Werd I/O

[Laura Hazard Owen at NiemanLab]

There are some interesting referral statistics embedded in this piece. Facebook referral traffic has fallen more than 40% over the last year; referrals from Reddit have increased by 88%.

But the focus is this:

"Search traffic, still dominated by Google search, has remained relatively steady during the period, Brad Streicher, sales director at Chartbeat, said in a panel at the Online News Association’s annual conference in Atlanta last week. Google Discover — the Google product offering personalized content recommendations via Google’s mobile apps — is increasingly becoming a top referrer, up 13% across Chartbeat clients since January 2023."

I think what's particularly notable here is the shift between kind of product. Google Search, despite the black box nature of its ever-changing algorithm, always felt like it was a part of the open web.

Discover, on the other hand, is an algorithmic recommendation product that tries to proactively give users more of what they want to read. It's much more akin to a Facebook newsfeed than it is a search index. There are likely editors behind the scenes, and a human touch to what gets surfaced. Publishers are even more in the dark about how to show up there than they were about how to rise through search engine rankings.

I'm curious about what this means for the web. Is this just an advertising / walled garden play from a company that wants to maximize advertising revenue and time on platform? Or is it a reflection of the web getting too big and too messy for many users, creating the need for a firmer hand to show them where the good content is? Is it a function of increased skittishness about an open web that might publish content and ideas that aren't brand safe? Or is it just changing user behavior in light of other apps?

Perhaps some elements of all of the above?

#Media

[Link]

[Paul Bradley Carr]

"I’ve worked at (and founded!) my fair share of billionaire-funded publications and I’ve always had a firm rule: You have to be more critical of the people writing the checks (and their cronies) than you are of anyone else. It’s the only way to offset the inherent bias of taking their money."

Paul Carr discusses quitting his column at the SF Standard because of its newfound apparent shyness when it comes to criticizing tech moguls - which is a serious journalistic flaw when you consider how important said moguls are to the culture and politics of San Francisco.

This is in the wake of fallout from its coverage of Ben Horowitz's conversion to MAGA, to which the subjects publicly objected. The SF Standard's backer, Michael Moritz, is another wealthy tech backer, who has actually been collaborating with Horowitz's partner Marc Andreessen to build a sort of city of the future on repurposed agricultural land in the North Bay.

As Paul points out, there must be a separation of church and state between editorial and business operations in a newsroom in order to maintain journalistic integrity. That doesn't seem to be something every newcomer understands.

#Media

[Link]

[Dan Kennedy at Media Nation]

Dan Kennedy picks up on a detail in Brian Stelter's Atlantic article about troubles at the Washington Post:

"The Post’s content-management system, Arc, which was supposed to be a money-maker, had instead turned out to be a drag on the bottom line."

He goes on to sing Arc's praises, but notes that 25% of its staff were just laid off, and wonders what went wrong there.

Here's what I think happened. There were two parallel forces at play:

• Newsrooms are not natural software companies (except for their own ends).
• Content management systems are a commodity technology.

It's notable that almost every newsroom that has built its own CMS has eventually left it in favor of a platform built by someone else - most commonly WordPress. Sinking resources into building your own means spending money to solve problems that someone else has already solved, and often solved well.

Particularly in tough times for the industry, newsrooms need to be spending money on the things that differentiate them, not by reinventing perfectly good wheels. WordPress isn't zero cost - most newsrooms partner with an agency and a managed hosting provider like WordPress VIP - but it's a lot cheaper than building all those features yourself would be. And the outcome by picking an open source platform is likely higher quality.

The exception is if the way you both think about and present content is radically different to anyone else. If you're truly a beautiful and unique snowflake, then, yes, building your own CMS is a good idea. But there isn't a single newsroom out there that is unique.

Likewise, if I'm a potential customer (and, as it turns out, I am!), I don't know why I'd pick a proprietary platform that's subject to the changing business strategies of its troubled owner over an open source platform which gives me direct ownership over the code and powers a significant percentage of the web. The upside would have to be stratospherically good. Based on sales emails I get that choose to focus on Arc's AI readiness, that case isn't being made.

The outcome is a bit sad. We need newsrooms; we need journalism; we need an informed voting population. Honestly, the Arc bet was worth trying: I can see how a platform play would have been a decent investment. But that doesn't seem to be how it's panned out, to the detriment of its parent.

#Media

[Link]

[Casey Newton at Platformer]

This fantastic round-up post focuses on Platformer's decision in January to leave Substack in protest of its content policies that permitted full-throated Nazis to earn money on the platform.

With a long-term view, it's been a good strategic move:

"We’re much less vulnerable to platform shifts than we were before. I had long worried that Substack’s unprofitable business would eventually lead it to make decisions that were not in the best interest of our readers or our business. (Besides not removing literal 1930s Nazi content, I mean.)"

This is the reason publishers should publish from a website they control. Sure, you can syndicate out to meet readers where they're at, but owning your own space makes you much less subject to the whims of someone else's platform.

And even that syndication to social platforms is becoming more controllable. One hope for the future that Casey notes:

"One way I hope [Platformer] will evolve is to become part of the fediverse: the network of federated sites and apps that are built with interoperability in mind. The fediverse is built on top of protocols, not platforms, which offers us a chance to decentralize power on the internet and built a more stable foundation for media and social apps."

Ghost, the open source platform that now powers Platformer, is building fediverse support directly into its platform at a rapid pace, so this almost feels like an inevitability. The benefit will be that Platformer can reach its readers on platforms like Threads, Flipboard, and Mastodon and maintain full control over its relationships with them. That's a game-changer for publishers.

#Media

[Link]

[Matthew Hutson at Nature]

"Beyond the ability to fine-tune open models for focused applications, Kal’tsit says, another advantage of local models is privacy. Sending personally identifiable data to a commercial service could run foul of data-protection regulations. “If an audit were to happen and you show them you’re using ChatGPT, the situation could become pretty nasty,” she says."

Many organizations have similar privacy needs to these researchers, who simply can't send confidential patient data to third party services run by vendors like OpenAI. Running models locally - either directly on researcher laptops, or on researcher-controlled infrastructure - is inevitably going to be a big part of how AI is used in any sensitive context.

We have the same needs at ProPublica - unless journalists are examining public data, they shouldn't use hosted services like ChatGPT that might leak identifying information about sources, for example. Local models are a huge part of the future for us, too.

#AI

[Link]

[Kyle Orland at ArsTechnica]

"Microsoft and Constellation Energy have announced a deal that would re-open Pennsylvania's shuttered Three Mile Island nuclear plant. The agreement would let Microsoft purchase the entirety of the plant's roughly 835 megawatts of energy generation—enough to power approximately 800,000 homes—for a span of 20 years starting in 2028, pending regulatory approval."

This seems to be the new front in datacenter technology: purchasing or building entire nuclear plants in order to cover the energy cost. It is significantly better than high-emissions power from sources like coal, but it also speaks to the increased demand that new technologies like AI represent.

As ArsTechnica points out:

"Industry-wide, data centers demanded upward of 350 TWh of power in 2024, according to a Bloomberg analysis, up substantially from about 100 TWh in 2012. An IEA report expects those data center power needs to continue to rise in the near future, hitting the 620 to 1,050 TWh range by 2026."

AI is a huge and growing part of that, although let's not pretend that the internet industry overall has low emissions. We often pretend we're greener than we are, simply because we can't directly see the output - but there's a lot of work to do, and a lot of carbon footprint to own up to.

#AI

[Link]

[Tracy Durnell]

"Links — connections between ideas — are the magic system of the Internet. They power the open web, enriching online writing. Generative AI is the parasitic dark magic counterpart to the link."

I love Tracy's observation that "online, we think together", which also calls back to the original definition of the word blog ("weblog" = "we blog").

Links are context, further thought, community. Removing that context removes depth. They're inherent to the web: they're what the web is. When platforms want to strip-mine value from our work - our writing, our thinking - by lifting it away from its community and context, we need to fight back. And fight back we will.

#Technology

[Link]

[Chris Metinko at Crunchbase News]

"Last year, venture funding to Black-founded U.S. startups cratered — totaling only $699 million and marking the first time since 2016 that the figure failed to even reach $1 billion, Crunchbase data shows."

And:

"While last year did not see Black founders raise $1 billion in total, this year such founders and startups are on pace to raise less than even half-a-billion dollars. In fact, the combined total of funding to Black founders in the second half of last year and the first half this year is only $351 million."

While some of this is a reflection of the ongoing tightening in VC overall, that certainly doesn't account for a pull-back of this magnitude.

VC is often a connections-based business: investors like to have warm introductions from people they trust. It helps to be part of the in-group, and given the demographics and backgrounds of most investors, Black founders may be excluded. Open calls for pitches help, but the single biggest thing venture teams could to do widen their net and make sure they don't miss out on talented Black founders is for their own teams to be more representative. This article doesn't directly mention whether there's been progress on that front - but the numbers suggest maybe not.

#Business

[Link]

[Eva at kibty.town]

PSA for anyone who switched to Arc as their main browser (hey, that's me!): it had a giant vulnerability that the team, at the time of writing, doesn't seem to have acknowledged publicly, although it has been patched.

Aside from the lack of disclosure, perhaps the biggest ongoing concern for me is in the last few paragraphs:

"while researching, i saw some data being sent over to the server [...] this is against arc's privacy policy which clearly states arc does not know which sites you visit."

Sigh.

#Technology

[Link]

[Federal Trade Commission]

"A new Federal Trade Commission staff report that examines the data collection and use practices of major social media and video streaming services shows they engaged in vast surveillance of consumers in order to monetize their personal information while failing to adequately protect users online, especially children and teens."

None of this is particularly surprising, but it's frankly nice to see the FTC see it and recommend taking action. Lina Khan is doing great work actually holding software monopolies to task.

My favorite recommendation is the first one:

"Congress should pass comprehensive federal privacy legislation to limit surveillance, address baseline protections, and grant consumers data rights;"

This should have happened years ago, and even now, getting it done will be a struggle.

This one, on the other hand, falls into the "and pigs should fly" category:

"Companies should not collect sensitive information through privacy-invasive ad tracking technologies;"

Yes, companies should not, but they will until comprehensive privacy legislation is enacted with meaningful penalties. This report is a step in the right direction; that legislation must come next.

#Technology

[Link]

[Jason Koebler at 404 Media]

"The creator of an open source project that scraped the internet to determine the ever-changing popularity of different words in human language usage says that they are sunsetting the project because generative AI spam has poisoned the internet to a level where the project no longer has any utility."

Robyn Speer, who created the project, went so far as to say that she doesn't think "anyone has reliable information about post-2021 language used by humans." That's a big statement about the state of the web. While spam was always present, it was easier to identify and silo; AI has rendered spam unfilterable.

She no longer wants to be part of the industry at all:

"“I don't want to work on anything that could be confused with generative AI, or that could benefit generative AI,” she wrote. “OpenAI and Google can collect their own damn data. I hope they have to pay a very high price for it, and I hope they're constantly cursing the mess that they made themselves.”"

It's a relatable sentiment.

#AI

[Link]

[Myoung-Gi Chon in The Conversation]

"We found a disturbing link between work-related communication outside of regular hours and increased employee burnout. Answering emails after hours was linked to worse productivity, employees badmouthing their employers and other negative behaviors."

This is an important (if perhaps obvious) finding, but it's worth diving a little deeper and asking follow-on questions. Is it just the act of sending communications out of working hours? Or is it also an underlying organizational culture of disrespect for employees that allows such a thing to be normal?

The reason I ask is that one might be tempted to address the symptom - those out of hours emails - when there's likely something deeper to also take care of.

In the same vein, that's not to say that you shouldn't address the expectation of ubiquitous availability because the larger cultural work is still to be done. They clearly are bad in themselves, and do lead to exhaustion and burnout. But it seems to me that you have to do the bigger work, too.

#Business

[Link]

[Matt Mullenweg]

"Those of us who are makers, who create the source, need to be wary of those who would take our creations and squeeze out the juice. They’re grifters who will hop onto the next fad, but we’re trying to build something big here, something long term—something that lasts for generations."

Matt Mullenweg takes a strong stand for open source, and against companies that claim to be open but aren't quite.

Of course, not everything Automattic does is open source - its commercial operations were kicked off by the centralized Akismet anti-spam service, after all - but I agree that this clarity is useful.

It ends with a call to action: to support organizations that support ecosystems rather than abuse them. It's hard to disagree with that.

#Technology

[Link]

[Kavitha Surana at ProPublica]

"When the mother of three realized she had unintentionally gotten pregnant in the fall of 2022, Georgia’s new abortion ban gave her no choice. Although it made exceptions for acute, life-threatening emergencies, it didn’t account for chronic conditions, even those known to present lethal risks later in pregnancy."

This story - alongside Amber Nicole Thurman's - shows that the abortion bans really are leading to preventable deaths.

"Miller ordered abortion pills online, but she did not expel all the fetal tissue and would need a dilation and curettage procedure to clear it from her uterus and stave off sepsis, a grave and painful infection. In many states, this care, known as a D&C, is routine for both abortions and miscarriages. In Georgia, performing it had recently been made a felony, with few exceptions."

As Kavitha Surana points out, abortion bans haven't actually led to a decrease in abortions. Instead, they've made them harder and significantly riskier. It's a worse situation all round. Deaths like these are senseless: a tragedy at the hands of a fundamentalist ideology with no basis in science.

#Democracy

[Link]

[Sarah Perez at TechCrunch]

"Mozilla is exiting the fediverse. Though the concept of the open social web, also known as the fediverse, has been picking up momentum ever since Meta last year introduced its first-ever federated app, Instagram Threads, Firefox maker Mozilla on Tuesday announced it would be ending its experiment in running a server on the fediverse. The server, Mozilla.social, today connects users with the Mastodon social network, an open source rival to Twitter/X. It will be shut down on December 17."

I wish Mozilla had taken a more ambitious approach to the fediverse, rather than running a Mastodon instance for a handful of people. An organization of its size could have prototyped different kinds of social media on the fediverse, or incubated disparate projects running on the protocol. It could even have experimented with adding social functionality directly to Firefox. Instead, precisely none of that happened, and its instance was apparently used by 270 people or so.

Sarah Perez points out that this isn't the only initiative that's been shuttered recently:

"Among those products affected by the pullback were its VPN, Relay, and Online Footprint Scrubber, in addition to its Mastodon instance, the company said at the time. Meanwhile, its virtual world Hubs was shut down."

Mozilla itself has a lot of potential but never seems to quite realize it: it doesn't seem to be very good at building a joined-up product strategy, which has led its existence to become increasingly at risk. The vast majority of its funding (nearly 90%) has come from Google's payments to the organization in exchange for being the default search engine pick - and now that's under threat.

There's a need for a mission-driven organization with Mozilla's values that executes more fearlessly, where product voices hold more weight vs open source engineering discussions. But, right now, I don't think it exists.

#Fediverse

[Link]

[Simon Willison]

"Whether or not you enjoy MrBeast’s format of YouTube videos, this leaked onboarding document for new members of his production company is a compelling read."

It really is fascinating. It's also really badly written, which says a lot about the priorities MrBeast instills in his team. Simon points out that video is ingrained in the culture:

"Which is more important, that one person has a good mental grip of something or that their entire team of 10 people have a good mental grip on something? Obviously the team. And the easiest way to bring your team up to the same page is to freaken video everything and store it where they can constantly reference it. A lot of problems can be solved if we just video sets and ask for videos when ordering things. [...] Since we are on the topic of communication, written communication also does not constitute communication unless they confirm they read it."

MrBeast will be studied for decades to come: a piece of the culture that, like him or not, is genuinely new. This document is a key to understanding what he does.

#Media

[Link]

[Media Economies Design Lab at University of Colorado Boulder]

"The Media Economies Design Lab at the University of Colorado Boulder is launching a 5-month process of mentorship and peer-to-peer learning, empowering veteran community builders to adopt emerging open social networks."

This seems like a wonderful initiative for the right people, as well as for new social networks. The incubator covers a broad set of networks that include the fediverse, Matrix, Bluesky, and Nostr.

From the site:

"Participants will meet as a full group monthly and receive ongoing, 1:1 project-specific technical support as they seed and grow new communities using open social media tools of their choosing. Participation is fully remote and all sessions will be conducted in English. All participants should commit to making strides toward community-building in open social media by the end of the program. Completion of the full program, from November 2024 to April 2025, will be compensated with a stipend of $3,500 USD."

No prior experience is needed.

I love that the Media Economies Design Lab is doing this. I'm very curious to see the cohort as it emerges!

#Fediverse

[Link]

[Matt Burgess at Wired]

"Today, a group of six computer scientists are revealing a new attack against Apple’s Vision Pro mixed reality headset where exposed eye-tracking data allowed them to decipher what people entered on the device’s virtual keyboard. The attack, dubbed GAZEploit and shared exclusively with WIRED, allowed the researchers to successfully reconstruct passwords, PINs, and messages people typed with their eyes."

Fascinating stuff. This attack doesn't work with a normal laptop or device because we tend to look at the screen as we type instead of the keys. But on the Apple Vision Pro, your gaze is your pointer. By tracking what you're paying attention to, attackers can understand exactly what you're typing, including sensitive information.

Apple has patched the problem, presumably by making its virtual avatars just a little bit more dead in the eyes. But as more eye-based interfaces roll out, more exploits will surely be discovered. As we reveal more of ourselves in virtual space, more of our secrets become apparent, too.

#Technology

[Link]

[David Allen Green at The Law and Policy Blog]

"In essence: this endorsement is a masterpiece of practical written advocacy, and many law schools would do well to put it before their students."

This is a fascinating breakdown of Taylor Swift's endorsement of Kamala Harris: not just the what of her endorsement, but the linguistic how. As David Allen Green says, it's worth studying.

It comes down to this:

"The most effective persuasion is often to lead the listener or reader to making their own decision – and to make them feel they are making their own decision."

Taylor Swift's endorsement really matters, and was clearly planned carefully. This wasn't a dashed-off Instagram description, and I'll certainly be learning from it.

#Culture

[Link]

[Mathias Felipe de Lima Santos at NiemanLab]

"To survive, journalism must continue to embrace technology. But doing so should never cost newsrooms their independence. News organizations should prioritise building direct relationships with their audience to reduce reliance on third-party platforms. They should also stay informed about evolving regulations, and actively participate in policy discussions shaping the future of the news-tech relationship."

Exactly.

There are good points made here about paying for news technology. My belief is that there's value in sharing resources between newsrooms. Building and supporting technology as a commons could help newsrooms further their goals while staying independent. This shared model also prevents newsrooms from each developing the same commodity technology, which across the industry is a huge misuse of resources.

Clearly, too, a strong independent social web benefits both newsrooms and news consumers. A strong fediverse helps empower newsrooms to build direct relationships. Newsrooms should support this movement.

Regardless of the solution, it's good to see these topics being brought up in the news space. These represent the problems and the existential threats: now it's up to the industry to act.

#Media

[Link]

[Lucas da Costa at Briefer]

"Some people call our strategy "open-core" and that's technically right. Still, I'd rather say that we have two pieces of software: one that is open-source and another that is not. I think that's more honest because we're not trying to hide the fact that we're selling a non-open-source version of our software."

This is a pretty honest take on open sourcing a product in a VC-funded startup, which needs to maintain a certain level of valuation growth to justify its investment.

Someone in edtech once told me that if I held back any of a product I was building that they would tell their substantial network not to use it. I don't think that's fair: I'm not sure there's much to be gained by making features that are mostly used by wealthy companies free. This is particularly true when owning your licensing means you still retain optionality to provide a lower-cost or zero-cost license for certain organizations.

I also like this reason for open sourcing their core product:

"Finally, by going open-source we commoditize our competitors' core functionality. This means they now have to compete against us in terms of innovative features, performance, and price, all of which are usually not their strong suits, let's be honest."

When executed well, and used against high-priced enterprise software in particular, this approach deflates closed-source business models and can be a real competition lift. I like that Briefer is naming that.

The one piece I don’t agree with is this:

"Open-source helps us manage Briefer's roadmap along with our users because there will be more of them, and because they'll have access to the source code. That way, they can help us figure out where to go, and help us get there by implementing what they need."

My experience in open source is that it doesn't absolve you from needing to keep a tight hand on the product steering wheel. Your open source community can actually muddy the water here, because open source users aren't always the same thing as customers, and may need a different set of features or functionality. Maintaining a coherent product vision is harder in open source, not easier.

Still, this was a lovely post to read, and I appreciate the open thinking. It certainly made me want to check Briefer out.

#Technology

[Link]

[Hunter Walk]

"I asked some investor friends to share, as the title suggests, one thing they wished people better understood about venture capital. There were no ground rules other than to specify that ‘people’ could be founders, politicians, LPs, etc and that it would be default attributed but anonymous if they desired."

Hunter Walk's ongoing series of inside perspectives from venture capital is brilliant: both nuanced and real. I didn't spend long as a mission-driven investor - two years - but many of these perspectives resonate.

I wish I'd read these before I started my investment journey: so much more of this job is about building funnels and ensuring that your portfolio has adequate opportunities for follow-on investment. You need to have a point of view / thesis on the future of technology and applicable markets, for sure, but there's much more on-the-ground sales work than is popularly discussed.

If you're raising money - or investing - they're worth checking out. It's an ongoing series, so I recommend just subscribing to Hunter's blog for more.

#Business

[Link]

[iA]

"In a text editor, chapters are files. Organizing your files is work, but in a large text body it’s essential work. Your book or thesis will grow from it and get stronger as you clarify the structure. With iA Writer 7.2, structuring large writing projects has become a lot easier."

In other words, my favorite text editor just got a big upgrade for anyone writing large projects (hey, that's me!).

I've long been an iA Writer superfan: all my blog posts are written in it, and I use it as the starting point for most meaningful documents. This new update brings it into direct competition with Ulysses, another markdown text editor I love. I've been using iA Writer for short-form writing and Ulysses for longer-form writing (I have a very large book draft in there right now). But now, potentially, I can do it all from one app.

What it doesn't seem to do - yet - is the kind of file re-ordering that Ulysses excels at, so I can move scenes and chapters around each other with ease. From this post, it sounds like that will come:

"Tree view is the first step toward a document outline. Tree view is the technical foundation for offering a more detailed view of the document structure. All we can say for now is that it will work very much like tree view, just inside the document."

It's all great work. This level of care and attention in a text editor really matters. I'm grateful that iA exists.

#Technology

[Link]

[Elahe Izadi at The Washington Post]

I adore The 19th. It's making big moves, and that's good news for everyone.

"What [Emily] Ramshaw and fellow co-founder Amanda Zamora started in January 2020 — a newsroom with just one reporter and no website — has grown into a digital operation that has raised nearly $60 million and employs 55 people. And in a sign of its growing ambitions, the 19th has now hired veteran news executive LaSharah Bunting, CEO of the Online News Association, as its first vice president, a role created to build up the 19th’s fundraising and budget operations."

It's also grown an endowment, which allows it to have a safety net and continue to grow and experiment. The ambition for the endowment to underwrite the newsroom's operations is meaningful: this would represent a fund designed to allow reporting on gender, politics, and policy to be undertaken sustainably. I don't know of any other similar fund in media.

Not mentioned here but extremely relevant: the amazing work Alexandra Smith, its Chief Strategy Officer, has been doing to redefine how to think about audience and reach on a fragmented web.

These are all signs of a forward-thinking newsroom that isn't content to simply accept the status quo - and, crucially, plans to stick around.

#Media

[Link]

[Paul Graham]

"In effect there are two different ways to run a company: founder mode and manager mode. Till now most people even in Silicon Valley have implicitly assumed that scaling a startup meant switching to manager mode. But we can infer the existence of another mode from the dismay of founders who've tried it, and the success of their attempts to escape from it."

Please forgive the Paul Graham link: this is a genuinely good point about running companies. And I don't think it's limited to startups: the dichotomy isn't between "founder mode" and "manager mode", but between purposeful companies built to be communities aiming at a focused goal and institutions that can move slower and less efficiently.

Skip-level meetings should be normal. Flat hierarchies are good. Everyone in a company should have the ability to have the ear of the CEO if they need it - and, likewise, the CEO should be able to freely talk to anyone in a company. A good idea can come from anyone; people with exceptional talent can show up anywhere on the org chart. Less regimentation and less bureaucracy allow those people to flourish - and, in turn, allow the organization to make better choices.

It's also a representation of what matters to an organization. Hierarchies emerge from people who care about hierarchy and chains of command; flatness emerges from people who just care about getting stuff done. The latter, in my view, always makes for a better place to work.

#Startups

[Link]