Ben Werdmuller

I consider myself really lucky that people stop by and read my posts. Thank you!

Every year I like to pause and ask folks a little bit more about themselves, what they’re worried about, and what they’re interested in. It’s a really short, anonymous survey that helps me out on many levels. Feedback is always a gift.

This year, there’s another reason for this survey, too: I’m considering offering services to organizations that want to tackle specific technology challenges — particularly educational institutions, newsrooms, and technology companies for whom engineering is not their primary activity.

I’ve led product design and innovation sprints for many different companies and have been directly involved in multiple innovation accelerators. I have a well-tested process that really works, and this is one way I might be able to add value.

This survey will help me figure out which problems and ideas people are thinking about, which will help me figure out how helpful I can be.

But I’d also just love to know what you’re thinking about.

To fill in the survey, click here. Thank you for your feedback!

Over the weekend, I started to notice a bunch of artists moving to Cara, a social network for artists founded by Jingna Zhang, herself an accomplished photographer.

The fediverse is a decentralized cooperative of social networks that can interact with each other: a user on one network can follow, reply, like, and re-share content from a user on another network. The whole thing depends on an open standard called ActivityPub, shared community norms, and a cooperative culture.

Of course, my first reaction was that Cara should be compatible with the fediverse so that its content could be more easily discoverable by users on social networks like Threads, Flipboard, and Mastodon. Cara is explicitly set up to be a network for human artists, with no AI-generated content, which will be increasingly valuable as the web becomes flooded with machine-made art. The fediverse would allow them to publish on sites like Cara that are set up to support their needs, while finding a broad audience across the entire web.

From its About page:

With the widespread use of generative AI, we decided to build a place that filters out generative AI images so that people who want to find authentic creatives and artwork can do so easily.

[…] We do not agree with generative AI tools in their current unethical form, and we won’t host AI-generated portfolios unless the rampant ethical and data privacy issues around datasets are resolved via regulation.

I’d love to follow artists on Cara from my Mastodon or Threads accounts. But how does Cara’s AI stance square with the fediverse? How might artists on Cara find a broad audience for their work across the web without risking that art being used as training data without permission?

The first thing a site can do to prevent its content from being used as training data is to add exclusion rules to its robots.txt file. These theoretically prevent crawlers owned by model vendors like OpenAI from directly accessing art from the site. There is nothing that legally binds crawlers from obeying robots.txt; it’s less enforceable than a handshake agreement. Still, most claim that they voluntarily do.

But even if robots.txt was an ironclad agreement, content published to the fediverse doesn’t solely live on its originating server. If Cara was connected to the fediverse, images posted there could still be found on its servers, but they would also be syndicated to the home servers of anyone who followed its users. If a user on Threads followed a Cara user, the Cara user’s images would be copied to Threads; if a user on a Mastodon instance followed that user, the images would be copied to that Mastodon instance. The images are copied across the web as soon as they are published; even if Cara protects its servers from being accessed by AI crawlers, these other downstream fediverse servers are not guaranteed to be protected.

By connecting to the fediverse, one might argue that servers implicitly license their content to be reused across different services. This is markedly different from RSS, where this is explicitly not the case: there is legal precedent that says my RSS feed cannot be used to republish my content elsewhere without my permission (although you can, of course, access its content in a private feed reader; that’s the point). But on the fediverse, the ability to reshare across platforms is core functionality.

The following things are all true:

• Content published to the fediverse may be both re-copied to and served from other peoples’ servers
• Those servers may have different policies regarding content use
• In the absence of a robots.txt directive, AI crawlers will scrape a website’s data, even if they don’t have the legal right to
• Some servers may themselves be owned by AI vendors and may use federated content to train generative models even without the use of a scraper

As a result, there is no way an author can protect it from being used in an AI training set. The owners of a fediverse site wouldn’t have the right to make a deal with an AI vendor to sell the content it hosted because they wouldn’t have the copyright to all of that content in the first place. But because AI crawlers greedily scrape content without asking for permission, unless the site explicitly opts out with robots.txt, it doesn’t matter.

This leads me to a few conclusions:

• It is a moral obligation for every fediverse site to prevent crawling of federated content by robustly setting robots.txt directives at a minimum
• Discussions about adding content licensing support to the fediverse are even more important than they appear
• Someone needs to legally prevent AI vendors from using all available data as training fodder

A fediverse (and a web!) where Cara can safely join while adhering to its principles is a more functional, safer network. To build it we’ll need to support explicit licensing on the fediverse, create a stronger standard for user protections across fediverse sites, and seek more robust legal protections against AI crawler activity. While these are ambitious goals, I believe they’re achievable — and necessary to support the artists and content creators who make the web their home.

I’m enamored with Matt Webb’s unoffice hours: a way to chat with him about anything, without needing to email him first, for 30 minutes.

As Matt says:

I loved those open conversations over coffee in the Before Times. There’s an ostensible reason to connect, so you talk about work, or compare notes about an idea, or whatever. But then the unexpected emerges. (Sometimes you have to hunt for it.) There are things in your head that you only know are there when you say them. And there are encounters with new ideas and new perspectives.

Exactly. So let’s do it.

Introducing my own Unoffice Hours: I’ve set aside a little time on Fridays to connect about anything.

Here are some topics that might be interesting to chat about:

• Feedback on a project you’re working on (startups, software, a writing project)
• Following up on something I’ve written in this space
• Product and technology strategy in the public interest (news, education, libraries, other mission-driven organizations)
• The indie web
• Fostering a collaborative organizational culture
• Saying hello

Matt calls the effect “manufactured serendipity”; I call it “intentional serendipity,” but the intent is the same. It’s good to chat and meet people, and you never know where it will lead.

To book a 30-minute chat, click here.

Without serious intervention, newsrooms are going to disappear. Changes to social media and the advent of generative AI threaten their businesses and the impact of their work. They need to own their online presences outright and build direct relationships with their communities—and they need to do it now.

Social media audiences are plummeting. Less than 35% of internet searches lead users to click on a website. The views and engagement that newsrooms depend upon to survive are disappearing.

It’s happening quickly. Semafor’s Max Tani reported recently:

Washington Post CEO Will Lewis is introing the paper’s new “Build It” plan today. In a meeting with staff, he noted that the paper lost $77 million over the past year, and saw a 50% drop off in audience since 2020: “To be direct, we are in a hole, and we have been for some time."

Addressing this challenge will require radical changes to how newsrooms invest in and build technology.

In this post, I’ll attempt to describe the challenges in more detail and then discuss how they can be more adequately addressed.

Some context: my move into news

I’ve recently gained a new perspective on these challenges. For over a decade, I’ve worked adjacent to news and journalism. I’ve seen the industry as an engineer, startup founder, product lead, investor, and advisor. More recently, I decided I could be more useful in directly leading technology efforts inside newsrooms. It’s been eye-opening, rewarding work.

My experience alongside news was diverse. I built product for newsrooms, founded a startup used by public media, invested in early stage media startups, and have taught human-centered product design to teams at organizations like the New York Times and the Associated Press, as well as at institutions like the Newmark School of Journalism and the Harvard Kennedy School of Government. I’ve built software, founded, grown, and supported startups, and taught product design to some of the biggest names in journalism.

My immersion inside newsrooms has been much more recent. ProPublica investigates abuses of the public trust by government, businesses, and other institutions. I’ve worked on technology strategy for the last year, first as a contractor, and now as its Senior Director of Technology. Before that, I was the first CTO at The 19th, which reports on the intersection of gender, politics, and power.

I made this career shift at a pivotal moment for journalism—though it seems every moment for journalism over the last fifteen years has felt pivotal. The industry has struggled to weather the seismic shifts brought about by the internet, which have impacted its business, the state of our politics, and public discourse. It’s been a struggle for decades.

The audience threat

It’s getting harder and harder for newsrooms to reach their audiences — and for them to sustain themselves.

I’ve often remarked that journalism treats the internet as something that happened to it rather than something it can actively shape and build, but it at least had some time to adjust to its new normal. The internet landscape has been largely static for well over a decade — roughly from the introduction of the iPhone 3G to Twitter’s acquisition by Elon Musk. People used more or less the same services; they accessed the internet more or less the same way. Publications and online services came and went, but the laws of physics of the web were essentially constants.

Over the last year in particular, that’s all changed. Shifts in the social media landscape and the growing popularity and prevalence of generative AI have meant that the rules that newsrooms began to rely on no longer hold.

At their heart, online newsrooms have a reasonably simple funnel. They publish journalism, which finds an audience, some of which either decide to pay for it or view ads that theoretically cover the cost of the work. Hopefully, they will make enough money to publish more journalism.

This description is a little reductive: there are lots of different revenue models in play, for one thing. I’m particularly enamored with patronage models that allow those with the means to support open-access journalism for anyone to read freely. Still, some are entirely ad-supported, some are sponsored, and others are protected behind a paywall (or some combination of the above). For another, journalism isn’t always the sole driver of subscriptions. The New York Times receives tens of millions of subscribers from its games like Wordle and Connections, as well as its Cooking app.

Still, there are two pivotal facts for every newsroom: their work must reach an audience, and someone must pay for it. The first is a prerequisite of the second: if nobody discovers the journalism, nobody will pay for it. So, reaching and growing an audience is crucial.

For the last decade and a half, newsrooms have used social media and search engines as the primary way to reach people. People share stories across social media—particularly Facebook and Twitter—and search for topics they’re interested in. It’s generally worked.

Over the last year, social media has radically fragmented. Twitter transformed into X under its new management; users began to flee the platform in the face of more toxic discourse, and active use plummeted. Facebook is slowly declining and referrals to news sites have fallen by 50% over the last year. Instagram is not in decline. Still, it’s harder to post links to external sites there, which means that while newsrooms can reach users, they have more difficulty converting them to subscribers.

On top of these changes, we’ve also seen the rise of Threads, Mastodon, and Bluesky, as well as a long tail of other social apps, platforms, and forums on which to reach people. Audiences on social media used to be found in a very small number of places and are now spread out across very different platforms. The fediverse and AT Protocol also yield different problems: which instance should a newsroom choose to make its home? How can it measure engagement in what it posts in a decentralized system so that it knows what’s working and where it should continue to invest its meager resources?

Much has been written about newsrooms’ inability to move away from X even as it has become a hotbed of white supremacy and far-right rhetoric. The honest truth is that it still drives significant traffic to their websites, and in an environment where traffic referrals are dropping overall, intentionally further deepening the traffic shortfall is understandably not a career risk newsroom leaders are willing to make.

Social media isn’t the only way newsrooms are finding it harder to find an audience. Even search engines, long the stalwarts of the web, are moving away from referring traffic.

As search engines move to make AI-driven answers more prominent than links to external websites, they threaten to reduce newsroom audiences, too. More than 65% of Google searches already ended without a click to an external site. Now, it’s planning to roll out AI-driven answers to over a billion people. It’s not that other links are going away entirely. Still, because AI answers are the most prominent information on the page, clickthroughs to the external websites where the answers were found initially will be significantly reduced.

A similar dynamic is at play with the rise of AI services like ChatGPT, emerging as stiff competition for search engines like Google. These services answer questions definitively (although not always correctly), usually with no external links on the page. ChatGPT could learn from a newsroom’s articles and display information gleaned from an expensive investigative story while never revealing its source or allowing readers to support the journalism.

Generative AI models seem like magic: they answer questions succinctly, in natural language, based on prompts that look a lot like talking to a real human being. They work by training a neural network on a vast corpus of information, often obtained by crawling the web. Based on these enormous piles of data, AI engines answer questions by predicting which word should come next: a magic trick of statistics empowered by something close to the sum total of human knowledge.

That’s not hyperbole. It’s not a stretch to say that OpenAI’s ChatGPT and Google’s Gemini were trained on most of the internet, including websites, published books, videos, articles, art, science, and research. They couldn’t function without this data — but, ironically, they rarely credit their sources for any of it. Users see the benefit of fast answers; the sources of that information are starved of oxygen.

We’re at the foothills of both changes: social media is likely to fragment further, and generative AI will become even more prevalent as it becomes more powerful. Newsrooms can no longer rely on their old tactics to reach their audiences, and they will need to build new tactics that take these trends into account if they hope to survive.

Some models are more resilient than others

The 19th’s Alexandra Smith recently wrote about the state of play in Columbia Journalism Review:

In our current reality, journalism exists in various formats splintered across platforms and products. People are just as likely to get their news on Instagram as from a news website. It no longer makes sense to rely primarily on measuring readership by traditional website metrics.

This is a depressing fact if you rely on paywalled subscriptions or ad impressions. Nobody’s looking at your ads if they’re consuming your journalism off-platform, and how can you possibly get someone to subscribe if they never touch your app or website? Instagram and TikTok don’t have built-in subscriptions.

Over the years, many people have suggested micropayments — tiny payments you make every time you read a news article anywhere — but this depends on everyone on the web having some kind of micropayment account that is on and funded by default and the platforms all participating. It’s a reasonable idea if the conditions are right, but the conditions will never be right — and, like subscription models, it shuts out people who can’t pay, who are often the people most in need of public service journalism to begin with.

For newsrooms like The 19th, the picture is much rosier: like most non-profit newsrooms, it depends on donors who support it based on its journalistic impact. (The same is true of ProPublica, my employer.) That impact could occur anywhere, on any platform; the trick is to measure it so donors can be informed. Alexandra developed a new metric, Total Journalism Reach, that captures precisely this:

Right now, it includes website views; views of our stories that are republished on other news sites and aggregation apps, like Apple News; views of our newsletters based on how many emails we send and their average open rates, reduced for inflation since Apple implemented a new privacy feature; event attendees; video views; podcast listens; and Instagram post views.

This is clearly valuable work that will help newsrooms like The 19th prove their impact to current and potential donors. The quote above doubles as a useful example of the places The 19th is reaching its audience.

It’s worth considering how these might change over time. Some of the media Alexandra describes are inside The 19th’s control, and some are less so.

Supplier power

In his classic piece ‌How Competitive Forces Shape Strategy, Michael Porter described five forces that shape competitive strategy. One of them is supplier power: the ability of providers of essential inputs to a business to exert influence over the organization. If suppliers to the industry have too much power — because there are few alternatives, for example — they can effectively force the company’s strategy by raising costs or enforcing adverse policies.

Newsrooms’ platforms for reaching their audiences, such as social media and Apple News, currently have outsized supplier power over the journalism industry. As a result, the industry is disproportionately vulnerable to the effects of business decisions made by the owners of those platforms.

In April, Instagram introduced a new automatic filter, switched on by default, to remove political content, which affected many newsrooms, and illustrates the kind of changes service providers can make on a whim.

Newsrooms on Apple News tend to see a multiple of the number of reads they see on their websites, but Apple could pull the product tomorrow. Even today, the number of views you get highly depends on which stories the Apple News team chooses to highlight. Ads in publications on Apple News need to use Apple’s ad network. It’s a closed shop. Apple News is only successful because it comes installed by default on Apple devices; hundreds of similar news aggregators have all failed to survive in their own right. It’s a precarious place to hang your hat.

We’ve already discussed the impact of search engine design decisions like prioritizing AI over click-through rates. Only one search engine is prominent enough to have disproportionate supplier power: a position Google has bought by spending over $21 billion a year to be the default search engine in every web browser.

However, not all conduits to readers have this outsized supplier power as a feature. Social media platforms, search engines, and news aggregators are all run by wealthy individual companies like X, Meta, Google, and Apple, who have the potential to exert their power. If you choose to leave them for any reason, you’re also leaving behind the relationships you’ve built up with your audience there: there’s no audience portability.

In contrast, email, podcasts (real podcasts, not the single-platform kind where you ink an exclusive deal with Spotify or Audible), and the web are well-used methods to reach audiences that aren’t owned by any platform. There are certainly market leaders for each communication type. Still, each is based on an open protocol that no single company controls — which means, for those methods, no supplier can exert adverse supplier power. If one service provider misbehaves, you can simply switch to another without losing functionality. You can bring your audience with you. They’re safer methods, as long as enough readers want to be reached in those ways.

That’s why so many publications have focused their strategies on their email newsletters. Everyone already has an email address, and (barring technical difficulties) if a publisher sends a subscriber a message, they’re guaranteed to receive it. Moreover, people engaged enough to hit the “subscribe” button are far more likely to convert to donors or upgrade to a paid subscription.

Newsletters, unfortunately, are also in decline. Open rates have fallen over the last decade; Gmail’s dominant position and aggressive filtering have made it harder for newsletters to be noticed; there’s more competition for attention. There aren’t any great ways for new readers to discover newsletters — those subscription pages are subject to the same internet traffic dynamics as articles. It’s getting harder and harder to direct new visitors to subscribe, which is why we see more overt “please subscribe” popup overlays on news sites. The focus has needfully shifted to converting more existing subscribers into donors or customers rather than widening the funnel and finding more newcomers.

Newsrooms need alternative media that allow them to make direct connections with their audiences. These media must be free from undue supplier power and have a large base of existing users that can be tapped into.

So what else is out there?

The answer is not much. Yet.

The innovation squeeze

Most non-profit newsrooms have tiny technology teams. The 19th, when I was CTO, had two engineers in addition to me; ProPublica has four. (Other interactive developers work on standalone stories but don’t address platform needs.) In contrast, I led a team of twenty-two engineers at the last startup I worked at, and we had over a hundred at Medium.

To bridge that gap, there is a small community of digital agencies that make supporting newsroom platform needs a core part of their business. Probably the most famous are Alley and Upstatement, but there are around a dozen more that are actively used by newsrooms.

They do beautiful work and are an excellent way for a newsroom to start strong with a modern brand and a well-functioning web platform. I strongly recommend that a new newsroom consults with them.

There is an emerging dynamic, though, where the technology vision for a newsroom is outsourced to the agencies. As we’ve discussed, a newsroom’s success and impact depend highly on core internet technologies like the web and email. Newsrooms quite reasonably spec and build a platform based on what will work well today. However, because the vision and expertise for harnessing the internet lie with the agencies, they don’t have any meaningful technology capability for innovating around what will work well tomorrow.

Newsrooms absolutely need to focus on today. That’s an obvious prerequisite: they must meet their audiences, subscribers, and donors where they’re at right now. However, they also need to be aware of what is coming down the road and prepared to experiment with, engage with, and potentially help shape new technologies that could impact their businesses in the future. If the internet changes, they need to be ready for it. To reference an overused Wayne Gretzky quote: you need to skate to where the puck will be, not where it is right now.

Nobody knows for certain where the puck will be. That means newsrooms need to make bets about the future of technology — which, in turn, means they must have the capacity to make bets about the future of technology.

Most newsrooms already have technical staff who maintain their websites, fix broken platform stacks, and build tools for the newsroom. These staff must also highlight future business risks and allow them to experiment with new platform opportunities. In a world where newsrooms rely on the internet as a publishing mechanism, technology expertise must be integral to their strategy discussions. And because technology changes so quickly and unpredictably, maintaining the time, space, and intellectual curiosity for experimentation is critical.

Nothing will work, but anything might

Experimentation doesn’t need to be resource-intensive or time-consuming. Alongside in-house expertise, the most important prerequisite is the willingness of a newsroom to test: to say “yes” to trying something out, but being clear about the parameters for success, and always rooting success or failure in a concrete understanding of their communities.

I’ve written before about how, if the fediverse is successful, it will be a powerful asset to media organizations that combines the direct relationship properties of email with the conversational and viral properties of social media. At the same time, there’s no doubt that the network is relatively small today, that the experience of using Mastodon falls short of corporate social networks like the Twitter everyone remembers, and that features like blocking referrer data makes life much harder for audience teams. There are lots of good reasons for a resource-strapped management team to say “no” to joining it.

At the same time, because it has the potential to be interesting, some newsrooms (including my employer) are experimenting with a presence. The ones who make the leap are often pleasantly surprised: engagement per capita is dramatically higher, particularly around social justice topics. Anecdotally, I discovered that posting a fundraising call to action to the network yielded more donations than from every other social network — combined.

It’s worth looking at Rest of World’s “More Ways to Read” page — a massive spread of platforms that runs the gamut from every social network to news apps, messaging platforms, audio, newsletters, and RSS feeds. The clear intention, taken seriously, is to meet audiences where they’re at, even if some of those networks have not yet emerged as a clear winner. All this from a tiny team.

However, experimenting isn’t just about social media. It’s worth experimenting with anything and everything, from push notifications to website redesigns that humanize journalists to new ways for communities to support the newsroom itself.

On the last point, I’m particularly enamored with how The 19th allows members to donate their time instead of money. Understanding that not everyone who cares about their mission has discretionary spending ability, they’re harnessing their community to create street teams of people who can help promote, develop, and share the work in other ways. It’s brilliant — and very clearly something that was arrived at through an experimental process.

I learned a formal process for human-centered experimentation as a founder at Matter, the accelerator for early-stage media startups, which changed the way I think about building products forever. A similarly powerful program is now taught as Columbia Journalism School’s Sulzberger Fellowship. If you can join a program like this, it’s well worth it, but consultants like Tiny Collaborative’s Tran Ha and Matter’s Corey Ford are also available to engage in other ways. And again, the most important prerequisites are in-house expertise and the willingness to say “yes”.

To achieve this, they must shift their cultures. The principles of experimentation, curiosity, and empathy that are the hallmarks of great journalism must also be applied to the platforms that power their publishing and fundraising activities. They must foster great ideas, wherever they come from, and be willing to try stuff. That inherently also implies building a culture of transparency and open communication in organizations that have, on average, underinvested in these areas. As Bo Hee Kim, then a Director of Newsroom Strategy at the New York Times, wrote back in 2020:

Companies will need to address broader issues with communication, access, and equity within the workplace. Leaders will need to believe that newsroom culture has a bigger impact on the journalism than they understood in previous years — that a strong team dynamic is as important as their sharp and shiny stars. Managers are key to this transition and will need to reset with a new definition of success, followed by support and training to change.

Gary P. Pisano in Harvard Business Review:

Too many leaders think that by breaking the organization into smaller units or creating autonomous “skunk works” they can emulate an innovative start-up culture. This approach rarely works. It confuses scale with culture. Simply breaking a big bureaucratic organization into smaller units does not magically endow them with entrepreneurial spirit. Without strong management efforts to shape values, norms, and behaviors, these offspring units tend to inherit the culture of the parent organization that spawned them.

Creating an innovative culture is complex, intentional work. But it is work that must be done if news organizations are to innovate and, therefore, survive.

Conclusion

The internet is changing more rapidly than it has in years, creating headwinds for newsrooms and jeopardizing independent journalism’s viability. We need those organizations to exist: they reduce corruption, inform the voting public, and allow us to connect with and understand our communities in vital ways.

These organizations must own their digital presence outright to shield themselves from risks created by third parties that wield outsized supplier power over their business models. They must build direct relationships with their communities, prioritizing open protocols over proprietary systems. They need to invest in technology expertise that can help them weather these changes and make that expertise a first-class part of their senior leadership teams.

To get there, they must build an open culture of experimentation, where transparency and openness are core values cemented through excellent, intentional communication. They must be empathetic, un-hierarchical workplaces where a great idea can be fostered from anywhere. They must build a mutual culture of respect and collaboration between editorial and non-editorial staff and ensure that the expertise to advise on and predict technology challenges is present and well-supported in-house.

Experimentation and innovation are key. Newsrooms can discover practical ways to navigate these challenges by testing new strategies, technologies and mindsets. The road ahead is challenging, but with strategic investments and a forward-looking approach, newsrooms can continue to fulfill their vital role in a well-functioning democratic society. The best time for action was ten years ago; the second best time is now.

I recently wrote a short aside about stock options:

But in general, for regular employees, I think options are rarely worth it. They typically require an up-front investment that many employees simply can’t make, so it’s a bit of a fake benefit to begin with, and their future value is little more certain than a lottery ticket.

Hunter Walk kindly reshared it on a few networks with some of his own thoughts; a conversation with Tony Stubblebine arose in the comments that Hunter wrote up as its own post. In particular, he says it helped him articulate the ups and downs of private stock to the average person:

For much of a startup’s life new FUNDING VALUATIONS are LEADING indications of POTENTIAL. They are what someone is willing to pay for shares today based on what they believe the company CAN DO in the FUTURE.

DOWN ROUNDS and RECAPS are LAGGING indications of PERFORMANCE. They are what someone is willing to pay for shares today based upon what the company HAS DONE in the PAST.

It’s a great post, and the comments from Tony were thoughtful. Which led me to feeling a bit bad about how flippant and imprecise my original post had been.

So, on that note, I’d love to define options, make some corrections, and dive a little deeper into my core argument.

The ins and outs of options

First, let’s define options and explain why they’re so common as a factor of startup compensation.

An option is the right to buy a specified number of shares in a company at a specific price. That price is typically defined by an external auditor. It’s good practice for this to happen once a year, but it’ll also be triggered when the company raises a round of equity funding (i.e., when it sells shares to outside investors in order to raise significant capital).

If a startup were simply to grant stock directly to employees, it would be taxable as compensation. Options are almost always non-taxable at the point where they are issued, so they’re a favorite way to give employees the ability to see some of the potential upside in a venture.

Typically in a startup you’ll receive an option grant as part of your compensation package. So, for example, you might receive the right to buy (“exercise”) 40,000 shares at 50 cents a share (the “strike price”). This is almost always on what’s called a vesting schedule: you won’t be able to buy any shares in the first year, but then when you cross that threshold (the “cliff”), you’ll be able to buy 25% of your allocation (the first 10,000 shares in my example). Over the next three years, the amount of your allocation that you can exercise will increase proportionally, until you can buy them all at the end of four years.

If you leave the company, you usually only have 90 days to exercise whichever options have vested. Some particularly progressive companies extend that exercise window — sometimes to a couple of years. But for 80-90% of startups, it’s 90 days.

If the startup is excited about keeping you, you may find that they’ll grant you more options periodically, each with their own vesting schedules. This, they hope, will keep you at the company.

In my example above, you might have done the math to realize: 40,000 shares at 50 cents a share is $20,000. You would need to lay out that amount of money to acquire the shares — and you need to hope that the company’s shares increase in value in order to see any upside.

If the company’s share price has increased in the time between the options were granted and when the employee exercises them, the difference is taxable. In the above example, recall that my options are for 40,000 shares at 50 cents a share. Let’s say I choose to exercise them all at the end of my four year vesting period: as we’ve discussed, I pay $20,000. But let’s say that the real fair market value has risen to 75 cents a share. The difference between 40,000 shares at 50 cents and 40,000 shares at the market value of 75 cents is $10,000 is usually taxed as income. So I’m actually paying $20K + income tax on another $10K. (This isn’t by any means the full extent of potential tax implications; I’m not going to touch ISOs and AMT in this post, for example.)

Early employees, who join before most funding rounds have taken place, will receive options with a very low exercise price. Later employees will usually receive options with a higher price, because more growth and fundraising has taken place in the interim. (Down rounds and recaps are certainly possible, though: many startups go through tough times where their valuation decreases. Not every graph always goes up and to the right.)

In both cases, any stock they buy is largely illiquid. Because the startup is likely a private company rather than a publicly traded one, their shares are not liquid. They will need to wait for the company to go public or hope that management will allow them to trade their stock on the secondary market.

Some corrections

So the first thing to say is: no, options are not really like a lottery ticket. They are a sort of gamble, but it’s one where (depending on your position, seniority, and what size the company was when you joined) you have a say in the outcome.

The second, which I’ve already corrected in the original post is: as Hunter pointed out in his post, a recap is not the thing that actually lowers the stock price. It’s a trailing signal of what the company has already done. A change in stock price is an effect of what has already happened.

And a clarification: options don’t require an up-front investment at the time that they’re granted. You invest at the time when you exercise them, which may still be as a lump sum.

Why I think exercising options isn’t worth it for many employees

If you’re on a rocket ship startup, exercising your options is almost certainly worth it (depending on the strike price of your particular options grant). The problem is: how do you know you’re on a rocket ship? Or, given that most startup employees won’t be part of a startup with hockey-stick growth, how can you be reasonably sure that your company will grow in such a way that exercising your options is worth it?

90% of startups fail. That doesn’t mean that every startup has an equal 1 in 10 chance of success: a lot depends on a range of factors that include internal culture, management expertise, execution quality, and market conditions. Still, there is not a small amount of luck involved. Most startups won’t make it.

You should never make an investment that you can’t afford to lose. As Hunter says in his post:

Don’t behave as if they’re worth anything until they actually are

Don’t over-extend yourself to exercise [options] in scenarios which put your financial well-being at risk.

If you’re obviously, unquestionably on a rocket ship: by all means, buy the options. (Yes, sometimes it really is obvious.)

If it’s not clear that you’re on a rocket ship, but you’re feeling good about the startup, and you can definitely afford to spend the money it would take to exercise your options: knock yourself out. Honestly, I don’t really care what people with wealth do in this scenario. My worries do not relate to you.

If it’s not clear that you’re on a rocket ship and spending the money to exercise your options would be a stretch: I would suggest you think twice before doing so. I also would warn you to never take out debt (which many startup employees do!) in order to exercise your options.

And that’s really the crux of my argument.

Startup employees without significant independent spending power who work for a venture with an uncertain future and who did not join their ventures at a very early stage — which I would argue describes most startup employees — should think long and hard before exercising their options.

It’s more than a little bit unfair that the people who can most easily realize upside from the startups they work for are people who already have wealth. Granting the ability for employees to buy shares directly at their fair market value is limited, too: this would make them investors, who the SEC says mostly need to be accredited. The definition of accreditation is either being a licensed investor, earning over $200,000 a year for the last two years, or having a net worth of over a million dollars excluding the value of their home. So the door is effectively closed to people from regular backgrounds.

I wish more equitable systems were commonly in use. Some different tactics are in use, which include:

• Restricted Stock Units. Here, stock is granted directly as part of an employee’s compensation. Upside: the employee has the shares. Downside: they’re taxed on them as soon as they vest, and selling them is restricted. So the employee effectively receives an additional tax bill with no way of recouping the lost funds until much later (if they’re lucky). RSUs are common in later-stage companies but very uncommon in riskier, early-stage companies for this reason.
• Phantom stock. Really this is a bonus plan tied to stock performance, income tax and all.
• Profit sharing. Which is only useful if the startup makes a profit (most don’t).

While some have value in their own right in particular contexts, I see them as compensation strategies that might sit alongside stock options, rather than replacing them.

I would love it to be less risky for the employees who are actually doing the work of making a startup valuable to see more of the upside of that work. But, at least for now, my advice remains to take those inflated Silicon Valley salaries and bank them in more traditional investments.

It’s been a little over a month since I launched ShareOpenly, my simple tool that lets you add a “share to social media” button to your website which is compatible with the fediverse, Bluesky, Threads, and all of today’s crop of social media sites.

You might recall that I built it in order to help people move away from their “share to Twitter” buttons that they’ve been hosting for years. Those buttons made sense from 2006-2022 — but not so much in a world where engagement on Twitter/X is falling, and a new world of social media is emerging.

People have been using it, and I’ve had lots of great feedback.

So, today, I’m pleased to announce releases for two of the biggest requests people have made for the tool.

A share icon

A share button needs an icon. That was clear from the very beginning. It needs to be something distinctive — this is a different kind of social media share tool — but also immediately recognizable as a share icon.

I reached out to one of the best designers in the field: Jon Hicks, whose excellent work includes the new Thunderbird logo, Disney’s SpellStruck, Spotify’s icon set, and Truck, an excellent record store in my hometown. I was delighted when he agreed to create a share icon for ShareOpenly.

This icon works really well at small and large sizes: in sidebars, in footers, and wherever you need to help people share. Click the version embedded here to share this very post:

A WordPress plugin

Lots of people have asked me for an easy way to embed a ShareOpenly link into WordPress.

David Artiss, a support lead at Automattic’s excellent WordPress VIP service, has written a WordPress plugin that is now available in the official WordPress plugin directory. He writes more about it in an announcement blog post on his site:

Simply download the plugin, activate it and you’ll find a link added to the bottom of every WordPress post or page. A simple settings page allows you to change the sharing text, as well as whether it appears on posts and/or page content.

Boom! It couldn’t be easier.

I really hope that the new icon and the WordPress plugin make it easier to include more open sharing to your website. ShareOpenly is suitable for everything from small blogs to large publishers.

Manually creating a share link

Of course, you don’t need to use the WordPress plugin. You can embed a share icon onto any web page using this code:

<a href="#" id="shareopenly"><img src="https://shareopenly.org/images/logo.svg" alt="Share to social media"></a>
<script>
  document.querySelector('#shareopenly').addEventListener('click', (e) => {
    e.preventDefault();
    let href = 'https://' + 'shareopenly' + '.org/share/?url=';
    href += `${encodeURIComponent(window.location.href)}&text=${encodeURIComponent(document.title)}`;
    window.location.href = href;
  });
</script>

Or you can construct the URL yourself by following the instructions on this page.

Have fun, and please keep the feedback coming! You can always email me at ben@werd.io.

When our son reached nine months old, I published this UsesThis-style list of products and services we were using; I had previously published one before he was born.

He’s twenty months old now: walking around and using words. He straddles the worlds between being a baby and a little boy, but becomes more like the latter every day.

I thought this was a good time to update the list.

Hardware

Stroller: Our Uppababy Cruz V2 is still holding strong, although we use it less and less. We’re in the lucky position of being able to walk to daycare, and he quite often just wants to walk alongside us. (The stroller comes in most handy when we’re late.)

Travel Stroller: New to our mix is the Joolz Aer (now superseded in their lineup by the Aer+). It’s light as hell and folds up to easily fit in an airplane’s overhead bin, which has been useful for both short and long-haul trips. When we’re not on the move, it permanently lives in the trunk of my car.

Car Seat: We still use the Clek Foonf, which is like the tank of car seats. It’s solid and he seems to find it comfortable, although it’s beginning to be an effort to get him in and out of it. That’s more of a function of my tiny car than the seat itself; it’s more likely that I’ll upgrade the latter than the former.

Travel Car Seat: We’ve added a Cosco Scenera car seat to our mix. It’s light and cheap, and is really straightforward to install into a new car (for example a rental) or into an airplane seat. The thing we haven’t cracked yet is traveling on a plane without this kind of car seat: it’s bulky enough that it would make one-parent flights with him next to impossible. (If you’ve made this transition, please tell me how. I’m considering a CARES airplane safety harness, but how safe is, it, really?)

Bed: We’re in the final months or weeks of the Ikea Sundvik crib and Naturepedic Classic Organic Cotton Crib Mattress, which have served us incredibly well. If he wasn’t using a sleep sack, he would have easily scaled its walls and climbed out bed already. It’s time for a Real Bed (and therefore time for a Real Bedroom). Model TBC.

White Noise: We’re still using the Hatch Rest, although I’m not fully sure why. Habit at this point? He does love boffing the top of it to get it to change color and sound.

Baby Monitor: We’re still on the Nanit Pro. It’s been fairly good for us, but it definitely has shortcomings. The humidity sensor almost never works, and the device itself will fairly often disconnect from the wifi. He’s also been known to shake it like a wizard’s staff. We’re not measuring how long he’s been awake, etc, anymore, so we don’t use the recordings as much as we used to. But it’s nice to be able to get a notification when he’s stirring.

Changing Mat: Any surface will do at this point. Towels, standing up in a chair with a rubber mat to protect it — whatever. About one out of twenty times he’ll end up on the Peanut, but it was much more useful when he was a little potato.

Kitchen Stool: The Cosco Kitchen Stepper has become a mainstay: he can stand on a protected tower while watching us cook, or helping out himself.

High Chair: The Stokke Tripp Trapp has grown with us. We’ve stopped using the built-in tray, which has made it a little easier to clean: instead, we push him in to the table, and he eats alongside us. Having everyone be first-class participants around the dinner table has been lovely.

Travel Booster: We’ve been known to bring the Fisher-Price booster seat with us (it’s in the back of my car right now), although it’s happened less and less lately. Most restaurants we go to — when we eat out, which is rarely — have high chairs, and I suspect future friends and family visits will be marked by him eating on one of our laps.

Food: He’s eating what we eat more often than not. We keep Annie’s Mac and Cheese stocked just in case, and Trader Joe cottage cheese and apple sauce have become staples. But we’re mostly off the food pouches, powders, and specialized food products.

Toys

This gets its own section at this point. There are a few real standouts:

‌Uppstå toddler walker: I can’t overstate how much he loves this thing. It took him a little while to get used to it, but it’s grown with him. He’ll put his favorite things in it and walk them around the house, sometimes at speed. Just incredible value for money based on how often he plays with it.

‌Fisher-Price Little People Airplane and School Bus: I’m almost certain these didn’t play songs and noises when I was a kid. They certainly do now, but it’s easy to switch off the sounds if you don’t want them. He plays with both continuously. It’s sort of weird to see how Fisher-Price toys have evolved — I kind of miss the peg-like figures they used to come with — but he’ll have the same nostalgic feelings for these that I have for mine. Just, um, look out underfoot.

‌Plantoys Wonky Fruits & Vegetables and Assorted Vegetables: These are simple enough — wooden fruits and vegetables joined with little velcro pads — but he loves taking them apart and putting them back together again. They come with a little wooden “knife” that makes it easier to split them.

‌Melissa & Doug Pull-Back Vehicles: He thought they were cool when he was much younger (they were a much-appreciated first birthday present), but these days he knows how to pull them back and make them go. We find them all over the house.

‌Busy Bee: Apparently the best-selling toddler toy in Aotearoa New Zealand. He loves pulling it along behind him, to the extent that it can usually be used as a distraction from just about anything else.

Books

Many, many times a day, he’ll demand to be read to. Seems like a good thing to me. Our collection of books is always rotating, but I thought it would be fun to list our stalwarts.

‌Hooray for Birds!, by Lucy Cousins: I could recite this verbatim. At this point, so can he, so I’ll often stop to let him fill in the blanks with the various bird noises. It’s lyrical, offers lots of opportunities to the reader to make silly voices, and is lots of fun.

‌Little Blue Truck, by Alice Schertle and Jill McElmurry: Rhythmic and perfectly wordsmithed; almost a song. You’ll find that you quickly get your engine noises, animal sounds, and “beep beep beep!”s to a professional level. Like Hooray for Birds!, he’ll often now come in with the right noises at the right time.

‌Each Peach Pear Plum, by Janet and Allan Ahlberg: I loved this book as a kid, and now he does, too. He calls it “Each Peach” and asks for it by name. (“Each Peach! Each Peach!”)

‌Guess How Much I Love You, by Sam McBratney and Anita Jeram: I like, half-jokingly, to call this the classic tale of toxic one-upmanship. But it’s sweet, and easy to give heart and voice to. I love him to the moon and back, too.

‌Jamberry, by Bruce Degen: This book makes no sense! It’s like a commercial for berries! But that doesn’t matter - he loves it.

I’m looking forward to introducing him to books for slightly older children that we have waiting in the wings; in particular, The Story of Ferdinand, Make Way for Ducklings, and Dogger, which are my top-three favorite picture books of all time.

Screen Time

Yes, we’re a Ms Rachel household. Lately, he’s been asking for my old pal Grover by name. And there’s a channel on YouTube called Helper Cars — origin unknown — that seems to be designed to appeal to his psyche.

We’re not completely comfortable with screen time, but it’s definitely become a thing. I figure it’s fine in moderation, and all of the above have some educational content to them.

Software

I’m retiring this portion of the baby list: we don’t use any software to track or help parent him anymore, save for the ubiquitous MyChart from the pediatrician and the Procare daycare-management app that lets us know when he’s eaten and sends us photos of his day from time to time. We just, you know, listen to him, and try to figure out what he needs. We’ll look up recipes and activities from time to time, but there’s nothing special about that; we do it for ourselves, too, and you all know how Google works.

Photo by Daiga Ellaby on Unsplash

I’m paralyzed by the world. We seem to be at a kind of crossroads.

There’s so much to be appalled by, so much to be worried about, and I worry that not saying something might be considered to be acquiescence or approval.

So, in this moment, I thought I’d actually take a step back and explain what my worldview actually is. It’s perhaps overly ambitious, but I want to declare I think is important, and what drives me to say the sorts of things I do. And, yes, these same factors also drive the decisions I make about where I work and how I build software.

As always, I would like to read yours.

My view on the world — as is true of yours, and of everybody’s — is a function of my lived experiences, and the lived experiences of the people I care about.

It’s not about leading the world; it’s about living in a peaceful one

I was born in the Netherlands, grew up in England, spent time in Austria, went to school in Scotland, and have been in the United States since my early thirties. My dad is Swiss-Dutch-Indonesian; my mother was American on one side and Ukrainian on the other. They met in Berkeley in the early seventies and were heavily active in various progressive causes as activists. My dad in particular, who was drafted into the US Army after his family moved to America as a teenager, organized Vietnam War protests in the Bay Area, and was often harassed by the police.

My parents intentionally left the US to raise me. The closest thing I have to a hometown is Oxford, famous for its universities, where academic families are constantly coming through. My peers at school came from all over the world — including from behind the iron curtain — and experiencing the different smells and tastes of peoples’ homes was completely normal.

At the same time, I was a third culture kid for almost all of my childhood and early adulthood, and identified with no national identity. I never felt any real ties to any particular geographic place for its own sake. I was raised an atheist and have never felt religious ties (even though, as a child in England, I went to a Church of England school where we were made to pray every day). What I did identify with, very strongly, was family.

My dad is one of the youngest concentration camp survivors: he spent his early years in a Japanese-run camp in Indonesia, which still colors the way he sees the world. His mother had nightmares every single night for the rest of her life; I will remember hearing her screams though the walls forever. His father, who I never got to meet, was a resistance leader who was forced to dig his own grave multiple times. I’ve heard stories about the camps and what happened afterwards for my entire life. Even when they returned to the Netherlands, Indo people like my dad’s family were an ethnic minority and treated poorly. They eventually moved to California, thanks to a local sponsor, where they ran a gas station on highway 12 in Sebastopol. My uncle was severely beaten up for daring to serve Black people. The gas station itself was routinely shot at — once killing the family dog — simply because it was run by immigrants.

My great grandfather escaped Ukraine twice. His village was destroyed by the White Army as part of vicious Pogroms. When he emigrated to the US, he secularized, changing his name in the process in order to sound less Jewish. Eventually he became the General Manager of the Pennsylvania Joint Board of the Amalgamated Shirt Workers. I’ve previously posted excerpts from my grandfather’s obituary that discuss that experience as well as my grandfather’s experience as a Jewish POW in Germany during WWII.

My grandfather, by the way, ended up translating Crime and Punishment into English, and taught in the Slavic and Eurasian Studies and History departments at the University of Texas at Austin for forty years. He met Albert Einstein, had coffee with Sylvia Plath, discussed philosophy with Hannah Arendt, and never quite realized his dream of being a poet. In the end, he married into an institutional American family: my great grandfather was a WWI test pilot and eventually became a diplomat who negotiated the US withdrawal from Haiti. (This fact of my family history is, I want to be clear, not an endorsement of the US’s behavior overseas, including in Haiti.)

So, all of this is to say: I have no interest in patriotism, let alone nationalism. It’s not a value I hold, and I’m not excited by any country having a leadership position in the world. I find flag-waving to be petty. What I care about are values: the democracy, inclusion, and co-operation that can lead to a lasting peace. I’m repelled by military strength, because I’ve seen what various militaries did to my family. I’m repelled by anti-immigrant sentiment, because I come from refugees and immigrants. I don’t like the idea of assimilation, because I’ve seen the richness inherent in lots of cultures. Forced assimilation — which is usually into a conquering culture — is tantamount to subjugation.

National exceptionalism — American exceptionalism, or European exceptionalism, come to that — is ridiculous on its face. Cold wars and imperialist foreign policies are things to avoid, not things to perpetuate. No country is the “best”, and even the idea of “a best country” is narrow-minded. No religion is the “best”; please enjoy practicing yours, but please don’t impose it on anyone else. There are definitely people who think McCarthy’s witch hunt against communism was right in spirit, even if they condemn the historical event itself — let’s just say they and I harbor some very different ideas about what an open, democratic society should look like.

Nations aren’t what’s important. Principles are. Specifically, the principles of openness, inclusion, fairness, peace, equity, and democracy.

What matters is that everyone can live a good life, wherever they are, whoever they are, and however they identify, free from threat of violence or exploitation. Ideological or national superiority aren’t useful values. What matters is the experience of being a human, everywhere. What matters is avoiding the killing and horror of war. What matters is honoring the beautiful diversity of the world.

A strong operating system for all

I was only half-joking when I compared governments to operating systems. While they certainly don’t map perfectly to actual software operating systems, I do think they provide a very similar purpose: to create a bedrock of services and infrastructure in order to ensure society runs smoothly.

What does “society runs smoothly” really mean? I’ll return to my definition above. What matters is that everyone can live a good life, wherever they are, whoever they are, and however they identify, free from threat of violence or exploitation. Freedom of expression, association, and to pursue one’s best interests are important here: what John Locke called the pursuit of happiness. To ensure the safety of that pursuit, I think John Locke’s version of a social contract — the idea that we surrender a little personal liberty in order to make evolving common agreements in the best interests of everyone — is important.

I can’t be a libertarian because I see the importance of the trade-offs here. One role of the operating system is to prevent the vulnerable from exploitation: public goods like universal healthcare, public education, and integrated public transit ensure that people who are not wealthy have the opportunity to build a great life. One of the most visceral reactions I’ve ever had in my life was discovering Ayn Rand, and then, to my horror, discovering that beyond just getting into her novels, some people actually believed in her ideology of everyone for themselves.

Healthy communities are an important part of all of our well-being. Once again: every person deserves to live a good life. We all live in a complex, interconnected network of people, and what happens to someone else also affects us. Caring for the whole network is also in our own best interests. It can’t be everyone for themselves.

It’s hopefully obvious from my definition, but I don’t think GDP (or money at all) is a great way to measure a society, either. It doesn’t say much about what an ordinary person’s experience actually is. It doesn’t measure human well-being, and that’s how we should be thinking about how well we’re doing. I’m less interested in is the stock market going up? than is being poor a death sentence? as a question — and I don’t think the first necessarily leads to a reduction in the second. More and more people agree.

One important function of the social operating system is welfare, which ensures that people don’t fall though the cracks. There are others, some of which I’ve already mentioned: education, transit, and healthcare.

I couldn’t have founded my first startup if I hadn’t had the benefit of the excellent National Health Service. Millions of PR dollars have been spent in the US to paint social infrastructure as being a bad thing, but I never once had to worry about going to the doctor under universal healthcare. I didn’t have to worry about losing health insurance when I quit my job. I could just do it. Say what you want about free markets, but I think that freedom of optionality — having broad choices regardless of income or personal net worth — comes closer to real personal freedom than a world without that kind of social infrastructure.

Here in the United States, it doesn’t come automatically: you need people to fight for you. I’ve lost five members of my family to an incurable genetic disease. One of them was my mother, who I helped care for over the course of a decade — which was, in fact, the reason I moved to the US to begin with. She was a teacher, and the great medical care she received was only possible because of the incredible negotiating power of her teacher’s union. While there should have simply been universal healthcare to look after her, their incredible negotiating power literally lengthened her life by eight years. Unions can be amazing institutions; while not every union is great, the concept of them is. And in a world without the social infrastructure to care for the vulnerable, they are vital.

This should be the purpose of the law, too: to prevent harm and exploitation, particularly of the vulnerable, in service of maintaining the ability to have a good quality of life. But the law itself, alongside tradition and the twin ideas of unity and stability, often does the opposite. It has often used as a way to maintain a status quo where vulnerable people are exploited for other peoples’ benefit.

This runs deep: some of the earliest police forces in America were slave patrols. A law that only benefits the powerful or upholds an unjust status quo is, in itself, unjust. Unity that depends on adherence to the values of the powerful (and on the silence or silencing of the vulnerable) is a sham. Stability based on prioritizing the needs of an in-group to the exclusion of others is definitionally fascism. A claim that moderate values are more reasonable only makes sense to people who don’t need more radical change in order to achieve equity.

Being awake to those injustices is not a binary: it’s not something you either are or not. It’s an ongoing, uncomfortable process of education and coming to terms. There are lots of ways to deal with and redress them, the comparative merits of which are up for discussion. What’s clear to me, though, is that dismissing their existence outright, and painting them with a reductionist brush in order to rob them of importance, is in itself a perpetuation of those injustices.

When people started talking about being “woke” and taking to the streets to demand restorative justice, I was relieved and excited. This is what moving forward looks like. In contrast, I see people harping on about the harms of “woke-ism” as being part of the dying gasps of the twentieth century: that adherence to tradition and unity and stability in service of the same old inequalities. Change is good; particularly here.

Change can also be easy. Using a preferred pronoun costs you nothing except for letting go of a tradition. The tradition, in other words, gets in the way of someone’s chosen identity being recognized. Same-sex marriage costs you nothing except for letting go of a tradition. The tradition once again gets in the way of someone being able to realize their needs. Your religious beliefs might forbid same-sex marriage; then simply don’t get same-sex married. Practice your own religion to your heart’s content, but don’t enforce your traditions on anyone else. Refer to someone as they would like to be referred; treat everybody with respect. It seems foundational. A fear of change or adherence to a tradition should not be a barrier to making the world more just or treating our fellow humans with respect.

Let’s return again to the core idea: every person deserves to live a good life. Of course, who we consider to be a “person” is important. Thomas Jefferson incorporated Locke’s version of a social contract into the Declaration of Independence, even going so far as to say that “all men are created equal,” but he famously kept slaves. These days, we might ask about our spheres of concern: do we care about people in our families? Our neighborhoods? Our towns and cities? Our churches? Our ethnicities? Our value structures? Our states? Our countries? Our regions? The world? How do we relate to people outside of those spheres?

For reasons I’ve tried to explain above, I’d love it if we considered the world to be what we care about; the welfare of a person in Gaza is just as important as the welfare of a person next door, even if we might not share a religion or care for the regime they live under. In fact, depending on their context, their welfare might be more important, because they need more help to bring them to that reasonable standard of living, free from violence and exploitation.

Because our definitions of what a good life is vary, and because no government can possibly claim to represent or understand the complete set of needs and lived experiences in its populace, participative democracy is the only equitable model for government. What’s important here: everyone can vote without hindrance, votes are fair, anyone can become a representative, decisions are actually made at the ballot box rather than in court, and there is real choice. (If there’s a candidate whose values you hate, do what you can to persuade your fellow voters to vote for someone else. That’s what democracy is.)

Those principles are core. It might surprise you to learn that I’m not inherently against the idea of billionaires, and certainly not against the idea of starting businesses and finding success in doing so. But it must be done without exploiting other people and preventing them from being able to live a good life. It must be done without perpetuating injustices, for example by eroding workers’ rights, forcing a minimum wage that is too low to live well on, lobbying for unequal laws, or fighting against their ability to negotiate for better working conditions. Can it be done without those things? I don’t know. But if it can’t, then it shouldn’t happen.

So what does this have to do with the internet?

I see the web as a platform as being rooted in the kind of internationalism I believe in. The internet itself is a physical manifestation of the idea that we are all connected.

Anyone can publish, anywhere, and be read by anyone, anywhere. That’s amazing! Anyone can start a business and find users all over the world. That’s also amazing! It’s the most borderless, open platform we’ve ever created. The potential to learn about the lives of people we would never otherwise meet, in places we would never otherwise visit, is colossal. We can share ideas and, even more importantly, build empathy globally. I couldn’t be more excited about that. That’s what keeps me building.

It’s easier to dehumanize someone you don’t know. The internet has the potential to allow everybody to become knowable. I see that as a route to peace, and to a better world where exploitation can no longer happen in the shadows.

What I’m not enthused by is the idea that the internet is here as an exercise in furthering any one country’s interests: that one nation’s worldview should trump another’s. At its best, it’s an international commons: an overtly progressive space by design.

I support the indieweb and the fediverse because those technologies harness for the benefit of the public, rather than for the profit and entrenched power of a tiny few. I see silos and centralized services as being anti-democracy, because the whims of a monarch-like figure can have a profound impact on which information we’re allowed to see. We’ve seen that most obviously recently with Elon Musk’s acquisition of Twitter, but it was previously also true of Facebook, and of every large service that aimed to intermediate peoples’ connections with each other. If one entity controls what we see and can learn about, they will abuse it, always.

I’m imperfect. Of course I am. I’ve made terrible mistakes and, from time to time, I’ve hurt people. But that doesn’t mean I can’t try.

I’ve built open source platforms for organizing educational institutions and non-profits; I’ve supported newsrooms that help to create a more informed voting population; I’ve worked in newsrooms that help speak truth to power. It’s not because I love social networking in itself, or because I want to get rich by building software.

It’s because I remember the sound of my Oma having nightmares through the walls. I see the nationalists and isolationists as trying to divide people into in-groups and out-groups. I see hoarding wealth as akin to building walls. I see conservatism as being a way to preserve the kind of bigotry that can grow and explode into the kinds of hatred that swallow whole families. Quixotic as it might be, I see connecting people as a way to help prevent it all from ever happening again.

The requirement for TikTok to relinquish Chinese ownership or face a nationwide ban was signed into law today, as an add-on to a foreign aid bill:

But even as ridiculous as it is to tack on a TikTok ban to foreign spending support, Biden had made it clear he supported the TikTok ban anyway. Still, it does seem notable that, when signing the bill, Biden didn’t even mention the TikTok ban in his remarks.

This is a couple of years after Meta, clearly threatened by the app’s enormous growth, started employing a PR firm to kick off an anti-TikTok campaign.

There are a few worries at play:

• TikTok will irresponsibly collect enormous amounts of data about hundreds of millions of Americans, something no other social network would ever do
• There’s a possibility that TikTok will be used to spread propaganda, unlike every other social network
• TikTok is Chinese, not American

The third is most pertinent. These are clearly things — gathering data, spreading propaganda — that only American companies should do to Americans (or to the rest of the world).

Ironically, banning a service from the open internet nationwide is exactly the kind of thing that China has done again and again through its Great Firewall. Rather than protect American users through the kinds of far-reaching privacy legislation that we need, government chose to address TikTok alone on the basis of what amounts to xenophobic protectionism.

It’s the kind of xenophobia we saw at a Senate hearing on child safety earlier in the year:

“You often say that you live in Singapore,” Cotton said before demanding to know where Chew’s passport was from (Singapore, obviously) and whether he’d applied for citizenship in China or the US (no, said Chew). “Have you ever been a member of the Chinese Communist Party?” he then asked abruptly, as if hoping to catch Chew by surprise. Chew’s response wasn’t shocked so much as fed up. “Senator! I’m Singaporean!” he reiterated. “No.” (Singapore is not part of China.)

The Verge further made this point:

It’s not even necessary to make the case that China might have undue influence over TikTok. Apple, for instance, has weathered years of critiques about its relationship to the Chinese government; no reasonable person has ever suggested this hinges on Tim Cook being a secret communist. Instead, it’s a line of questioning that seems simply designed to play on Chew’s foreignness — even when it’s got nothing to do with the topic at hand.

It’s not that TikTok is particularly harmful compared to other similar apps: it’s that we’re deathly afraid of China.

I find it unsettling that a global platform - the internet, that is - which seeks to connect everyone in the world is being undermined for Americans in this way. Should this precedent spiral, it’s not unreasonable to think that more foreign services that threaten American incumbents will be banned or forced to divest. The result would be a National Internet, culturally and economically cut off from the rest of the world: something so dystopian that I’ve written it into science fiction stories.

Beyond nationalism, the propaganda argument may ironically cover for the fact that viewers are likely to encounter a wider range of international viewpoints:

Anti-China hysteria may be completely grounded in xenophobia, but some legislators clearly also don’t care for the content of TikTok. Raja Krishnamoorthi, an Illinois Democrat supportive of the anti-TikTok legislation, recently raised concerns over TikTok’s content recommendations. TikTok has been criticized for presenting its users with “pro-Palestinian” content. It has been a major source of videos of the unfolding atrocities in Gaza, and a ban of TikTok could, as the Independent’s Io Dodds reports, “clobber the pro-Palestine movement,” supporters of which have used the platform effectively for communication.

Current Affairs concludes:

The disconnect between the American people’s interests and the priorities of national politicians has never been more stark. We need to resist their attempt to get us to be afraid of Chinese people, and to control platforms where viewpoints unfavorable to U.S. imperialism are given a public airing.

This in itself feels un-American: a violation of the democratic rights enshrined in the First Amendment and the principle of free speech. We have the right to learn about those viewpoints, and to receive media created elsewhere in the world. Just as on the web itself, which allows us to learn about the world from a variety of perspectives, we’re richer for it.

I imagine there will be a battery of lawsuits from the international investors who actually own TikTok on the back of this legislation. It’ll be interesting to see what happens, and what the response from other nations will be.

Personally, I just think it’s stupid.

It’s my mother’s birthday. She would be 72 today.

The week we lost her, I wrote this piece, which I re-read today.

In it, our friend Anita Hurrell remembered her like this:

One time you drove us in the van to the seaside and we ate sandwiches with cucumber in them and I thought they tasted delicious and I felt this strong sense of deep content sitting with Hannah in the back listening to her singing and humming for the whole journey. I have no idea where we went, and in my head it was nowhere in England, but rather part of the big-hearted, loving, funny, relaxed, non-conformist world of your family in my childhood - full of your laughter and your enormous kindness.

[…] I look back and see […] what a true feminist you were, how much of parenting you seemed to do much better than we do these days, how generous and homemade and fun and kind the world you and Oscar [my dad] made was.

I feel so privileged to have had that childhood. To have had a mother like her.

In the piece I read at her memorial, I said:

Before I was born, both my parents were involved in struggles to support affirmative action and tenants’ rights. She described herself as having been radicalized early on, but it’s not particularly that she was radical: she could just see past the social templates that everyone is expected to adhere to, and which perpetuate systemic injustices, and could see how everything should operate to be fairer.

That was true on every level. She wanted she and [her siblings] to all be treated equally, and would make it known if she thought the others were getting a raw deal. She tried her best to treat Hannah and I equally. If someone made a sexist or a homophobic remark around her, she would call it out. If someone was xenophobic, or unthinkingly imperialist, she would bring it up. She was outspoken - always with good humor, but always adamant about what really mattered.

When our son was born, I wrote:

The last time I saw you, just over a year ago, you were in a bed in the same institution, your donated lungs breathing fainter and fainter. I kissed you on the forehead and told you I loved you. You’d told me that what you wanted to hear was us talking amongst ourselves; to know that we’d continue without you. In the end, that’s what happened. But I miss you terribly: I feel the grief of losing you every day, and never more than when my child was born.

[…] In this worse universe that doesn’t have you in it, I’ve been intentionally trying to channel you. I’ve been trying to imagine how you would have shown up with them, and what your advice for me would have been. I’ve been trying to convey that good-humored warmth I always felt. You made me feel safe: physically, yes, but more than that, emotionally. I want to make them feel safe, too: to be who they really are.

That first piece from three days after we lost her:

I want to honor her by furthering what she put into the world. The loving, non-conformist, irreverent, equity-minded spirit that she embodied.

Her values are the right ones, I’m sure of it — her non-conformity, her progressivism, her intellectual curiosity, her fearlessness and silliness (and fearless silliness), her gameness for adventure, her internationalism, her inclusive care and love for everybody and absolute disregard for the expectations other people had for her, or for nonsense tradition, for institutions, or for money.

She is the best person I’ve ever met and could ever hope to meet. I’ve been so, deeply sad, every day, but grief isn’t enough to honor her, and I wish I’d been better at doing that.

I miss you, Ma. I love you. I’m so, so sorry.

You know all those “share to Facebook” / “share to Twitter” links you see all over peoples’ websites? They’re all out of date.

Social media has evolved over the last year, yet nobody has “share to” links for Mastodon, Bluesky, Threads, etc. There have been a few attempts to create “share to Mastodon” buttons, but they haven’t taken the larger breadth of the new social media landscape into account.

So I’ve built a prototype, which I’ve called ShareOpenly.

At the bottom of every article on my site, you’ll see a “share to social media” button. Here’s the button for this article.

If you click it, you’ll be taken to a page that looks like this one:

You can select one of the pre-set sites in the list, and you’ll be taken to share a post there. For example, if I click on Threads, it will take me to share there:

But if you, for example, have a Mastodon instance, or a Known site, or an indieweb site at a different domain, you can enter that domain in the box, and ShareOpenly will try and find a way to let you share the page with that site.

ShareOpenly will do a few things first:

1. If it’s on a “well-known” domain — eg, facebook.com — it’ll send you to the share page there.
2. It checks to see if it can figure out if the site is on a known platform (currently Mastodon, Known, hosted WordPress, micro.blog, and a few others). If so — hooray! — it knows the share URL, and off you go.
3. It looks for a <link rel=“share-url”> header tag on the page. The href attribute should be set to the share URL for the site, with template variables {text} and (optionally) {url} present where the share text and URL should go. (If {url} is not present, the URL to share will be appended at the end of the text.) If it’s there — yay! — we forward there, replacing {text} and {url} as appropriate.

Once you’ve shared to a site, the next time you visit ShareOpenly, it will be in the quick links. For example, I shared to my site at werd.io in the example above, and now here it is in the links:

It’s early days yet — this is just a prototype — but I thought I’d share what I’ve built so far.

If you want to add ShareOpenly to your own site, please do! Just replace the URL and test in this link - https://shareopenly.org/share/?url=url&text=text - with your own. You can also just visit the ShareOpenly homepage to share a site directly.

Syndicated to Indienews

I’ve spent most of my career — now well over two decades of it — building things on the web. I’ve worked as a software developer, I’ve founded a couple of my own companies, and I’ve often found myself leading teams of engineers. Right now I’m the director for both engineering and IT (although there are teams of people who write code who aren’t under my wing — newsrooms are complicated).

Over time, a lot of my work seems to have become less about “what shall we build?” or “how shall we build it?”. Those questions are always vitally important, but there are prerequisites that sometimes need to take center stage: things like “what are we here to do?”, “how should we work together?”, and “how do we think about what matters?”

I’ve been sharpening my thinking about the necessary conditions to do good work, and how to achieve them. Here’s a window into how I’m thinking about these ideas across three dimensions: Organizational Context, Team Leadership, and Technology Trends.

Organizational Context

There’s an unattributed but often-quoted management strategy cliché that says: culture eats strategy for breakfast. I’m a believer. Culture contains the fundamental building blocks of how an organization acts as a community: its values, beliefs, attitudes, norms, processes, and rules. Without a strong one, you cannot succeed — regardless of what your strategy looks like. Conversely, a great strategy, by definition, is one that incorporates building a great, intentional culture. Without one, your team is more likely to burn out and leave, you’re much less likely to build something high-quality, and you’re unlikely to foster new ideas.

Software engineering, at least in the places I’ve practiced it, is all about innovation. The focus is rarely on maintaining the present, although some degree of maintenance is always necessary. Instead, it’s about building the future: figuring out what we’ll need our platform to look like in two to five years and finding ways to get there. It’s a creative pursuit as much as it is about rigor and craft, and it’s about values and taste as much as it is about business necessity.

This dynamic is well-served by some organizational cultures and actively undermined by others. The trick is figuring out which you’re in, and finding ways to either embrace the former or build a buffer zone for the latter.

One popular way of looking at organizational culture is the Competing Values Framework, which defines four distinct overall culture types — all four of which are usually present to different degrees inside an organization.

• Adhocracy: an organic, unbureaucratic way of organizing work that challenges the status quo, formal titles, and hierarchy in favor of a focus on risk-taking and innovating at speed.
• Clan: a family-like culture that, again, is relatively unbureaucratic, without much structure, where rules tend to manifest as social norms rather than edicts or rigid process.
• Hierarchy: where an emphasis is placed on top-down control from upper management in order to create predictability and lower risk. Roles are clearly defined, rules are codified, and even internal communication tends to be stratified.
• Market: a culture optimized around competition, both with competitors and internally. Measurable results are central, but the workplace can easily become toxic because everyone is trying to better themselves vs their peers.

Like many frameworks, the reality is not actually as cut and dry as this. Instead, I think these categories are best thought of as facets of an organizational culture. In some organizations, hierarchy and market focus may have a heavier emphasis; in others, innovation and collaboration.

I vastly prefer working within organizations that look like the first two environments — adhocracies and clans — and I’d hazard to say that almost every single engineer, designer, and product manager I’ve ever worked with feels the same way. Hierarchical systems are inherently creatively stifling: innovation can’t take place in an environment with predominantly top-down control. The same goes for hyper-competitive environments: while the competition might be motivating for some in the short term, it’s really hard to collaborate effectively and build on each others’ ideas if everyone is trying to get ahead of each other.

Hierarchies in particular definitionally strip your authority in favor of top-down direction, forcing you to negotiate through layers of politics to make any kind of change. Most good engineers are collaborators, not cogs, with ideas, expertise, and creativity that should be embraced. But hierarchy demands cog-like behavior, and creates institutional fiefdoms that tend towards bureaucracy, inhibiting any really new work from being done if it hasn’t been rubber-stamped. These aren’t great places for a creative person to work.

As Robin Rendle put it recently:

This is the most obvious thing to say in the world, but: the hard work should never be the bureaucracy, it should be designing things and solving technical problems. If the hard work ain’t the hard work, ya gotta bounce. Don’t kill yourself trying to tell people that.

That isn’t to say that every team in an organization should work the same way or strive for the same culture. It might be that a legal, compliance, or safety team needs to work in a more rigid way as a system of control, or that a sales team needs to be intensely market-oriented. Or those things might not be true at all! My point is that it’s a mistake for engineers to assume that because they work best in a particular kind of environment, everyone should work that way. Every organization is comprised of a mix of culture types, and every team needs to work in a way that allows them to do their best work.

This may seem obvious, but we often talk about a single team’s working style setting the cultural norms for a whole organization. For example, it’s common for an organization to be described as engineering-led or sales-led. To be clear, this is a false choice: there should simply be people-led organizations that are inclusive of different interdisciplinary needs and styles of working.

For that to be a reality, top-level leadership in particular needs to acknowledge that not every team works the same way. For my purposes, this means acknowledging that engineering needs a particular kind of culture in order to thrive (and is important enough to have its own culture and be deserving of autonomy).

A prerequisite to this is understanding the potential for a technology team (or any team) in the first place. That’s less likely to happen in organizations where it’s treated as back-office, paint-by-numbers work. If an organization can’t see the importance of a team’s work, and if it inherently does not respect the effort and expertise inherent in those roles, it’s going to be very difficult for them to do good work.

My bias is to lean heavily on storytelling and listening as tools for fostering understanding: finding ways to explain why the work of product and engineering is important in the context of the whole organization, and how that expertise can be leveraged in order to benefit everybody. It’s okay to not understand what an engineering team has the potential to do from the outset, but if organizational leaders continue to not understand, that’s on me. The way to get there is through being transparent about what we’re doing, how we’re thinking, and which challenges we expect to encounter.

It really matters. Mutual understanding begets mutual respect.

There needs to be an explicit understanding between teams, mutual respect between parties that encompasses their expertise and different ways of working, and loose protocols for how everyone is going to communicate with each other that is compatible with their different styles of working.

I shouldn’t presume to tell a team from another discipline what they need to do their job, just as they shouldn’t presume to tell me. I should treat another team as the expert in its discipline, and they should treat my team as the expert in mine.

Throughout all this and despite our differences, we’re all in the same boat. We need to all be pulling in the same direction, motivated around a single, motivating mission (why we’re all here), vision (what is the world we’re here to try to create), and strategy (what are we going to try and do next to make it a reality).

The role of upper management is to set the direction, foster a culture that supports everyone, and help to build those protocols (all while not running out of money). One role of team leaders is to navigate those protocols and act as a buffer where there is friction.

Team Leadership

Vulnerable, open leaders make it safe for everyone to take risks and show up to work as they are.

So far I’ve written a lot about how engineering teams need organizational support that starts with a compatible culture that is founded on respect. But even in an environment that is un-hierarchical, transparent, informal, respectful, and open, with clear organizational goals and a defining mission, there’s more work to be done in order to create an environment where engineers can do their best work.

As I wrote last year:

The truth is that while some of the tools of the trade are drawn from math and discrete logic, software is fundamentally a people business, and the only way to succeed is to build teams based on great, collaborative communication, human empathy, true support, and mutual respect.

Leaders need to be stewards of those values. I believe — strongly — that this is best achieved through servant leadership:

[Servant leadership] aims to foster an inclusive environment that enables everyone in the organization to thrive as their authentic self. Whereas traditional leadership focuses on the success of the company or organization, servant leadership puts employees first to grow the organization through their commitment and engagement. When implemented correctly, servant leadership can help foster trust, accountability, growth, and inclusion in the workplace.

Each of these are important; I would also add safety. A blame-free environment where everyone can speak openly, be themselves, make mistakes, and not feel like they have to put on a mask to work is one where people can take risks and therefore innovate more effectively.

When you’re facilitating a brainstorming exercise, you might intentionally throw in a few out-there suggestions to make participants feel comfortable to take risks with their own contributions. Similarly, one of the roles of a leader is to push the envelope, and maybe risk looking a bit silly, in order to allow other people to feel more comfortable taking risks with their work — and when they do, to cheerlead them, support them, and help them feel comfortable even if their ideas don’t work out.

In a hierarchical team, the leader might ask if team members are adhering to their standards. In a supportive team, the leader might primarily ask how they are doing at supporting their team. It’s not that you don’t ever ask if someone isn’t performing; it’s more to do with the center of gravity of assessment. Supportive teams put the employees first.

Fostering that sort of team culture heavily depends on how a manager shows up day to day. A manager who isn’t vulnerable, doesn’t reveal much of themselves, and requires homogeneity is — probably unintentionally — fostering a hierarchical culture where masking is the norm rather than a supportive one where people are free to to be themselves.

The same sorts of fractal dynamics that affect inter-team collaboration apply to inter-personal collaboration, too. Everyone is different and has different working and communication styles, and homogeneity should never be the goal.

You can tell a lot by a team’s approach to feedback. If it is given in one direction — from managers down — then you likely have a hierarchical culture where team members may be less able to speak up and share their ideas. (The same is true if feedback is sometimes given to managers but rarely acted on.) I’ve observed that the most successful teams have clear, open, 360-degree feedback loops, where everyone’s feedback is directly sought out and incorporated — from team members to managers, between team members, and from manager to manager.

Another observable difference in team cultures can be seen through the kinds of norms that are enforced. To the extent that there are hard and fast rules on a team, they should be grounded in a purpose that supports forward motion, rather than to provide comfort to leadership or simply to enforce sameness.

As illustration, here are two contrasting examples of norms I’ve often seen enforced on teams:

• Source code is written to adhere to common style guidelines, and is peer reviewed.
• Cameras should be turned on during video calls.

Common coding style rules are a social contract that lower the cognitive load of working with code that someone else on the team has written, removing important roadblocks to everyone’s work; peer review is a really great channel for feedback, learning, and preventing bugs. Meanwhile, enforcing that cameras should always be on during video calls only serves to make some people less comfortable on the call.

Ultimately, success here is measured in what you ship, how happy your team is, whether they recommend working at your organization to their friends, and how long people stick around for.

Technology Trends

It’s important for an engineering team to not just have a competence in working with technology but to have strong opinions about it, its implications, and how it intersects with the lives of the people it touches. They should strive to be experts in those issues, learning as much as they can from relevant publications, scholars, and practitioners.

It would be ludicrous to examine the use of AI but not study its ethical issues. Not only is there a moral hazard in not understanding the subject holistically, but by leaving out topics like bias, intellectual property violations, and hallucinations as you investigate bringing AI into your work, you actually create liability for your organization. It’s both an ethical duty and good due diligence.

Similarly, imagine studying blockchain a few years ago but not covering its environmental impact or its potential for use in money laundering. Leadership might have been excited by the potential for financial growth, but by not examining the human impacts of the technology, you would have missed substantial risks that might have created real business headwinds later on.

Or imagine relying on developing code as a core function of your organization and not staying on top of new techniques, approaches, exploits, and technologies to build with. Your team would effectively be stuck in time without any real way to progress and stay relevant, creating a risk that your product would suffer over time.

Or, come to that, imagine working in a fast-moving field like technology and not forming a strong, informed opinion about how it will change that is rooted in learning, experimentation, and active collaboration with experts and other organizations.

This is another area where an open, collaborative, inclusive culture can be helpful. Giving space to team members who want to share their knowledge and ideas about a subject, and entrusting them to cover it from their perspective, helps allow for topics to be covered through the lens of a variety of diverse lived experiences. But by practicing and championing the idea of inclusion as a core team value, you encourage team members to actively go and speak to diverse experts and gather a variety of viewpoints. The gene pool of ideas is widened as you investigate a subject and your own ideas and resulting products and strategies will be stronger as a result.

In a hierarchical culture where strategy is set from the top down, this kind of broad, inclusive learning might not be as effective, or it might not be present at all. Servant leadership helps ensure that everyone has the space to learn and grow with respect to topics they may not have mastered yet, or that their perspectives are championed. You simply have access to fewer ideas from fewer perspectives, and you’re wildly limited as a result.

Those same open feedback channels that create well-functioning, communicative teams can also serve as a way for team members to learn from each other. The principles of openness, inclusion, respect, openness to risk, and collaboration can serve as guiding lights as teams navigate new technologies and help their respective organizations get to grips with these topics. Leaders have a role in fostering learning and knowledge-sharing on a team, and ensuring that it is a first-class activity alongside writing and architecting code.

Overall

A lot of the things that are important to get right with engineering aren’t really about engineering at all. The best teams have a robust, intentional culture that champions openness, inclusivity, and continuous learning — which requires a lot of relationship-building both internally and with the organization in which it sits. These teams can make progress on meaningful work, and make their members valued, heard, and empowered to contribute.

At a team leadership level, servant leadership is a vital part of fostering a culture of innovation and adaptability. By prioritizing the well-being and development of the people on their teams, leaders are making an investment that leads to higher performance, more nuanced strategy, more resilience, and lower churn.

At an organizational leadership level, a clear strategic direction and a focus on inclusivity help to provide the leeway to get this work done. I don’t know if you can succeed without those things; I certainly know that you can’t create a satisfying place for engineers and other creative people to work.

The most interesting and successful organizations have an externally-focused human mission and an internal focus on treating their humans well. That’s the only way to build technology well: to empower the people who are doing it, with a focus on empathy and inclusion, and a mission that galvanizes its community to work together. And, perhaps most importantly to me, that’s the only way to build a team that I want to work on.

That’s how I’ve been thinking about it. I’d love to read your reflections and to learn from you.

I’ve been thinking about the risks and ethical issues around AI in the following buckets:

• Source impacts: the ecosystem impact of generative models on the people who created the information they were trained on.
• Truth and bias: the tendency of generative models to give the appearance of objectivity and truthfulness despite their well-documented biases and tendency to hallucinate.
• Privacy and vendor trust: because the most-used AI models are provided as cloud services, users can end up sending copious amounts of sensitive information to service providers with unknown chain of custody or security stances.
• Legal fallout: if an organization adopts an AI service today, what are the implications for it if some of the suits in progress against OpenAI et al succeed?

At the same time, I’m hearing an increasing number of reports of AI being useful for various tasks, and I’ve been following Simon Willison’s exploratory work with interest.

My personal conclusions for the above buckets, such as they are, break down like this:

• Source impacts: AI will, undoubtedly, make it harder for lots of people across disciplines and industries to make a living. This is already in progress, and continues a trend that was started by the internet itself (ask a professional photographer).
• Truth and bias: There is no way to force an LLM to tell the truth or declare its bias, and attempts to build less-biased AI models have been controversial at best. Our best hope is probably well-curated source materials and, most of all, really great training and awareness for end-users. I also would never let generative AI produce content that saw the light of day outside of an organization (eg to write articles or to act as a support agent); it feels a bit safer as an internal tool that helps humans do their jobs.
• Privacy and vendor trust: I’m inclined to try and use models on local machines and cloud services that follow a well-documented and controllable trust model, particularly in an organizational context. There’s a whole set of trade-offs here, of course, and self-hosted servers are not necessarily safer. But I think the future of AI in sensitive contexts (which is most contexts) needs to be on-device or on home servers. That doesn’t mean it will be, but I do think that’s a safer approach.
• Legal fallout: I’m not a lawyer and I don’t know. Some but not all vendors have promised users legal indemnity. I assume that the cases will impact vendors more than downstream users — and maybe (hopefully?) change the way training material is obtained and structured to be more beneficial to authors — but I also don’t know that for sure. The answer feels like “wait and see”.

My biggest personal conclusion is, I don’t know! I’m trying not to be a blanket naysayer: I’ve been a natural early adopter my whole life, and I don’t plan to stop now. I recently wrote about how I’m using ChatGPT as a motivational writing partner. The older I get, the more problems I see with just about every technology, and I’d like to hold onto the excitement I felt about new tech when I was younger. On the other hand, the problems I see are really big problems, and ignoring those outright doesn’t feel useful either.

So it’s about taking a nimble but nuanced approach: pay attention to both the use cases and the issues around AI, keep looking at organizational needs, the kinds of organic “shadow IT” uses that are popping up as people need them, and figure out where a comfortable line is between ethics, privacy / legal needs, and utility.

At work, I’m going to need to determine an organizational stance on AI, jointly with various other stakeholders. That’s something that I’d like to share in public once we’re ready to roll it out. This post is very much not that — this space is always personal. But, as always, I wanted to share how I’m thinking about exploring.

I’d be curious to hear your thoughts.