The Signpost
← Back to Contents
View Latest Issue


Special report

Search and destroy: the Knowledge Engine and the undoing of Lila Tretikov

Contribute  —  
Share this
By William Beutler


The introduction to an unreleased WMF presentation about the Knowledge Engine

This article was originally posted on the author's blog and is republished with his permission. The views expressed in this article are his alone and do not reflect any official opinions of this publication.

The Wikimedia Foundation is in open revolt. While the day-to-day volunteer efforts of editing Wikipedia pages continue as ever, the non-profit Foundation, or WMF, is in the midst of a crisis it’s never seen before. In recent weeks, WMF staff departures have accelerated. Within just the past 48 hours, employees have begun speaking openly on the web about their lack of confidence in the leadership of executive director Lila Tretikov.

All in all, it’s been a terrible, horrible, no good, very bad start to 2016. Controversy in the first weeks of the year focused on the unexplained dismissal from the WMF Board of Trustees of James Heilman, a popular representative of Wikipedia’s volunteer base, before shifting to the unpopular appointment to the WMF Board of Arnnon Geshuri, whose involvement in an anti-competitive scheme as a Google executive led him to resign the position amidst outcry from the staff and community (see previous Signpost coverage).

But other issues remained unresolved: WMF employee dissatisfaction with Tretikov was becoming better known beyond the walls of its San Francisco headquarters, while questions mounted about the origin, status and intent of a little-known initiative officially called Discovery, but previously (and more notoriously) known as the “Knowledge Engine”. What was it all about? How do all these things tie together? What on Earth is going on here?

Deep breath.

The strange thing about the Knowledge Engine is that, until very recently, basically nobody knew anything about it—including the vast majority of WMF staff. Not until Heilman identified it as a central issue surrounding his departure from the Board had anyone outside the WMF staff ever heard of it—though in May 2015, Risker observed that a team called “Search and Discovery” was “extraordinarily well-staffed with a disproportionate number of engineers at the same time as other areas seem to be wanting for them”. This despite the fact that, as we know now, the WMF had sought funding from the Knight Foundation of many millions of dollars, receiving just $250,000 in a grant not disclosed until months later. As recently as this month, a well-considered but still in-the-dark Signpost article asked: “So, what’s a knowledge engine anyway?”

After several months of not knowing anything was amiss, followed by weeks of painful acrimony, we think we have the answer: as of February 2016, the mysterious project is in fact a WMF staff-run project to improve Wikipedia’s on-site search with some modest outside funding, which sounds like a good idea; sure, Wikipedia’s on-site search engine isn’t maybe the best, but we also know at some point it was an ambitious project to create a brand new search engine as an alternative to Google. Sometime in 2015, the WMF submitted a proposal to the Knight Foundation asking for a substantial amount of money to fund this project. It is described in still-emerging documents from this grant request as a “search engine”, and several early mock-ups seemed to suggest this was in fact the idea, as seen in internal documents revealed for the first time by the Signpost last week.

A slide from the April 2015 presentation for the Knight Foundation featuring "Wikipedia Search"

Why would Wikipedia consider building a search engine, anyway? The most likely answer is fear of being too dependent on Google, which sends Wikipedia at least a third of its total traffic. In recent years, Google has started providing answers to queries directly on the search engine results page (SERPs), often powered by Wikipedia, thereby short-circuiting visits to Wikipedia itself. Tretikov herself, in a rambly January 29 comment on her Meta-Wiki account page, identified “readership decline” as Wikipedia’s most recent challenge.[1]

It’s an understandable position: if you are the leader of an organization whose success has been largely described in terms of its overall traffic, any decline in traffic may be equated with a decline in Wikimedia’s ability to fulfill its mission. I submit this is short-sighted: that Wikipedia has an educational mission whose impact cannot be measured solely in terms of traffic. That Google borrows information from Wikipedia—though they are not alone in this—in such a way that it answers people’s questions before they have to actually click through to en.wikipedia.org is still a win for Wikipedia, even if it reduces the (already low) probability that a reader will become a Wikipedia contributor.[2]

The logic is twisted, but you can follow it: most readers find Wikipedia through a search engine, so if the search engine that helped make Wikipedia the success it is today changes its mind and starts pointing elsewhere, better to get ahead of things and create a new alternative that people will use. I guess? If we accept this reasoning, we still have to confront questions like: Is this actually something the WMF can accomplish? Is this within the WMF’s scope? Is this something that will help Wikipedia accomplish its mission? These are much harder questions for WMF to answer—in part because the answers are “no”, “no”, and “no”—and would absolutely have to be shared with the Wikimedia Board of Trustees ahead of time and, for political reasons, socialized within the Wikipedia community itself. The incident surrounding Heilman’s departure suggests the former was an issue, and the ongoing furor is because the latter obviously did not occur.

Meanwhile, the extreme unwillingness of Lila Tretikov and even Jimmy Wales to talk about it is, in fact, tearing the Wikimedia Foundation apart. Tretikov has lost all remaining credibility with Wikimedia staff and close community observers, not that she had much to begin with. As this week comes to an end, more staffers are quitting, remaining ones are complaining in public, and it seems impossible to imagine Lila Tretikov remaining in charge much longer.

Lila Tretikov and Jimmy Wales

If you need a detailed timeline of events to understand how we got here, I am pleased to say you’ll find just what you’re looking for below, although I’m afraid this whole thing is too large and multifaceted to do proper justice within the space of this already very long article. A full accounting may go back to the mid-2000s, when Jimmy Wales harbored ambitions of building his own search engine—Wikiasari in 2006 and Wikia Search in 2008. It certainly would include a full accounting of the many high-profile WMF staffers to leave since late 2014, and the role Tretikov played in each. It would include a careful examination of what the WMF can and should do in Wikipedia’s name, and an evaluation of how the evolving app-focused Internet raises questions about Wikipedia’s own future.

I think that’s more than I can accomplish here.

Instead I want to focus on what’s happening this week. But first we have to fill in some of the blanks. To do so, you’ll want to wind back the clock a few weeks:

  • Let’s start on January 25, when Jimmy Wales called Heilman’s claims that transparency issues were at the core of his dismissal “utter fucking bullshit”. Jimmy Wales is known for occasionally lashing out at pestering editors on his Talk page, and this certainly seems to be one of those times.
  • On January 29 Tretikov made her first public, community-facing statement about the Knight Foundation grant, which was welcomed for showing some self-reflection but also raised more questions than it answered.
  • On February 1 WMF developer Frances Hocutt stated, on Tretikov’s discussion page no less, that employees were being “censured for speaking in ways that I have found sharply critical but still fundamentally honest and civil”.
  • We then jump ahead to February 11, when Wales was still doing his “Baghdad Bob” routine, publicly insisting to Wikipedia editors that any suggestion WMF had ever considered building a search engine was “a total lie”.[3]
  • Just hours later, WMF Communications uploaded the Knight Foundation grant agreement itself to the WMF’s own wiki, confirming for the first time, in public, that WMF was describing the project as “the Internet’s first transparent search engine”. The Signpost has the most detailed breakdown not only of the grant agreement, but also three supplemental documents which were leaked to the Signpost but have not been made public at this time.
  • You might then have a look at The Register, always snarky, but with a decent summary of where things stood last week, just before it became newsworthy. I definitely recommend this February 15 story by Vice’s Motherboard about the fiasco (and this follow-up) but skip this Newsweek story except to see how the media was, for a brief moment, cluelessly reporting that Wikipedia was taking on Google.
  • However incomplete, I think this upshot from The Verge is a good enough summary, at least for public purposes:

Whether Wikimedia’s plans just naturally evolved [away from the search engine project] or whether it was responding to the community’s response is difficult to say, but the organization is now, at least, claiming it does not want to square up to Google, but just improve its own product.

  • As all this was unfolding, the exodus of key WMF staff was accelerating. On February 8, Tretikov announced on Wikimedia-l that Luis Villa, head of the Community Engagement department and previously a member of the WMF’s legal team, would be leaving.
  • At least Tretikov seemed to be in control of that one. Because the next day Anna Koval, a manager of the education program, announced her own departure on the mailing list.
  • And then on Friday, February 12, a very big resignation letter dropped on the Wikimedia-l mailing list: that of Siko Bouterse, another veteran leader who had long provided a crucial link between the Wikipedia volunteer community and the professional WMF staff. Careful with her words, Bouterse wrote:

Transparency, integrity, community and free knowledge remain deeply important to me, and I believe I will be better placed to represent those values in a volunteer capacity at this time.

  • Messing up my timeline a bit, but still worth noting: Hocutt, the developer who had made public internal fears about silencing dissent, announced her own (albeit temporary) departure in yet another Wikimedia-l post on February 17, noting her leave was “due in part to stress caused by the recent uncertainty and organizational departures.”

Finally, on February 16, Lila Tretikov published an open letter (co-authored by Vice President of Product Wes Moran) on the Wikimedia blog titled “Clarity on the future of Wikimedia search”. Alas, it wasn’t terribly clarifying: it seemed aimed at the clueless mainstream journalists like the one at Newsweek, and not at the Wikipedia community who knew which information gaps actually needed to be filled in. It began:

Over the past few weeks, the Wikimedia community has engaged in a discussion of the Wikimedia Foundation’s plans for search and discovery on the Wikimedia projects.

Well, that is certainly one way to put it! Put another way, you have been backed into a corner defending the untenable proposition that Wikipedia has never considered building a search engine, and now that the mainstream press is reporting, based on your own documents, that you are building a search engine, one certainly has to say something about it.

After much boilerplate about the growth of Wikipedia and its many achievements, Tretikov and Moran finally get around to the point:

What are we not doing? We’re not building a global crawler search engine. We’re not building another, separate Wikimedia project. … Despite headlines, we are not trying to compete with other platforms, including Google.

This seems to be true, insofar as there is no search project currently. However, Wales had previously locked himself into the position that there was never a search project originating from WMF, and by now we know that is obviously false. Without any acknowledgement in this letter, it is useless. But it’s worse than that:

Community feedback was planned as part of the Knowledge Engine grant, and is essential to identifying the opportunities for improvement in our existing search capacity.

We are 10 months past the initial plans for this far-reaching, mission statement-busting project, six months past the award of a grant to pursue this quixotic effort, and not two months removed from the violent ejection of a Board trustee over the matter… and all you can say is “feedback was planned”?

Finally, the closest thing to acknowledging the Knowledge Engine was, at some point, actually a search engine:

It is true that our path to this point has not always been smooth, especially through the ideation phase.

And nothing more.

The first comment on the post was brutal, bordering on uncivil, from a retired editor. It concluded:

You are either:

a. Flat out lying, and hoping we don’t actually read the grant,
b. Have misled the Knight Foundation as to your intentions for their grant money, or
c. Seriously incompetent and should never be put in charge of writing a grant application
None of these options look good for the WMF.

A few hours later, a member of WMF’s Discovery team gamely stepped forward and tried to offer a plausible explanation for how the grant request did not necessarily imply a Google-competitive search engine project—damage control, essentially—but still had to concede the wording of the grant did not make Tretikov or WMF look good: “It is ambiguous. I can’t speak to the intent of the authors and while there are current WMF staff listed, they are not the sole authors of the document.”

Finally, a day later a true hero emerged in Max Semenik, another Discovery team engineer, mostly unknown to the community, and who was willing to take off his PR hat to say what everyone pretty much knew:

Yes, there were plans of making an internet search engine. I don’t understand why we’re still trying to avoid giving a direct answer about it. …

The whole project didn’t live long and was ditched soon after the Search team was created, after FY15/16 budget was finalized, and it did not have the money allocated for such work … However, ideas and wording from that search engine plan made their way to numerous discovery team documents and were never fully expelled. …

In the hindsight, I think our continued use of Knowledge Engine name is misleading and should have ended when internet search engine plans were ditched. No, we’re really not working on internet search engine.

Now that sounds like a real answer! What’s more, it also provides the outlines of a believable story as to why the Knight Foundation grant included language about the search engine, even if it wasn’t then the plan. This is transparency of a sort! But it’s transparency of the last-ditch kind. That it had to come from a low-level engineer indicates there is a major problem, and speaks to the fact that the WMF simply cannot go on this way.

At a time when Wikipedia has already-existing problems, the WMF was asking for money to basically create a whole new set of problems. That is the mark of an organization, if not a movement, adrift. Clearly, they pitched a search engine to Knight, and they asked for millions—I have heard the number placed at $100 million over 5 years—later reduced to $12 million, of which Knight provided $250K to build a plan—essentially a pat on the head: ‘since we like you, here’s a few bucks to come up with a better idea’.

Mysteries remain: where did the idea come from, who championed it, when did it die—or when did it recede and what happened afterward? One answer is supplied in another comment on this public thread (!) from yet another WMF team member (!) pointing a finger at former VP of Engineering Damon Sicore as having “secretly shopped around grandiose ideas about a free knowledge search engine, which eventually evolved into the reorg creating the Discovery team.” Sicore left in July 2015. A big remaining question, for which there is no answer at this time: when the actual grant was submitted to the Knight Foundation.

An argument I have heard in recent days is that it’s common in grant-making to try for everything you can and see what actually sticks. This may be true, but if so, it doesn’t seem to have been worth it. That WMF leadership felt they had to hide the fact later on also underlines the mistake they knew they were making.

Another big question: how does this affect Wikipedia’s public reputation, particularly among donors, most especially among foundations? You have to think the answer is a lot. The WMF looks like the Keystone Kops. Why would you give it money? And right now, the Knight Foundation specifically must be asking what it’s got itself into.

A slide from an already-public November 2015 presentation on Discovery by the Wikimedia Foundation

Within the last 24 hours, the trickle of public criticism about Tretikov has become a widening stream. Some of it is taking place in the above comment thread, plenty is still happening at Wikimedia-l, but a lot of it has moved to a semi-private Facebook group called Wikipedia Weekly, where staffers previously not known for voicing internal dissent have been speaking quite frankly about how bad things are at 149 New Montgomery Street.

Yesterday afternoon on the mailing list, a developer named Ori Livneh replied to a plea for calm by community Board trustee Dariusz Jemielniak by explaining why they could not remain silent:

My peers in the Technology department work incredibly hard to provide value for readers and editors, and we have very good results to show for it. Less than two years ago it took an average of six seconds to save an edit to an article; it is about one second now. (MediaWiki deployments are currently halted over a 200-300ms regression!). Page load times improved by 30-40% in the past year, which earned us plaudits in the press and in professional circles.

…

This is happening in spite of — not thanks to — dysfunction at the top. If you don’t believe me, all you have to do is wait: an exodus of people from Engineering won’t be long now. Our initial astonishment at the Board’s unwillingness to acknowledge and address this dysfunction is wearing off. The slips and failures are not generalized and diffuse. They are local and specific, and their location has been indicated to you repeatedly.

Shortly thereafter Asaf Bartov—one of WMF’s more outspoken staffers, even prior to the last 48 hours—voiced his agreement and turned his comments back to Jemielniak:

Thank you, Ori. +1 to everything you said. We have been laboring under significant dysfunction for more than a year now, and are now in crisis. We are losing precious colleagues, time, money, *even more* community trust than we had previously squandered, and health (literally; the board HR committee has been sent some details). Please act. If for some reason the board cannot act, please state that reason. Signal to us, community and staff, by concrete words if not by deeds, that you understand the magnitude of the problem.

And then, about 10 minutes later, Lila Tretikov posted to this very conversation thread, and this is all she had to say:

For a few 2015 accomplishments by the product/technical teams you can see them listed here:

https://meta.wikimedia.org/wiki/2015_Wikimedia_Foundation_Product_and_Technology_Highlights

That is the complete text of her emailed post. That is really all she had to say, in a public thread specifically criticizing her leadership and all but explicitly calling for her removal. One gets the feeling, at this point, even Lila Tretikov just wants it to be over.

In the early morning hours of February 19, a WMF software engineer named Kunal Mehta wrote an impassioned, rather forlorn post on his personal blog, titled: “Why am I still here?”:

Honestly, I don’t understand why the current leadership hasn’t left yet. Why would you want to work at a place where 93% of your employees don’t believe you’re doing a good job, and others have called you a liar (with proof to back it up) to your face, in front of the entire staff? I don’t know everything that’s going on right now, but we’re sick right now and desperately need to move on. …

I love, and will always love Wikimedia, but I can’t say the same about the current state of the Wikimedia Foundation. I’ve been around for nearly nine years now (nearly half my life), and it feels like that world is slowly crumbling away and I’m powerless to stop it.

And that’s why there is really just no way Lila Tretikov can continue to lead the WMF. A week ago, the thinking was: the Board of Trustees chose her over James Heilman, so they’re really sticking with her. At the time it also seemed like the Knowledge Engine was a going concern, and their support for her owed to their insistence on moving ahead with the project above community and staff objections. Knowing what we do now, it’s inexplicable. The thinking now is: she obviously has to go, and the only reason the Board might have for not acting on it would be legal considerations.

For the sake of Wikipedia’s future, the Wikimedia Foundation needs new leadership. Lila Tretikov must resign, or she must be replaced. This is the most challenging article I’ve ever had to write. The next one, I hope, will be about the start of the turnaround.

Notes
  1. ^ “Our aim was to begin exploring new initiatives that could help address the challenges that Wikipedia is facing, especially as other sources and methods arise for people to acquire knowledge. If you haven’t yet, please have a look at the recent data and metrics which illustrate the downward trajectory our movement faces with readership decline (since 2013), editor decline (since 2007, which we stabilized for English Wikipedia in 2015), and our long standing struggle with conversion from reading to editing. These risks rank very high on my list of priorities, because they threaten the very core of our mission.”
  2. ^ See this comment from WMF’s Dario Taraborelli, who argues: “[T]raffic per se is not the goal, the question should be about how to drive back human attention to the source”.
  3. ^ Full quote: “To make this very clear: no one in top positions has proposed or is proposing that WMF should get into the general “searching” or to try to “be google”. It’s an interesting hypothetical which has not been part of any serious strategy proposal, nor even discussed at the board level, nor proposed to the board by staff, nor a part of any grant, etc. It’s a total lie.”
  4. ^ “It seems to me extremely damaging that Lila has approached an external organisation for funding a new search engine (however you want to define it), without first having a strategic plan in place. Either the Board knew about this and didn’t see a problem, or they were incorrectly informed about the grant’s purpose. Either is very bad.”
← Previous "Special report"
Next "Special report" →
S
In this issue
17 February 2016 (all comments)
  • Special report
  • Op-ed
  • Featured content
  • Traffic report
  • Technology report
  • Blog
    • + Add a comment

    Discuss this story

    These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.
    • Motion of no confidence, anyone? TomStar81 (Talk) 22:33, 19 February 2016 (UTC)[reply]
      • Sadly, it did not help. Apparently, the board is still under the delusion that it is possible to salvage the situation whilst maintaining the status-quo, and even a consensus of nearly all the old hands on staff were unable to make them budge beyond assigning a "coach" that the ED feels is optional. — Coren (talk) 22:51, 19 February 2016 (UTC)[reply]
        • That's the "old hands" responsible for the mess Lila's been trying to clean up, right? That's the old hands who have been told to respect and respond to the community and readership, right; the crew who have had their pet folly projects shut down as silly, nonsensical wastes of time, right? I hope the board gives her enough time to finish the job. But then, the board is spineless. --Anthonyhcole (talk · contribs · email) 04:38, 20 February 2016 (UTC)[reply]
          • You'll have to be more specific, Anthony. Which "pet folly projects" have been shut down since the new ED's start date in this apparent grand spring cleaning process nobody has ever heard of? Flow is in maintenance mode (which is not a euphemism as its commit history attests) while a user talk opt-in beta on a few wikis is ongoing; it might indeed be best to ultimately shut it down if there's a better, less contentious way to address user experience issues with talk pages. I personally think that's hard to do without painting yourself into a corner long term, but that's another story. Beyond that?--Eloquence* 15:06, 20 February 2016 (UTC)[reply]
            • Eloquence, I had flow in mind in particular. But I recall some other thing (an automatic home page or something for mobile, that Jorm initiated) that was canned. Your surfacing here and on the Wikipedia Weekly Facebook page is creepy and unbecoming. You are largely responsible for the shambles Lila inherited. We haven't forgotten the contempt with which you treated us when you were in control at WMF. What are you trying to achieve here? Have you been stirring the shit in the background, drumming up this coup? --Anthonyhcole (talk · contribs · email) 00:23, 21 February 2016 (UTC)[reply]
              • Hi Anthony, the situation is very much out of control, and very respected and trusted individuals who are absolutely beyond your vitriol have made this abundantly clear. My former colleagues are distressed, taking medical leave, can't sleep, and are openly revolting. Perhaps you allow me the amount of humanity to believe that my reaction to this is shock and pain more than anything else.
              • As for your argument, it is simply false. The tech priorities have not shifted significantly, to the extent that they can be identified at this point. mw:Wikimedia_Engineering/2014-15_Goals/Q1 documents the priorities as they were just around the time Lila joined the organization; she had little influence on these goals as she was still brand new. Again, with the exception of Flow being backburnered, these projects have either been successfully concluded to generally broad satisfaction (SUL finalization, HHVM, Phabricator), or they are continuing (VisualEditor, Mobile Apps, Analytics).
              • This is not a criticism of her. It's a criticism of your narrative. We largely agreed on these priorities, and on many other things, including the need to focus more attention on community needs. She, Toby, Damon and I worked closely together on the engineering reorg, as well, which also had been in the making for a long time. And I do indeed even agree with the part of your narrative that states that Lila has generally herself made community relations more of an ED level issue than it was before, which is doubtlessly a good thing. The issues here, however, go deeper than that, and they reflect primarily the churn, communications issues and erratic decision-making of the last year, of which KE is only a symptom. That is the issue many people are raising, one which you are unfairly and unempathetically dismissing with a heroes/villains storyline. There will be no triumphant return of User:Eloquence to the WMF, don't worry. Been there, done that, got the closet full of t-shirts, and it was my call to leave the org. But this is a crisis, people I care about are affected, and I will speak out as and where I see fit.--Eloquence* 01:16, 21 February 2016 (UTC)[reply]
    Eloquence, I'm starting to wonder, what is your narrative? Wouldn't it be better to try to calm things down? Reading things like "unfairly and unempathetically dismissing with a heroes/villains storyline" isn't quite what I would've expected. Jeblad (talk) 04:00, 21 February 2016 (UTC)[reply]
    Erik is right. I'm being hyperbolical and manichean, largely due to the late hour and the loneness of my voice. As someone on wikimedia-l just put it, I'm looking at a black box (WMF) and deducing causes from correlations, which is a dangerous practice.
    Just thinking out loud here. It seems the board - or at least some members of the board - have been getting warnings from staff about the staff's relationship with the ED for over a year. Clearly it was an important issue. Yet the board did nothing. Might this have been mended, before it came to this, if the board had behaved in a timely, responsible fashion? That's rhetorical, to no one in particular. --Anthonyhcole (talk · contribs · email) 05:44, 21 February 2016 (UTC)[reply]
    Different people on the board had different opinions regarding how serious the issue was and how to fix it. Doc James (talk · contribs · email) 15:43, 21 February 2016 (UTC)[reply]
    James, you're awesome. Regarding the different views on the board about the seriousness of the problems: the breakdown in the staff-ED relationship is so clear now, it must have been clear six months ago, to any board member who did due diligence on the issue. I'm left with the only possible conclusion that those who didn't see it didn't look. What a failure. This is a tragic situation for all concerned - not least Lila - and it could have been handled gently and discretely if confronted early enough.
    In the "Bring back Sue Gardner" thread on Andrew's Wikipedia Weekly Facebook group, in the responses to Jonathan Cardy's comment [1] they're discussing the usefulness of rigorous post-mortems after crises like these. I hope the WMF commissions such an investigation, and I hope whoever performs it doesn't shy away from examining the board's role in this. --Anthonyhcole (talk · contribs · email) 22:53, 21 February 2016 (UTC)[reply]
    Might be even better to get an independent body to do this review. Personally I saw the issues between the staff and ED as obvious as of mid Oct. The staff survey in Nov/Dec clarified things further. What has happened since November has not come as a surprise to me. The longer this is taken to be addressed the fewer okay options will be available to the board. Not sure what the positions of my fellow board members are. Doc James (talk · contribs · email) 00:14, 22 February 2016 (UTC)[reply]
    independent body to do this review, would be a very good idea--Ozzie10aaaa (talk) 01:30, 24 February 2016 (UTC)[reply]
    This "motion of no confidence" stuff reminds me a bit of that Star Wars movie. We still don't know nearly enough to be confident who the villains and heroes (if any) may be.
    I may be crazy, but my guess is that there's a simple structural fix for this: WMF needs to reduce salaries at the top end. Lay out all the employees in terms of pay percentile, everyone below 75% the top 25% stays the same, everyone above that gets cut down to the 75th percentile what the lower 75% make. If some major personalities want to leave, great - if not, we can reevaluate and try a bigger cut. The problem with laying out big bucks for decision makers is they make decisions, and we don't really want people making decisions, we just want them closing out tasks in Phabricator that have been sitting around unanswered for years on end. The problem with laying out big bucks for career software designers is they want to make careers, have some big piece of software with a fancy name they can say they built for a top website, not say they closed out a bunch of miscellaneous user requests.
    What to do with the leftover elite decision maker money? Pay it to senior undergrads and selected volunteer editors. Give them each an internship, a simple project, a job reference. I'm thinking victory for Wikimedia isn't really coding a better MediaWiki - victory is teaching hundreds of people how to code a better MediaWiki.
    And as for decisions? Well, let's try and elevate community processes like the Wishlist into something that most users actually know about. Wnt (talk) 00:42, 20 February 2016 (UTC)[reply]
    Wnt, I have no idea whatsoever why you think this has anything to do with how much money anyone makes. Really and truly, I can't see any relationship at all between what you wrote and what this Signpost/blog article talks about. Risker (talk) 00:49, 20 February 2016 (UTC)[reply]
    @Risker: I haven't seen convincing evidence there's anything "the matter" with Lila Tretikov. I'm thinking what's "the matter" is that we keep hearing about an Executive Director at all, rather than an RFC or a volunteer starting a coding project. There has to be a way to downgrade the role of the people at the top. Wnt (talk) 01:04, 20 February 2016 (UTC)[reply]
    Ah. Sorry, I seem to have been reading a very different discussion than you have been. Risker (talk) 01:21, 20 February 2016 (UTC)[reply]
    @Wnt: : [ignoring for the moment that this seems unrelated to the ED issues] you want to cut all the experianced people, give their jobs to interns - and somehow this will result in having 100's of people who can code better mediawiki? Who will teach these interns how to do stuff? If gsoc/opw is any indication, most interns (not all) need significant hand holding, especially at the beginning. Replacing all experianced people with unsupervised interns would probably just result in a bunch of interns making unusable software. More generally I think you are assuming that everyone is a developer at wikimedia, and some developers eventually get promoted into decesion makers and that it would be possible for everyone to revert back to fixing bugs if needed. Where really there is a much more of a separation between managers (particularly at the upper level) and developers and they dont have the same skill set. Managers may or may not have the technical skills to close out technical phabricator tickets. Bawolff (talk) 02:17, 20 February 2016 (UTC)[reply]
    @Bawolff and Risker: Whoops! I realize now what I said was being read as a more radical plan than I'd initially intended, which was only to go after the top quarter, not the top three-quarters. My apologies. Wnt (talk) 03:39, 20 February 2016 (UTC)[reply]
    • Wales is quite often a liar, pure and simple. But every time someone raises this, even with examples, they find themselves very quickly sort-of ostracised because for some insane reason he has a lot of acolytes. I've not been around that much of late but the last example I can recall was, I think, Ched. Tretikov may be a part of the problem, and staff disagreements etc are not uncommon anywhere, especially when a new broom appears, but Wales is the real public face, ie: mainstream media etc. He needs to back off because he is only making matters worse with his wayward announcements and opinions, often supposedly in a personal capacity but inevitably read by those with news-power as being in some way official. - Sitush (talk) 01:39, 20 February 2016 (UTC)[reply]

    • I approve of the direction Lila's taking the WMF in.

      The community consultation was particularly effective in highlighting the needs and wants of our readers, as well as those of the editing community, and this has informed the ongoing strategy design process - a process that has deep community input. That strategy, in turn, informs funding decisions.

      The Community Resources Team surveyed the community and discussed with them their technical priorities, and tailored their current Idea Lab Campaign accordingly.

      The WMF have accepted the FDC's proposal that the WMF submit to the same reporting standard they expect of their chapters.

      Lila could and should have been more candid about the Knowledge Engine project as the idea was evolving, and I hope she's learned from that, but under her the WMF has developed institutional structures that are intrinsically respectful of and responsive to the volunteers and readers. I hope she survives this crisis but, if she doesn't, I hope those structures do. --Anthonyhcole (talk · contribs · email) 02:27, 20 February 2016 (UTC)[reply]

    • Amazing piece User:WWB. We had a consultation of our readers in 2015. Search and discovery was on the list at 13th. Efforts to improve our apps and content were the first five. Not saying internal search could not use some improvement but that was not our long term goal. Doc James (talk · contribs · email) 06:44, 20 February 2016 (UTC)[reply]
    • Not sure why you say Max Semenik is unknown to the community. He is an administrator here as User:MaxSem, and previously served as a steward and an administrator of the Russian Wikipedia.--Ymblanter (talk) 09:44, 20 February 2016 (UTC)[reply]
    • I worked at a nonprofit for several years, and received several grants from the Knight Foundation. I feel the specifics of this case have been very, very, very blown out of proportion. While the Wikipedia community may have important underlying concerns, there is nothing shady or unusual about this particular grant process. At all.

      Here’s some very important specifics pertaining to the grant process:

      The widely circulated grant agreement PDF was not written by the WMF. It was written by junior staff members at the Knight Foundation. Here’s how it works: you have a conversation with staff at the Knight Foundation about ideas you have for program areas you’d like to fund, throw these ideas back and forth through a phone call or two, and then send the Knight Foundation a 1 page summary of what you’re looking to do. Knight junior staff then turns this into a document that they send to their VP, and then once the VP signs it, it’s done.

      Importantly, the language on the grant agreement is not written by the grantee, but instead by the Knight Foundation, and usually by junior staff. I could show you some of our old grant agreements, and you’d be blown away by how “off” the language is on the agreement versus on the proposals we sent in. The grant agreement language is designed to be informal, and is written largely based on conversations.

      Given the funding amount of $250k, this was *not* a long, drawn out grant process. This grant must have gone from “quick first chat” to “grant agreement” in a week or less. Grants of less than 250k are not approved by the Knight Foundation Board and are instead approved by VPs. They happen very quickly. This is likely why many people at the WMF felt blindsided.

      The last Knight Foundation grant I got took two weeks from conversation to grant agreement. The final text in the grant agreement was written by staff at the Knight Foundation and had several important mistakes present in it.

      I’m not sure why senior WMF didn’t explain this more clearly. My best guess is they didn’t want to malign the Knight Foundation. Nothing about this grant process seems incompetent to me.

      I could not disagree more with your call for Lila Tretikov’s resignation. It’s completely ridiculous. This is just a growing pain associated with WMF applying for foundation funding, something they’ve only done a literal *handful* of times in the past. Grant funding is an extremely important part of WMF being able to innovate outside of existing budget areas. Grants are also typically opaque. I don’t think the WMF has applied for many grants, at least based on its size. I’m willing to bet that no other nonprofit the size of WMF has taken as little grant money as WMF has. So this seems like both a growing pain and a necessary growing pain. 2600:1010:B001:974D:A1FF:B07E:3146:626B (talk) 12:20, 20 February 2016 (UTC) (not logged in so as to not burn my own grant relationship bridges)[reply]

    Wikimedia has secured many larger & more complex grants, as early as 2008 when we were able to secure a $1M/year / three year grant from Sloan, which provided crucial support in the early years of WMF's growth. Since then WMF developed its major gifts capacity, and that team is generally excellent at what it does. See foundation:Benefactors which gives you a good view into the sheer scale of the program over the years. Even Knight has previously given a $600K grant for mobile to Wikimedia. The problems here are of a different nature.--Eloquence* 13:12, 20 February 2016 (UTC)[reply]
    The grant agreement’s language is amazingly close to the grant application’s language. The grant application was written and submitted in Aug 2015 by the WMF to the Knight Foundation.
    That grant application has been asked for multiple times by multiple people. No one has formally released it. Doing so would not however confirm your suspicions or reflect badly on the Knight Foundation.
    The grant application started as early as Apr. It involved a fair bit of communication and a number of meetings / presentations. It was awarded by Knight Sept 1st, and not approved by the board until Nov. So it was an 8 month process at least. Doc James (talk · contribs · email) 13:27, 20 February 2016 (UTC)[reply]
    I strongly object to the notion that I have been unwilling to discuss the Knight grant, "knowledge engine" or anything else. There is no evidence offered for this because it simply isn't true.--Jimbo Wales (talk) 13:40, 20 February 2016 (UTC)[reply]
    Jimmy, please weigh in on this thread publicly: https://lists.wikimedia.org/pipermail/wikimedia-l/2016-February/081997.html Based on public comments, private comments, and private discussions, the ongoing crisis of faith in the ED and board is a key factor in the increasing staff departures, as we continue to see Lila fail to improve in transparency, culture, accountability, strategic clarity, etc since the November meeting and the institution of a performance improvement plan. The KE 'scandal' was an opportunity for Lila to take charge, communicate clearly, and articulate how she and WMF are taking concerns about openness and planning seriously; instead at every step she, and you, have denied there was ever a problem. Staff have been watching and listening to upper management and the board, and people are voting with their feet. --brion (talk) 15:46, 20 February 2016 (UTC)[reply]
    Jimbo Wales, just before the Knight Foundation grant agreement was released, you said, "To make this very clear: no one in top positions has proposed or is proposing that WMF should get into the general 'searching' or to try to 'be google'. It’s an interesting hypothetical which has not been part of any serious strategy proposal, nor even discussed at the board level, nor proposed to the board by staff, nor a part of any grant, etc. It’s a total lie."
    The wording of the grant agreement did not bear you out. Since then, Max Semenik has said,
    To clarify:
    • Yes, there were plans of making an internet search engine. I don’t understand why we’re still trying to avoid giving a direct answer about it.
    • There has never been any actual technical work on this project.
    • The whole project didn’t live long and was ditched soon after the Search team was created, after FY15/16 budget was finalized, and it did not have the money allocated for such work (umm, was it in April? in such case, this should have been soon after the leaked document was created).
    • I don’t think anybody but the certain champion of the project has considered competing with Google with any degree of seriousness.
    • The scrapping was finalized in summer, after said champion and WMF parted ways.
    • However, ideas and wording from that search engine plan made their way to numerous discovery team documents and were never fully expelled.
    • Speaking of team name, “Discovery” is not about stage one from that leaked plan. The team was initially called “Search” then almost immediately after realizing it also works on non-search projects (like maps) it was renamed to Search and Discovery then just Discovery. At the time of the second renaming, we already had no plans of actually doing any internet search engine work.
    • In the hindsight, I think our continued use of Knowledge Engine name is misleading and should have ended when internet search engine plans were ditched.
    • No, we’re really not working on internet search engine.
    • And will not work in the future.
    • For shizzle.
    Lila has started an FAQ on Meta, in which she comments on the above statement by MaxSem, saying, "My recollection of events is close to Max’s."
    Brion Vibber says at the FAQ,
    Former VP of Engineering Damon Sicore, who as far as I know conceived the 'knowledge engine', shopped the idea around in secret (to the point of GPG-encrypting emails about it) with the idea that Google/etc form an 'existential threat' to Wikipedia in the long term by co-opting our traffic, potentially reducing the inflow of new contributors via the 'reader -> editor' pipeline. More ambitiously there was some talk about trying to capture more total web search mindshare/user-share... obviously since Google/etc have butt-tons of money they can much more effectively grab the user share, making our potential project unpopular until it gets canned... I guess? Given the secrecy at that stage, I assumed Damon was just a bit ... 'colorfully' paranoid about things like Google hiring people away or organizing their offerings to more thoroughly hide us... obviously if we'd gone through with a giant search engine it would have been public knowledge before we *did* it, so it never made much sense to me to hide it other than in coordinating an initial organizational/PR 'blitz'. It kind of feels like Lila stayed in 'KE is secret project' mode while everyone else moved away from it, but again I've not been in the loop for this stuff... --brion (talk) 17:50, 18 February 2016 (UTC)[reply]
    These statements are entirely incompatible with what you said: "To make this very clear: no one in top positions has proposed or is proposing that WMF should get into the general 'searching' or to try to 'be google'. It’s an interesting hypothetical which has not been part of any serious strategy proposal, nor even discussed at the board level, nor proposed to the board by staff, nor a part of any grant, etc. It’s a total lie."
    To believe you, and to believe that you only ever meant to tell the truth, the whole truth and nothing but the truth about this, I would have to believe that MaxSem, Lila, Doc James and Brion are all liars, and you are the only one telling the truth all along. --Andreas JN466 16:46, 20 February 2016 (UTC)[reply]
    At least "it simply isn't true" is an improvement over statements like "it f-ing simply isn't true". The statements may not be entirely incompatible. No one currently in top positions has proposed or is proposing that; Sicore's proposal (if this allegation is accurate) wasn't a "serious" strategy proposal, and it wasn't discussed at the board level, nor proposed to the board by staff, and the extent that remnants of such unserious, deprecated legacy language remained in actual grant documents was by an inadvertent lack of attention to detail, not by intentional fabrication. Wbm1058 (talk) 18:07, 20 February 2016 (UTC)[reply]
    Hold on, Wbm1058. Are you suggesting that someone reporting to the ED was able to apply for a multi-million dollar grant without the direct authorization of the ED? That the plan put forward in the grant application was not approved by the ED in advance of its submission? That a grant agreement that included clear language indicating that a broad search engine was to be developed, which was not completed until months after Sicore left the Wikimedia Foundation, was not negotiated under the purview of the ED? The ED is still an employee of the Wikimedia Foundation, unless something's happened that I missed entirely. I get that the plan changed, but the application for the grant didn't. Risker (talk) 19:16, 20 February 2016 (UTC)[reply]
    Risker, this is surely complicated. From what I've gleaned (I have no first-hand knowledge): Sicore didn't apply for the grant, b/c he was gone before the application was submitted. No, surely the ED negotiated and approved the plan, and is still an employee. It was alleged that only Sicore proposed that WMF should get into general 'searching' or to try to 'be google'. Language in the grant application that mislead in this direction may have not been intended to mislead quite that way. Clearly, with her talk of "TARDIS" the ED viewed this as a "really big thing", but not so clear that her vision was the same as Sicore's. Another vision might be something like Wolfram Alpha, which I hadn't heard of before today. A desire for secrecy might have been the rationale for not saying that more directly. Compare Wolfram search with Wikipedia search. I can't say one is better; they're just different. At this point I think, the party line is that they just don't know what their vision of improved search or "discovery" is, as they are just researching several avenues to look for promising roads to travel down. I'm mystified as to why apparently 90% of the staff, most of whom don't work in the Discovery department, are so stressed out about this. Seems there must be more rationale for their dissatisfaction than just this "Discovery" thing. Wbm1058 (talk) 20:14, 20 February 2016 (UTC)[reply]
    Wbm1058, you seem to be suggesting that we have to parse every statement made by Jimmy Wales to check whether it could potentially mean the precise opposite of what it appears to say. You may be happy to do that, but do you think that a communication style which requires such parsing is open, clear and transparent, and conducive to building trust?
    And are you happy for Wales to tell you it's a "total lie" to suggest that anyone "in top positions has proposed or is proposing that WMF should get into the general 'searching' or to try to 'be google'", or that it is "a part of any grant", when we now have WMF staff and the ED confirming that Damon Sicore did propose just that, and that this proposal did inform the grant application? Is it okay with you for Wales to call something a "total lie" which is actually true? And to call other community members like Doc James liars in the process? Andreas JN466 21:23, 20 February 2016 (UTC)[reply]
    No, I'm not happy about the need to parse Jimmy's comments in search of increasingly convoluted ways to find that they are consistent with what others are saying. I too find his communication style lacking, and am not pleased with some things he has said about Heilman. On the other hand, we should be careful about saying the same kinds of things about him. Wbm1058 (talk) 23:52, 20 February 2016 (UTC)[reply]
    Jimmy, could you discuss the reasons why the Board feels Lila continues to be a good choice for the ED position (I assume this continues to be the case, as the Board has made no statements to the contrary), when every single other "C-level" employee stated to the Board that they had no confidence in her, and many other WMF employees have cited her incompetence as a factor that is damaging staff morale? What are your feelings on the large number of WMF employees leaving the organization, with many citing as reasons for their departure organizational dysfunction and a feeling that they will be retaliated against for speaking out? Does the Board have any plans to address these issues? Could you tell us who recommended Geshuri to the Board? For that matter, could you do the same for the other appointed Trustees? Why do Board members, including you, apparently not read the Board Noticeboard on Meta? Do you think it is problematic that some members of the Board, such as Guy Kawasaki, rarely-to-never edit the projects and appear to have no familiarity with them? I look forward to a discussion of these questions. --71.119.131.184 (talk) 19:07, 20 February 2016 (UTC)[reply]

    A story in desperate search for a conspiracy. Sorry, but this is completely out of proportions. Jeblad (talk) 22:18, 20 February 2016 (UTC)[reply]

    Quite the opposite, Watson. Deduction. At last, it all comes together. Knowledge Engine was/is a secret and badly conceived program, James wanted to discuss it, meanwhile ED did not function (staff discontent) and when James interfered too much with that topic, he was kicked out. Left to explain: the appointment of Geshuri (though could be related to same top staff disfunctioning issue), and those extreme Jimmy Wales posts. -DePiep (talk) 01:09, 21 February 2016 (UTC)[reply]

    Nov. 20 Grant for payment within 60 days and the announcement

    The timeline some are discussing seems a little off. The Knight Grant was not completed until it was signed on November 20, 2015 (November was also when the Wikimedia FAQ page was created [2]). They then had 60 days to payout, and during that 60 days Knight made the joint announcement, which says it is:

    . . .to explore ways to make the search and discovery of high-quality, trustworthy information on Wikipedia more accessible and open with $250,000 from the John S. and James L. Knight Foundation. Funding will support an investigation of search and browsing on Wikipedia and other Wikimedia projects, with the goal of improving how people explore and acquire information Wikipedia includes more than 35 million articles in hundreds of languages. Its standards for neutral, fact-based and relevant information have made it a reliable resource for nearly half a billion people every month. With more than 7,000 articles created every day and 250 edits made per minute, Wikipedia is constantly growing and improving. Its open, nonprofit model, allows anyone to participate and contribute. The new Knight-funded project will help make it easier to discover information on this vast resource of community-created content.[3]

    So, regardless of the boilerplate in the agreement - Knight and WMF clearly have a common understanding that it is for early research into the Wikipedia.org search function of Wikipedia (aka., Wikimedia) projects, not any sort of Google. -- Alanscottwalker (talk) 22:24, 20 February 2016 (UTC)[reply]

    • As the article pointed out, Wittylama wrote a powerful piece on this matter. I agree particularly with him that the Knowledge Engine is actually a really good idea. It is really disappointing that it is brought down not by popular disagreement with the concept, but by this furore caused by opacity and disengagement within the WMF and between WMF and the volunteer community. Deryck C. 00:09, 21 February 2016 (UTC)[reply]
    It's the opacity claim that is rather opaque - if you accept a grant for study and open a Wiki page on the project - it's hardly opaque. I guess, some people want to know the date someone was not thinking about building a Google (if anyone can even remember when that was) but that seems untenable, all the ideas people decide not to pursue should be discussed? They would have no time to eat or sleep, let alone do any work. Alanscottwalker (talk) 00:19, 21 February 2016 (UTC)[reply]
    Is it still not opaque if the grant was discussed and worked on in secret for months before being publicly mentioned? (Recall that the first official disclosure of the grant came not from the WMF, but from the Knight Foundation listing it on their website.) I think you're setting up a straw man here. No one, as far as I can tell, wants disclosure of every single thing anyone at the WMF has ever mentioned or thought of, down to the level of one-off water cooler talk. It seems clear that this Knowledge Engine idea went far beyond that stage, with the WMF devoting time to putting together this grandiose proposal and initially asking the Knight Foundation for many millions of USD. As Beutler states in this article, Knight appears to have "let them down easy" by talking the WMF down to a small exploratory grant, which likely indicates the Knight Foundation shared some of the concerns raised by others about a lack of a detailed plan and the WMF's ability to follow through on such a proposal. And let's not overlook the broader context, with many community members having felt for years that the WMF is unresponsive to them, and WMF employees openly discussing their concerns about a lack of communication and leadership. The way this whole affair was handled appears to reinforce those views. --71.119.131.184 (talk) 01:38, 21 February 2016 (UTC)[reply]
    Well much of that all appears to be false - apparently, it was not very much pursued as a Google thing - that idea just never got off the launching pad. You discuss things with a funder precisely to spitball ideas - if your speculation is even true that it was someone at Knight who caused them to modify the idea, that shows it was not much of a deliverable idea or commitment by the WMF, as an organization, to even go there, in the first place. Alanscottwalker (talk) 01:49, 21 February 2016 (UTC)[reply]
    Alanscottwalker, that seems right to me, from my limited knowledge. Deryck, agreed that the underlying idea is really good, but opacity and lack of engagement caused many problems, including this furore. – SJ + 08:04, 22 February 2016 (UTC)[reply]
    Well, if you are Knight investing 250,000, you don't want to crush thought by being overly restrictive in a research grant, and if you're the WMF, you want a restricted grant to be as unrestricted as possible. As for engagement, there has got to be a balance, because you just cannot put every thought up for discussion to thousands of people, who may not know what they are talking about legally or technically, who have multiple diverse agendas (including, job security, or making money off of Wikipedia, or seeing Wikipedia fail), and maybe entirely adverse to every idea (which anyone who has seen Wikipedia discussions knows is de rigueur). Every idea has to go through tons of refinement before it's presented, that's just good process. Alanscottwalker (talk) 12:59, 22 February 2016 (UTC)[reply]

    "I strongly object to the notion that I have been unwilling to discuss the Knight grant, "knowledge engine" or anything else. There is no evidence offered for this because it simply isn't true.--Jimbo Wales (talk) 13:40, 20 February 2016 (UTC)"[reply]

    You have been willing to discuss these, you just haven't been willing or able to tell the truth about them (and evidence for that is provided in the piece above). In typical WMF fashion, your damage control has backfired completely. And for those still keeping track of the many WMF failures, the new Gather extension will be completely disabled here on 1 March 2016. Fram (talk) 08:16, 23 February 2016 (UTC)[reply]
    I recall it is established that James Heilmann was withhold background papers for the two topics Jimmy Wales mentions. They were not released (to certain Boardmembers) voluntarily. -DePiep (talk) 08:23, 23 February 2016 (UTC)[reply]
    • The thought that came to my mind is that Wikimedia should try to "focus on core competencies and outsource the rest". It's not Wikimedia's job to save us from Google. Building an internet search engine seems like an unnecessary and non-productive diversion of funds. Praemonitus (talk) 18:50, 23 February 2016 (UTC)[reply]
    It's your Signpost. You can help us.

















    Wikipedia:Wikipedia Signpost/2016-02-17/Special_report