After gushing with pride over Storyboard's many accolades, Mr. Karp pivoted, abruptly. The year-old concept "had run its course" and the editorial team, he announced, "will be closing up shop and moving on." Please, he asked, "join us in wishing them well."

But the Storyboard layoffs, which affected three staffers peripheral to internal operations, are hardly the only departures Tumblr has faced over the past six or seven months. Rather, they're the only ones Mr. Karp has spoken about publicly.

Sources close to the company, who requested anonymity, told Betabeat that a handful of high-level deputies have also quietly ended their tenure at Tumblr--leaving a noticeable absence around Mr. Karp where his leadership team should be.

"It’s like the fucking Argentinian government, people just get disappeared," said one source.

The most recent departure is Blake Matheny, Tumblr's VP of engineering, who gave notice in the past few weeks. Another source called it a "huge loss," adding, "Blake was the strongest tech leader there."

Fredrik Nylander, Tumblr's former executive vice president, gave notice last fall, we heard, but was asked to stick around by Mr. Karp, showing up at the office with less and less frequency. His LinkedIn profile states that he started as the CTO of Oscar, a New York City-based startup, this year. Mr. Nylander started as Tumblr's VP of technical operations back in 2011 and was promoted in less than a year to VP of engineering.

As EVP of Tumblr, Mr. Nylander replaced the role of former vice president Andrew McLaughlin, a veteran of Google, ICANN, and former deputy CTO for the Obama administration. Mr. McLaughlin lasted as VP at Tumblr for just nine months.

Marc LaFountain, Tumblr's former vice president of tech support, left in October, following his wife, an executive at R.J. Reynolds, to Switzerland. And last month, advertising veteran Rick Webb, who played a pivotal role consulting on the six-year old company's recent push toward profitability through advertising, announced that he too would be moving on after 10 months.

In the past, Tumblr has also lost lead developer Marco Arment, president John Maloney, and "vice president of people" Charlie Gray, an Xoogler who was with the company for only six months. So for tech leadership, that leaves Derek Gottfrid, VP of product, who may oversee some operations.

"We don't comment on staff moves, however all of the people you've referred to are very different circumstances spread out over months and not related at all to the closure of Storyboard," Tumblr spokesperson Katharine Barna said by email when she heard we were inquiring about the departures. She declined to comment further.

We reached out to each of the execs and will update the post when we hear back. But the sources we spoke with didn't attribute the recent shakeups to cost-cutting due to pressure from the board or even clearing the decks for Tumblr's rumored fundraising. Instead, they cited frustrations with Mr. Karp, who tends to marginalize deputies who disagree with him.

"Some people seem genuinely happy at Tumblr, but most are miserable, largely because Karp is wildly unpredictable and inconsistent," said one source. "Karp seems to treat Tumblr like a junior high lunch room--he sits with his five favorite people of the moment, and treats everyone else like a reject."

Mr. Karp has been encouraged to look for COO for Tumblr. Speculation says the role may go to former News Corp. chief digital officer and AOL CEO Jonathan Miller, who has been consulting for the company--one of the mentors Mr. Karp tends to collect. But Mr. Miller may still be hampered by "golden handcuffs" from his time at News Corp. Watchful observers don't have high hopes. "Even the COO search is a sham," said one source. "He’s not looking for Sheryl Sandberg, he’s looking to sideline that whole thing."

These staffing changes come at a critical time for Tumblr. Business Insider reported last week that Mr. Karp was in Silicon Valley "raising a big round of funding," even though it "doesn't need to raise more money." However, according to a profile in Forbes from January, Tumblr generated just $13 million in 2012, despite traffic of 18 billion page views per month. Forbes also noted that Tumblr shelled out an estimated $25 million in operations in 2012 and expects that figure to increase to $40 million in 2013. The company, which is headquarted in a cushy Flatiron clubhouse, last raised funding in 2011: $85 million at an $800 million valuation. Up until the end of 2012, Tumblr reportedly had a burn rate of $4 million to $5 million a month, before transitioning to its own data center, which lowered the burn rate to $2 million a month.

Given those financial concerns, some seem puzzled at the board's faith in Mr. Karp. "The fact that Tumblr is losing its most experienced people, at a moment when it wants people to believe it's succeeding in a big way, suggests some real incompetence on Karp's part," said a source.

The success Tumblr wants to project--as its Silicon Alley cohorts are trying to avoid down rounds--is related to the monetization strategy it launched for first time last year, which relies on native advertising via the Tumblr Dashboard, rather than traditional display or keyword ads. As head of sales Lee Brown told Bloomberg recently, brands pay for prominent placement of their posts, acting much like other users on the service:

“Marketers have become accustomed to buying scale as opposed to earning it,” Brown said. “We’re not really selling ads, we’re promoting their content.”

That approach has managed to attract big brands like Target, Coca-Cola, Adidas, Lions Gate, and Christian Dior, who average purchases of "just under six figures." Tumblr has begun offering the same service on its mobile app, which it says should lead the company toward since first annual profit since launching in 2007. But often the posts featured in the prominent Tumblr "Radar" spot on the dashboard come from regular users instead of brands, as evidenced by this unofficial archive.

One source was skeptical of the company’s emphasis on pageviews and number of blogs, given its advertising approach. "It has no strategy for monetizing anything other than logged-in dashboard users. Those numbers are much, much lower than what Tumblr's PR would suggest. So they're basically spending VC money to provide a free blogging platform, the vast majority of which can't be monetized."

Others disagreed with that assessment. "Monetizing the dashboard, actually, is the smart move. Why go and try and do deals with 100 million blogs, and share revenue, when you can just monetize the dash, which is bigger and you totally control?"

If Tumblr is indeed seeking another financing round, investors will no doubt be paying attention to its ability to monetize, as well as traffic itself. In November, Mr. Karp boasted that Tumblr had cracked Quantcast's list of top 10 U.S. websites. (It fact, Tumblr mistook the "top ten" badge on Quantcast as its website ranking. Instead, it had cracked no. 9 on Quantcast’s list of top networks and no. 15 on Quantcast’s list of top sites.)

Currently, Tumblr is ranked no. 21 in the top websites. A Quantcast graph shows that traffic appears to be plateauing, bolstered by an uptick in mobile users.

This story is developing, please email tips@betabeat.com with further information.

Mr. Pears wrote:

A lack of communication with the team running [the Tumblr API] means I’m going to have to take my app offline indefinitely because I won’t release it without the pieces I’m missing, specifically the pieces they were interested in seeing - that even with detailed information about my apps interaction I can’t even get a reply saying “yes, we’ve seen your posts, we’re swamped - but we’ll try and get someone to look at it soon” - and quite frankly, if I were a user of my app, I’d be angry at how long I’ve waited for the next release. I don’t want promises, I don’t want to be pushy or agressive. I just wish I didn’t feel so let down when the API has come so far so quickly.

Today, David Karp wrote on his own Tumblr, in a reblogged response to Pears:

You’re absolutely right — we’ve been letting down our developer community. This is an extremely high-priority for Derek [Gottfrid, Tumblr's Director of Product] and me, and something that our Engineering team (especially JB) has been pulling extra hours to help with. We don’t, today, have engineers dedicated to supporting our developer community — something we’re working to change as quickly as possible. You guys have been doing unbelievable work for Tumblr users, and we want to do everything possible to support your efforts.

Thank you so much for the thoughtful note. I’m sorry to let you down.

Karp then offers to Pears: "P.S. If empowering incredible developers is your kind of thing, we’d love to chat" replete with an email hyperlink and all.

A contrite, sincere apology! An admission of not doing something correctly! Refreshing, by any reasonable standard.

Of course, this isn't the first time Tumblr's had problems with their developer community—they threatened to shut down popular Tumblr extension Missing E, and pulled API access for another developer's Tumblr directory—but maybe, given the tenor of Mr. Karp's apology, the way they approach relations with their developer community may indeed be changing. Time will tell, actions speak louder, all that noise. But: it's a start.

Unless they cruelly drag along Tumblr devs like Lisa did Screech until they all finally just snap and grow out porny mustaches and make all these weird sex tapes and just become these ultimately sad caricatures of human beings who once brought joy to the world and held the potential to be something, nay, anything more than the fate they eventually resign themselves to. That'd just be mean.

But also, kinda funny.