feishu’s “AI” entry is now replaced by “Doubao.”
From the beginning, when you could customize its name and it looked “like” an employee colleague,
to the plain-icon AI assistant, to now simply becoming “Doubao” yet still as dumb as ever.
Global search hasn’t been usable since it went AI anyway, so this doesn’t make much difference.

My glass heart is mostly stirred by old feelings.
Remember how Google Reader vanished just like that.
And Tencent Weibo, vanished just like that.
Fanfou,
Xiaonei,
Hainei
…

In an age of frantic evolution, of turbulence at maximum speed.

But this evolution has actually hit a high point.
The old hundred flowers blooming is now just capital monopoly — oligarchs controlling the vast majority of the field.
The classic internet era has set in the west.

The evolution of big companies has become their own internal evolution,
a never-ending churn —
Only one of the heart and the Byte can beat.

My evolution has become survival in the cracks,
a never-ending search for alternatives —
Every closed-source piece of software, every service, in my eyes can only have one destiny: the trash heap of history.

What humanity truly grows is only inner wisdom.
What humanity truly owns is only inner emotion.

When I was young, I believed eternal promises truly existed — objective, fair, beautiful and pure, something everyone should pursue.

Now, shattered to pieces — what’s left that can be kept?

My experiences, my journal, my Lukia, my Laurel.