This is my magnum opus

hi! do you like supporting queer-owned/disabled-owned small businesses? do you like buying affordable stuff? awesome!

i have recently launched lavender lit, a used bookstore focused on uplifting the voices, stories, and experiences of marginalized folks of all kinds, with a specialty in queer books.

since so many of my formative experiences took place here on tumblr dot com, i decided to build in a special discount for tumblr users — you can enter the code tumblr10 at checkout for 10% off your first order on our online store! we’ve got a great selection of books, stickers, and zines to browse.

thank you so much for checking us out, spreading the word, and helping me get my dream business off the ground!

Translation of Internalspeak to Externalspeak

A P2 post I made internally a month ago on October 5th about what's next for Tumblr leaked as a screenshot. This is super rare, so I'll try to translate what was said for y'all here. It was also incomplete, so here's the full thing.

We are at the point where after 600+ person-years of effort put into Tumblr since the acquisition in 2019, we have not gotten the expected results from our effort, which was to have its revenue and usage above its previous peaks. It is better to have loved and lost than never to have loved at all. (Hat tip.) It is better to have tried to summit the peak and failed than never to have tried at all. We have learned a ton on the journey, and honed skills we can use to approach other summits.

This is me saying we've worked on Tumblr for four years with ~200 people full-time, and spent well north of $100M above revenue trying to turn the site around, but it hasn't yet. That sucks, but I also want to recognized the effort of everyone who tried and gave their best.

As we talked about in the past, if it doesn’t work we’ll have a backup plan and set up the business so we don’t need to let anyone go, we’ll just need to reflect and decide where else we should concentrate our energy together. This plan is happening now: the majority of the 139 people in Bumblr will switch to other divisions. No plans for any switches in Happiness or T&S.

Bumblr is the internal team name for the product side of Tumblr, and the link was to our internal company directory which showed the ~140 people currently in that team. There are also other teams that work on Tumblr that aren't affected or changing, so this tries to say specifically who's impacted. "Happiness" is our term for customer support, and "T&S" means trust and safety, which is the team that works on fighting bots, spam, dealing with illegal stuff, etc. While Tumblr was burning cash, we managed the rest of the products and company to support and subsidize it because we thought that would turn it around.

We assume the first choice for everyone working on Tumblr is to continue working on Tumblr, but we’re going to give everyone an opportunity to have a “top three” ranked list of what other things around Automattic they would be interested in working on. To infuse some Tumblr mojo.

As we've been telling the Tumblr team for over a year, if we can't get revenue up we need to switch some portion of them to work on other things within Automattic that do generate revenue and can support their salaries. We offered a survey where people could rank stack choose what they would be interested in working on instead of Tumblr. The team has actually been performing really well in their work, which is why we aren't letting people go, they just haven't been getting results, which usually means we're working on the wrong things and we should try working on different things. We've also learned a ton working on Tumblr that I think will make our other products better.

The leak was actually incomplete, here's the rest of the post:

This survey will be posted early next week, alongside some 2-minute videos from WP.com, WooCommerce, Jetpack, Day One, Pocket Casts, WP VIP, .Org, Applied AI, Texts, self-serve advertising (Blaze), Newspack, Pressable, and Gravatar talking about why you should consider picking them as one of your choices. In the meantime, feel free to bounce around the products linked on Automattic.com to get a feel for them as a consumer and see what tickles your fancy. We’ll crunch through those preferences with what the opportunities for growth are in the various products and businesses, come back with a plan, get feedback on that, and then post the final plan. The switchovers will happen on December 31, so we start 2024 completely fresh. We are shifting from the mode of “surging” on Tumblr with tons of people to get it to exciting growth, to working on how we can run Tumblr in the most smooth and efficient manner. Pretty amazing things in the social and messaging space have been accomplished with small teams, so I’m actually quite curious to see a smaller and more focused Tumblr’s performance in 2024.

What followed were a bunch of comments that actually caused us to update the survey, other teams around Automattic posted videos pitching Tumblr folks to work on them, we did the survey, and then posted a first draft of where people could go actually just this week. (Which is maybe why the leak happened, perhaps someone didn't like the proposed changes first draft.) As a reminder, there still are no people switching, it's just planning for what will happen on December 31st, 2023.

As I mentioned this was posted on October 5th, a few weeks before we announced the acquisition of Texts on the 24th, which actually didn't leak. I do appreciate whoever shared the screenshot trimmed it down to not prematurely break the Texts news.

So now you have the full post I made, it also got 45 comments which I'm not going to share out of respect to the privacy of colleagues, but you can imagine it generated a vigorous debate internally and a lot of discussion about how to make sure we're setting up Tumblr for success in this next chapter, a lot of tech discussion about maintenance, libraries, open source, and how small teams can move faster than big teams, if they have the right environment. I'll take further questions in the Asks.

Also, only tangentially related to Current Events… sometimes it feels like tumblr users are especially hostile to the service they’re voluntarily using- is this something that’s just expected with social media sites, or is there something about the tumblr user base that’s especially vitriolic?

Again, thanks for doing asks!

Tumblr has an especially young userbase, and it has definitely reminded me of how cruel kids can be to each other, given me flashbacks to elementary and middle school. I would truly encourage anyone who thinks attacking everyone working on the site or any attempt to make it sustainable to consider the long-term ramifications of that approach.

By the numbers, however, I believe the super-trolls are a very small sub-0.5% percentage of Tumblr's userbase, which is largely creative, supportive, artistic, and assumes good intent with the changes we make. However they can have an outsized impact, and this is one reason we are maintaining and even increasing our investment in Trust & Safety, which is the team that counters hate speech, bots, spam, trolling, attacks, and other anti-social behavior on Tumblr. Those bad actors will lose their accounts, and we will improve keeping them from registering new accounts right away and repeating the behavior.

The artistic, supportive part of Tumblr is one of the best places on the Internet. I've experienced moments of beauty and connection here that I haven't anywhere else, and it's why I love Tumblr deeply. It's why the entire team cares, and we'll do our best to support and allow that part of Tumblr to not just survive, but thrive.

Lord of the Rings Travel Posters - Created by Beverly Arce

Limited edition prints available for sale at Gallery Nucleus. You can follow the artist on Instagram.

28 days of exposure

Christmas with the Addams Family by Charles Addams

I discovered I can make chatgpt hallucinate tumblr memes:

This is hilarious and also I have just confirmed that GPT-4 does this too.

Bard even adds dates and user names and timelines, as well as typical usage suggestions. Its descriptions were boring and wordy so I will summarize with a timeline:

I think this one was my favorite:

Finding whatever you ask for, even if it doesn't exist, isn't ideal behavior for chatbots that people are using to retrieve and summarize information. It's like weaponized confirmation bias.

more at aiweirdness.com

Daily Mirror, England, January 22, 1909 Image © The British Library Board. All Rights Reserved.

Test

So talented