It's Useless for Office Agents to Tout Their DAU

Office agents touting their DAU right now is meaningless. Take Doubao Work's 800w DAU — it's essentially just Feishu and the Doubao app funneling users in endlessly. Open Feishu and there's a big banner right there telling me to assign tasks to Doubao Work. The only thing that made me happy was that Doubao Work's logo changed from an ugly Möbius strip to Dou Jie, who loves to work. As for WorkBuddy's 1000w DAU, all I can say is that if you blanket the offline world like that, even the shittiest app should be able to hit ten million DAU. If you really love DAU so much, why not add a QQ streak-revival feature to WorkBuddy? DAU would straight up double. Who knows how many of those two companies' DAU numbers are users who just log in to claim points, plus the default-checked "launch at startup" box. If you guys like looking at data, I can show you a few more: Baidu Dazi's MAU is approaching 10 million, Kuku AI has surpassed 50 million users, and SenseTime's Little Raccoon has over 9 million MAU. You can absolutely imagine how many users traditional apps like Baidu Wenku and the Baidu app have stuffed into Baidu Dazi and Kuku AI. To reiterate our earlier point: if Baidu can still buy its way onto a leaderboard this early, it means competition hasn't reached a boiling point yet. Once Robin realizes that buying traffic is pure comedy, or once Baidu Dazi's DAU asymptotically approaches ERNIE Bot's DAU from this past spring and buying traffic becomes completely useless — that's when it means all-out war has arrived. Although flaunting DAU is worthless, flaunting revenue is valuable — for example,

Office agents boasting about DAU right now is meaningless. Take Doubao Work's 800w DAU — it's essentially just Feishu and the Doubao app funneling users in endlessly. Open Feishu and there's a big banner right there telling you to assign tasks to Doubao Work.

The only thing that made me happy was that Doubao Work's logo changed from an ugly Möbius strip to Doubao, the work-loving mascot.

As for WorkBuddy's 10 million DAU, all I can say is that if you carpet-bomb offline like that, even the shittiest app should hit ten million DAU.

If you really love DAU that much, why not add a QQ streak-revival feature to WorkBuddy? Your DAU would literally double.

Who knows how many of those two companies' DAU numbers are users who just log in to collect points, plus the default-checked auto-launch-on-boot.

If you guys like looking at data, I can show you a few more: Baidu Dazi's MAU is approaching 10 million, Kuku AI has surpassed 50 million users, and SenseTime's Little Raccoon has over 9 million MAU.

You can absolutely imagine how many users traditional apps like Baidu Wenku and the Baidu app have crammed into Baidu Dazi and Kuku AI.

To reiterate our earlier point: in the early days of a leaderboard, the fact that Baidu can still buy its way up means competition hasn't reached a boiling point yet.

Once Robin realizes that buying traffic is pure comedy, or once Baidu Dazi's DAU asymptotically approaches ERNIE Bot's DAU from this spring and paid acquisition becomes completely useless — that's when the real war begins.

Even though advertising DAU is worthless, advertising revenue is valuable. Doubao's enterprise ARR surpassing 2 billion, sure — but Tencent calculating its ARR as surpassing $500 million based on seven days of revenue... I really have to hand it to them.

Here's the exact quote from the earnings call: "WorkBuddy's revenue over the most recent seven days, annualized, exceeds $500 million — more than double that of Tencent Meeting." This methodology reminds me of a certain two model companies.

All I can say is, if you calculate it that way, LibTV and Tipsy's ARR should exceed $1 billion.

Let me convert it for everyone using the above methodology: WorkBuddy's best seven-day revenue is 65 million RMB, its best single-day revenue is under 10 million RMB, which works out to less than 1 RMB contributed per person to WorkBuddy — proving that building a consumer office agent is pure comedy.

So let me follow up on my take on WorkBuddy bragging about consumer DAU: it has value, but not much. Its main value is trash-talking on earnings calls and intimidating Doubao Work into flashing its own DAU.

One sincere suggestion: office agents should hurry up and go B2B!

On a side note, Tencent's Q3 capex was 81.4 billion RMB, up 5x year-over-year. Building large models really burns money. Here's hoping Honor of Kings and Delta Force can keep covering the Hunyuan team's expenses 😭