Skip to main content

Can we build the dog?

I work at the intersection of technology, media, and democracy.

I've been a Director of Investments, a CEO, a CTO, and an engineer.
I co-founded Elgg and Known, worked on Medium and Latakoo, and invested in innovative media startups to support a stronger democracy at Matter.


A platform engineer's dirty secret: deleting users is hard

3 min read

There's rightly been a lot of discussion over the last few weeks about GDPR and its companion, the ePrivacy Directive. Internally, tech companies are scrambling: the architecture changes needed to support these changes need to affect every single user. Although they might provide related user-facing features only to people in the EU, the underlying data layers don't have meaningful differentiation between users from different countries, so the changes need to apply to everyone.

This is great news for proponents of individual privacy here in the US. I definitely count myself in that number.

One requirement making waves beneath the hood is the need for users to have their data completely deleted from a service. This isn't as easy as it might sound: a user's personal information typically isn't stored in one spot in a database, and isn't discrete from other users' information. Finding it and then ensuring it is removed without harming anyone else's experience is non-trivial in large systems, so perhaps understandably, most developers simply deactivate a user instead, leaving their data trail largely intact (but publicly inaccessible). That's not enough under this legislation.

The same may apply to files. Some years ago, researchers discovered that photos deleted from Facebook were lingering on their servers. It can be easier and cheaper just to remove access to a file than to actually physically remove it from disk. Content Delivery Networks also pose a problem: these are widely employed to optimize download speeds for content like photos and videos. This involves making copies of those files at "edge" locations that are geographically close to users around the world - so if you're accessing from Australia, you'll probably download it from an Australian node on the CDN. Sometimes, those copies linger long after those files are deleted.

Engineers are incentivized to provide fast, reliable implementations of required features and move onto the next thing. Storage is incredibly cheap, while processing time is less so. That means, in general, that they're likely to take the cheap, easy path and simply deactivate access to content rather than removing it. That's fine from a user experience perspective, but not from a user privacy and data rights perspective. GDPR, ePrivacy, and related legislation provide a much-needed stick to make content deletion do what the user expects it to do.

This sort of transparency of action is vital if we're going to have any sort of privacy online: if a user deletes content, they reasonably have the expectation that the content will really be deleted. If access is restricted to a few people, the user reasonably has the expectation that only those people can access it. Anything else is a breach of trust, not matter which terms may be hidden in the depths of the privacy policy. And if legislation is needed to bring about this transparency, then so be it.


What you're proud of

3 min read

I've always struggled with resumés.

The paper, career-orientated version of my life is one-dimensional at best. Here's what it looks like, more or less:

Built one of the first local classifieds websites. Graduated with an honors degree in Computer Science. Worked in educational technology at the University of Edinburgh. Co-founded a startup and an influential open source community. Worked for the Saïd Business School at the University of Oxford. Was CTO at Latakoo, a video transfer startup for newsrooms. Became Geek in Residence at the Edinburgh Festivals. Co-founded a startup and an open source publishing platform. Worked in engineering at Medium. Became Director of Investments (San Francisco) at Matter Ventures.

I'm proud of those things, for sure, but none of this really describes who I am. Even if I added clubs, programs, or volunteering, it would remain a very transactional list. I don't think the people who know me best would even recognize me in it. Where is the human behind the jobs?

That's what I wonder every time I look at a LinkedIn profile or receive a resumé as part of a hiring process.

Traditional resumés also do a grave disservice to people who have had a more eclectic journey. It's often seen as negative if you've tried a bunch of things that aren't quite a linear career progression. I don't think that's the owner's fault: everyone walks their own journey, which is a combination of luck, opportunities, creativity, and highly emotional decisions that are a product of their circumstances. But those factors, that underlying humanity, is completely lost on the page.

I wish resumés told a story. I want to know the narrative of a person. The why is often more important than the where. Not why did I take this job?, but why do I make the decisions I do? What motivates me?

And most of all: what am I really proud of? For me, it runs the gamut:

I'm proud of moving to California to be closer to my mother when she got sick, and having to be kicked out of the ICU because I wouldn't leave her side. I'm proud of building an online community that was a safe space for teenagers to come out. I'm proud of not being money-driven. I'm proud of financially supporting social justice organizations like Planned Parenthood and the SPLC. I'm proud of a short story I wrote a couple of years ago. I'm proud of cooking my Oma's Indonesian recipes and helping them live on. I'm proud of refusing to fall into the trap of traditional masculinity. I'm proud of always working mission-driven jobs. I'm proud of my fundamental belief that everybody is connected. I'm proud of my terrible puns.

All of these things are much more me. They don't fit on a resumé, but they also don't fit on a social media profile. They're also not just things I've made or organized; some are just characteristics, positions, or actions. But, together with the work I've done and other things I've made, they form a more three dimensional picture.

I wish there was a place where I could read the story of a person. Everybody's journey is so different and beautiful; each one leads to who we are. It would be the anti-LinkedIn. And because you wouldn't "engage with brands", it would be the anti-Facebook, too. Instead, it would be a record of the beauty and diversity of humanity, and a thing to point to when someone asks, "who are you?"


Becoming more interested in ICOs

4 min read

I started looking at blockchain from a position of extreme skepticism. Over time, mostly thanks to friends like Julien Genestoux and the amazing team over at DADA, I've come to a better understanding.

I've always been interested in decentralization as a general topic, of course - the original vision of Elgg had federation at its core, which is something I experimented with in Known as well. I'm also an active Mastodon supporter. It just took me a lot longer than it should have to see the implications in blockchain to actually bring those ideas about - mostly because of the very broey, Wall Street veneer of that scene. I don't need to be associated with the modern day Gordon Gekkos of the world; that's not what I went into technology to do.

What I did go into technology to do is empower people. I want to connect people together and amplify underrepresented communities. I want to help people speak truth to power. And I want to help create a fairer, more peaceful world. Speak to many founders from the early era of the web and they'll say the same thing.

By decoupling communications from central, controlling authorities, decentralization has the potential to do that. For example, the drag community was kicked off Facebook en masse because they weren't using their government-sanctioned names; that couldn't happen in a decentralized system. On the other hand, it's almost impossible to flag problematic content in such a system, so it could also allow marginalized voices to become even more marginalized with no real recourse.

But ICOs are really interesting. There is a well documented demographic bias in venture capital: it's significantly easier for well-connected, upper middle class, straight white men to receive funding. That's because most funding comes via existing connections; reaching out to investors cold is frowned upon and rarely works. The result is that only people who have connections get funding (except at places like Matter and Backstage that explicitly have an open application policy).

ICOs might be a different story. They are (theoretically) legal crowdfunding mechanisms that allow anyone to raise money, potentially from anyone - without diluting ownership of the company. Assuming you can pull it off (which is likely also dependent on having the right connections), you could potentially raise tens of millions of dollars without having to prostate yourself to Sand Hill Road. It's potentially very liberating.

But I need help understanding some of the mechanics - and I suspect the community in general does, too. 

In a traditional venture relationship, investors don't just bring money. They also bring expertise, connections, ideas, and sometimes even a shoulder to cry on. Your investors almost become like cofounders, and you build a relationship that lasts for many years.

In an ICO relationship, it seems to me that the incentive is for investors to dump their tokens almost immediately. You put your money into a presale, you wait for the price to go up, and then you immediately sell, because you don't know what's going to happen in the future. The good news is that you have your presale takings, but the potential for the post-ICO dump to irreversibly crash the price of your tokens seems high - which would effectively prevent you from being able to raise money in this way again. Not to mention the fact that you don't really have any kind of relationship with any of these investors. It's dumb, fickle money.

Equity is scary - you're giving away part of your company. But it also aligns investors with your mission. You're in the same boat: if you succeed, they succeed. At the extreme end, there's potential for certain kinds of investors to push you into unhealthy growth so they can see a return (sometimes employing toxic practices like installing their own HR team), but in general, I do believe that most investors are in it for the right reasons, and want to see companies succeed on their terms. I don't see an equivalent to the non-monetary side of the equation in the ICO world, and I worry that teams will suffer as a result.

But potentially I just don't understand. Just as a my friends helped me get my head into blockchain, I'd love some help with this, too.


I’m done with syndication. Let’s help people be themselves on the web.

2 min read

The IndieWeb has long promoted the idea of POSSE: Publish on your Own Site, Syndicate Elsewhere. In the wake of the Cambridge Analytica scandal, lots of platforms are re-evaluating their API policies.

This is kind of rearranging the deck chairs on the privacy Titanic, because the problem was that all this data was collected in one place, not that there was an API that allowed third party apps to publish on a user’s behalf. (To be fair, the publish API possibly enabled algorithmic propaganda / marketing campaigns to operate more efficiently.)

Still, here we are. I think this is a good opportunity to reconsider how the independent social web thinks of itself. I’ve long stopped syndicating posts to Twitter, and instead just post there directly. But I do try and post anything of substance on my blog.

POSSE requires participation from the networks. I think it might be more effective to move all the value away: publish on your own site, and use independent readers like Woodwind or Newsblur to consume content. Forget using social networks as the conduit. Let’s go full indie.

The effect of independence is practical, not just ideological: if you publish on your own site, your words are much more likely to stand the test of time and still be online years later. Social networks come and go, adjust their policies, etc. And there’s a business value to being able to point to a single space online that holds your body of thought and work.

Back when I was working on Known, investors would ask about the supplier risk of being so heavily dependent on third party APIs to provide a lot of the core value. They were right. Time to stop trying to integrate, and to double down on helping people own their own identities online in a way that helps them achieve their goals.


Not taking VC for the wrong reasons

3 min read

I come from open source communities and bootstrapped my first startup for the first few years. I've also been heavily involved in the ethics of data collection and the implications of high-growth models. Although I am one today (albeit at what I consider to be an ethican firm), I understand why people choose to avoid VC.

I'm worried, though, that some people have decided that venture investment is wrong for reasons that don't hold up.

I often see this from first-time founders who are used to having a paid salary that allows them to build product all day. Often, they would like to continue to do the same thing, but on a product that they control. It's a nice idea: frankly, I'd like that too. I'd love to be fully in control of a product I spend all day making.

But there's literally no model that allows you to do this as a full-time founder.

Whether you're bootstrapping, doing an ICO, taking venture investment, crowdfunding or soliciting donations, you still have to do the hard work of actually building a business. The same is true whether you want to make a multi billion dollar business, or whether you want to create something sustainable that pays for you to live well (the dreaded "lifestyle business", which is actually a perfectly fine and honorable thing to build). It's also true, for what it's worth, if you're building a non-profit: how are you going to keep getting enough donations on an ongoing basis so you can make a profit and grow when you need to?

It forces you into some uncomfortable positions - particularly if you've never run a business before. Data-driven testing barely makes sense when you're starting out (how do you get to statistical significance?), but that's what most developers seem to want to do; in fact, a whole bunch of qualitative, real world understanding is required before you write even a line of code. And then you need to keep doing it, while you figure out your growth strategy, your pricing, what your user journey looks like, how you retain users, and so on.

Those aren't things that VC businesses need to work on. That's something every internet business needs to figure out. If you're sitting and building code all day, as fun as that would be, you're doing it wrong. The code exists in service to the business. You need to figure out your core risks and address them: your user risk, your business and financial risk, and finally, your technology risk. You need to be able to build something that people want and which will viably make money - with the time, resources, and expertise at your disposal.

Investors can give you not just the financial runway to figure that out, but also the expertise. They've seen most of this before; they can connect you to people who can help you. The wrong investors will absolutely lead your company to horrible places, but the right ones, who interact with you with a service mindset, will help you achieve your goals - whatever they are.

If you're doing something good, you need to make it sustainable so you can keep doing it. Smart, ethical investment can help with the money, and it can help with the network and skills to actually build your business. Sitting and building product all day absolutely won't.


A backdoor to democracy

1 min read

In 2001, I built viral personality tests that for a while were the most  shared content online. In 2016, this same mechanism was used as a backdoor to democracy.

I wrote a piece over on Medium about how Cambridge Analytica created psychographic profiles for 50 million users:

Elections have become information warfare battlegrounds, fought using all of our personal details without our consent. Here, the weapons are vast data silos like Facebook and Twitter, alongside open, anonymous marketplaces for highly targeted advertising.

Read the whole piece here.


A day in the life of an engineer turned investor

4 min read

When I talk to former colleagues about my life at Matter, and in particular how much of my day I spend talking to people. As an engineer, maybe I had three meetings a week; these days it's often eight a day. And I love it: as a former founder, I'm excited to meet with hundreds of people who are all working on things they care deeply about - and I'm excited to find them.

This is what yesterday looked like for me:

7am: I finished a blog post draft that will be published on Thursday. I'm excited about intelligent assistants and the shift to ambient computing, and I was able to back up my piece with sources from an internal investment trend document I wrote.

8am: Headed into work, listening to On the Media, my favorite podcast.

9am: Caught up with email. I'm still figuring out a process for this: I get more than I can really handle, and I don't feel good about sending one-line responses.

9:30am: A standup with the team, talking about the day, and any new developments.

10am: I welcomed a group of foreign journalists who were interested in Matter. We talked for an hour about new trends, how we think about products vs teams (hint: we invest in teams), and whether there's still a future for print.

11am and 11:30am: I jumped on the phone with some founders who wanted to learn more about Matter, and whether it would be a good fit for their companies.

12pm: More email, including outreach to some startups that I'm hoping will apply. There are a lot of people out there who don't think of themselves as working on a media startup, but who are exactly what we're looking for, and who could be substantially helped by the Matter program.

1pm: I joined in on a workshop with our Matter Eight teams, thinking about how to pin down the top-down trends that make their startups good investments. Key question: why is now the right time for this venture? Our Directors of Program are, frankly, geniuses at helping people think their way through these kinds of questions, and I'm always excited to learn from them.

2pm: I sat down with the CEOs of one of our portfolio companies to give them some feedback on how they're describing their venture to investors.

3pm: I spoke to another founder who didn't join Matter, but wanted to give me an update about where they were. It's always exciting to hear about how a team has progressed.

3:30pm: I took an audit of our application process on the web. Some applicants drop off while they're filling in the form, and I wanted to know where that might be happening. At the same time, I did some SEO work on the website. (SEO work follows me in every role, wherever I go.)

4pm: I have a personal goal of reaching out to at least five startups a day - so I spent more time doing research and uncovering both communities to visit and events to attend, as well as individual startups that I would love to see join the program.

5pm: Facilitated introductions for some portfolio founders who wanted to meet certain investors. I always do double blind introductions, asking the investors first if they want to connect. Then I turned to going over our applicants, reading through their decks, and doing some research on their markets and founders.

7pm: I went home to eat.

8pm: I caught up on my RSS subscriptions, reading about the various industries and founders I'm interested in.

There's no time for coding anymore - but there's a lot to do, and I couldn't be happier to support these amazing founders. If that's you, applications are open now.


The GNU Peaceful General Public License

3 min read

I've been thinking a little bit about repurposing of software. One line I've always said I won't cross - inspired by one of the original investors in Elgg, who said the same thing - is that I won't build software that will be used directly or indirectly to kill people. That rules out working on defense contracts, or anything involving weaponry.

The trouble is, software can be repurposed. You could write an algorithm that identifies objects in photographs in order to improve search results, for example, and come into work one day to discover that it could be used for drone targeting. Algorithms can be used for evil. (I would argue that drones, at anybody's hand, fit the definition of "evil".)

I mentioned this on Twitter this morning, and Julien Genestoux made a really important point:

 Open source software can be used by anyone for anything, as long as the four freedoms are adhered to:

The freedom to run the program as you wish, for any purpose (freedom 0).

The freedom to study how the program works, and change it so it does your computing as you wish (freedom 1). Access to the source code is a precondition for this.

The freedom to redistribute copies so you can help your neighbor (freedom 2).

The freedom to distribute copies of your modified versions to others (freedom 3). By doing this you can give the whole community a chance to benefit from your changes. Access to the source code is a precondition for this.

There is nothing to prevent unethical use of the software. This is a real gap: while I applaud the principles of freedom at work in open source licensing, I would be appalled if Elgg or Known or anything else I'd written were used to cause harm to others. I want no part in that.

Specific modifications to open source licenses exist to achieve certain goals. For example, if software is released under the GNU Affero Public License, running it on a server for people to use counts as redistribution, and any modifications to the code must be made public.

So what if there was a version that refused use for military / defense applications? That would allow software to continue to be used freely, but would deny the license to anyone directly working for, or contracted by, military or defense organizations. Those parties would need to negotiate a specific license, allowing the softare vendor to make decisions on a case by case basis.

The license wouldn't be universal - not everyone has the same objections I do. But for developers like me, it would provide some peace of mind.


How we run the Matter application process using Typeform, AirTable, Zapier and Slack

7 min read

Applications for Matter Nine are open. It's my job - together with my New York City counterpart, Josh Lucido - to run the process, source candidates, and find the twelve teams that will walk through our San Francisco garage door on August 13.

We get many hundreds of applications for every class, which almost all arrive via our website. The trick is to ensure that everyone is handled fairly, robustly, and with transparency internally to the team. Nothing happens based on a whim, and nobody can fall through the cracks.

Inspired by Nick Grossman's piece about how Union Square Ventures ran their analyst application process, I thought it might be interesting to show off how we're using a collection of tools to drive our Matter accelerator application process.

The application form

The entire application to our accelerator takes place on a single form. We don't ask for a video, although we do want to see links to external resources like your website - and we definitely want to see a deck.

We've used Typeform to power our application form for years. The interface is both simple and pleasant to use. For a while, we had it embedded on our site, but a few users reported that the embed didn't work well on mobile devices, so I decided to link directly to the form instead.

Although the form is designed to be quick to fill in, we ask for a lot of information that will be useful to us as we make our decisions. (It's early stage, so these answers are more than likely imperfect, and that's fine.) Do you know who your user is? Who is the team, and can you execute? What is the mission, and why is that important? What are the trends that make this the right time to start this venture? How do you think you'll make money? We also ask diversity and inclusion questions to help us track our progress on our goal to build a more diverse and inclusive kind of startup community.

All of this data is used to make decisions in the sourcing process individually. It's also used in aggregate to examine trends in the startups that apply to us, and to help us figure out where the gaps in our sourcing might be, as well as how to iterate our process.

So storing it in a way that can be analyzed easily is vital. I don't have time to write my own scripts, and the investments team shouldn't need to have a computer science degree or know how to code in order to do this.

Luckily, AirTable exists.

The database

AirTable looks like a spreadsheet (at least, by default), but is much more like a database. Datasets are split up into "bases", which each contain "tables". Each table in a base can reference each other. And while a traditional database might have field types like text and integers, AirTable adds file-sharing, images, tagging, spreadsheet-style formulae, and a lot more.

Our ecosystem base has two core tables: People and Companies. These contain all the people and all the companies in Matter's ecosystem; not just those who have come through the application process.

To that, we add Applications and Assessments. Almost every question from our form is represented here. For example, we use a tag field (technically a "multi-select") for the areas of focus for the venture, a text field for a link to the deck, and long-text for the qualitative questions.

Our form asks about each member of the team, and these are represented in the People table. Similarly, the startup itself is added to the Company table. Each Application links to a Company, which in turn links to several People. That way, if a company applies to several classes, we can easily see each of them, and see how the company has evolved from one application to the next.

Because AirTable allows us to view a table using a Kanban view, we can easily create a view that starts applications in Inbox, allows us to drag them to Under Consideration, Invite for Pitch, and so on. It looks like this (I've hidden our actual applicants, and there are closer to 15 statuses in total):



For every single startup that applies, we assess the applicant using a special set of questions that we also use in our Design Reviews throughout the program itself. The answers to these questions get stored in the Assessment table, which links to the Application table. AirTable lets us structure this as a form, which I keep linked from my browser bookmarks tab:


(This is a subset of the questions.)

So to assess an incoming application, and at each stage of the application process, each reviewer's feedback is captured on the form, which is them recorded in AirTable. The investment team meets every week to decide who to advance through the process, based on the feedback.

Connecting Typeform to AirTable (and letting us know about it)

I built a Zapier zap to automatically translate incoming applications from Typeform into AirTable (as well as to notify us in a special investments-incoming channel in Slack).

It looks at the company in the application; if it doesn't already exist in AirTable, it builds a new entry in Companies. Otherwise, it updates the existing one.

It looks at each individual in the startup; if they don't already exist in AirTable, it builds new entries in our People table. Otherwise, it updates the existing ones.

And finally, it always creates a new Application entry, sets the Status to Inbox, and sends a summary of the information to Slack, so we're immediately notified that something new has come in.

In summary

We can now track every application for every company, including all our assessment notes, from a simple interface that also allows us to perform operations on the quantifiable information we capture. From this, we could theoretically create live dashboards that chart our process; we can (and do) also create static summaries of how our applications pool breaks down across themes, stages, team skills, intersectional diversity and inclusion statistics, and more.

I wish some of these steps were easier (for example, if AirTable's own forms were prettier, we might not need to use Zapier etc at all). And there are definitely things we could improve. Still, it's a robust process that allows us to run a very competitive application process in a data-driven way using a small team.

In the future, this structure will allow us to add new interfaces - for example, why not apply to Matter with a conversational chatbot? - that talk to this AirTable back-end. We can also easily perform experiments with the application process to make it more streamlined, brand application forms for specific events or partnerships, or better support certain communities.

In particular, I've been incredibly impressed with AirTable, and I've started recommending it to everyone. I'd love to hear your experiences.

And of course: Applications are open. Join Matter Nine today.


Blogging and newsletters

2 min read

I'm doing a lot more writing on my own blog this year. Writing has always helped me think through a problem space and socialize ideas; it's a good way to get feedback on something you're thinking about early. Unlike an article or a project, a blog is deliberately imperfect, and it works best if you do it regularly.

I'm not a lone blogger. There's something nice about reading my news feed away from the noise of social media (and on a reverse-chronological stream rather than somebody's algorithm), and I find myself learning about things I never would have otherwise discovered. It's a struggle - social media absolutely is addictive - but it's really worth it.

Here's my reading stack:

I use NewsBlur to power my subscriptions. It's $36 a year, and absolutely worth it. One great feature is that it allows you to subscribe to email newsletters: you can create a Gmail filter to forward your newsletters to a special address, and essentially keep them confined to a special inbox.

Because the reader ecosystem is pretty open, a few native apps are available. I've settled on Reeder for my Mac and iPhone; it's slick and gets out of my way. (The mobile app is $4.99 and the desktop app is $9.99.)

I'm trying my best to only follow individuals for now, but I expect I'll start adding some particularly insightful corporate / startup blogs over time. I'd love to hear recommendations.

Finally, I've noticed that Fred Wilson (who has blogged every single day for years) also allows readers to subscribe to his posts via email. If I'm going to continue to write on a regular basis, this seems like a pretty good idea; sadly the RSS ecosystem, as wonderful as it is, is very far from being mainstream at this point. He uses Feedblitz, and I'm thinking of giving that a try too. I'd love to hear whether you'd find that useful.


"Rock ain’t nothing but a white version of rhythm and blues, motherfucker." This is such an amazing interview.


"Facebook has created a centrally designed internet. It’s a lamer, shittier looking internet." How Facebook is killing comedy:


My friend Hugh Hancock suddenly passed away this week. He was instrumental in inventing Machinima. Here he is on Charlie Stross’s blog announcing his film Death Knight Love Story, starring Brian Blessed and Joanna Lumley.


WebSkills - a proposal for open intelligent assistants

3 min read

It's clear at this point that intelligent assistants - and more broadly, ambient computing devices that you interact with naturally, rather than holding like a smartphone or laptop - are going to play an important part in our digital future.

Of the platforms doing the rounds at the moment, I'm most excited by Alexa, because of its relative openness: Amazon has made it available as an operating system for manufacturers, so it'll start showing up in cars and offices, and they've treated their product line-up as a series of proofs of concept. Nice.

Still, you need to plug Alexa Skills (their name for apps) through their APIs in a relatively closed way. Back-end deals need to be done for new functionality, and so on.

What if that didn't need to be the case? Picture this:

1. I'm using my favorite web service. It lets me know that I can install its functionality into my WebSkills-compatible intelligent assistant, either using UI on the site itself, or through a strip at the top of the page, a bit like how Safari on the iPhone tells you about relevant apps. I push the button, because I'd love to be able to talk to this service whenever I need.

2. My device prompts me to make sure I want to authenticate with this skill and install it in my assistant. Sure I do.

3. What's actually happening is that an endpoint, referenced in the website's HTML through a <link rel="webskill" href="..." trigger="service name"> tag, is being registered with my device. (No, trigger isn't a valid link propertt right now, but bear with me.) The trigger is the unique service word that can be used to trigger the request. For example, if the trigger was "Wolfram Alpha", the request to the assistant might be of the form, "Alexa, ask Wolfram Alpha what is the GDP of Bhutan?"

4. When a request is made, the intelligent assistant looks to see if the trigger word has been registered. It then calls the associated URL from the link tag using a GET request with a q property that contains the full text of the request.

5. The endpoint returns either text to be read out, or the contents of a WAV or MP3 audio file. The intelligent assistant dutifully plays this out.

This is one example of a simple mechanism that would allow any provider on the internet to add intelligent assistant skills in a cross-platform way. It's unsophisticated, but it would allow a thousand intelligent assistant platforms to bloom, with the web at their core, rather than a few monopolistic platforms.

I'd love feedback! It's easy to talk about these kinds of projects, but talk is cheap, so my next plan is to build a proof of concept.


The Prototype Fund is looking for 25 open source projects to fund with up to €47,500. You need to be resident in Germany, but honestly, this sounds great.


Stellar as a platform (not a speculative investment)

2 min read

I'm starting to more closely follow developments over at Stellar. The platform forked from Ripple a few years ago, and while the former has made hard connections to a bunch of traditional banks (Santander, UBS, American Express are all partners), the latter is trying to create a genuine platform that can be used to underpin a variety of real services that allow people to quickly move funds in a decentralized way, including across borders.

As a result, one of its areas of interest is helping to support the underbanked. That absolutely piques my interest: if services can be built to lift people out of poverty in a non-predatory way (that last part is key), it's a genuine good for the world. LALA, which helps migrants send funds back to their unbanked families, is one such service that just announced it'll be using the platform.

Open Garden is another: a way for people to share internet connections with each other. You share your internet via your phone's wifi hotspot, and earn value that can in turn be traded to use someone else's bandwidth while you're on the move. Theoretically, this should prevent people from taking without also contributing to the network.

A lot has been made of its partnerships with IBM and Deloitte; Stripe is also on the list - and has been since the beginning - which implies some interesting payment integration possibilities in the future. Of course, it might also just be watching the market.

But what's most exciting to me is the advisory board: the founders of WordPress and Stripe are both represented, as is the Director of the Apache Software Foundation, and Sam Altman from Y Combinator. That's a solid combination of platform builders, openness experts, and startup supporters.

I do have a small number of Lumens (840 at the time of writing; my only cryptocurrency holding), because I decided I wanted to learn more about the space. I'm very turned off about speculative cryptocurrency investments; it's just not something that's interesting to me, in the same way that hedge fund trading is not in any way my bag. What I've always cared about is open platforms that have a positive societal impact, and I think that's what Stellar has the potential to be. Once the Wall-Street-like crypto buzz has died down, these are the things that will matter.



Edinburgh, Austin, San Francisco - a startup tale of three cities

3 min read

I sometimes tell the story of the three places where I’ve been involved in founding startups.

in Edinburgh, the cost of living was low, and I never worried about healthcare. I don’t think I would have founded a startup, or entered this world, without these kinds of democratic socialist protections. But at the same time, everybody told me it would never work and that I should get a real job. And although it’s changed since, in 2003 there was absolutely no infrastructure for starting this kind of business: precious little money or expertise.

in Austin, there was a lot of enthusiasm; very little “get a real job”. But with the exception of certain pockets, my perception is that investors were primed for more traditional businesses, and didn’t quite have the risk appetite or the value-add in terms of expertise they could offer. (This is changing rapidly, too.)

In San Francisco, there’s money and expertise everywhere. You can get funded and have coffee with people who have been on the journey many times before. Sometimes, you bump into those people in the burrito line. It’s a completely different universe. But, correspondingly, the cost of living is much higher and it’s harder to stand out, because there are a million other startups vying for everybody’s attention.

It’s not quite Goldilocks and the Three Bears, but it’s not three even choices, either.

I don’t think it’s possible to build a technology business and not at least visit the Bay Area regularly. Should you live here full time? I’m actually not sure - although you’re maximizing your opportunities for serendipitous meetings, the whole area is absolutely beautiful, and everyone’s really just a short walk away, you’re also meaningfully shortening your runway. There are lots of people here who don’t work in startups, so it’s absolutely possible to stay grounded, but some people only travel in those circles, and that’s an existential danger, too. And obviously there are the people who are only in it for the hope of VC money, absolutely everywhere.

Like everything, I think you’ve got to work out what’s best for you, your team, and your mission. But start with the individuals. What nourishes you? What kind of place will make you feel supported even when things are going wrong? Where does your joy come from, and where can you be in a place that makes you feel passionate about something, where you feel like being human is beautiful and not something flawed that needs to be improved? Where will you not just work best, but live best?

I’ve found that here, but it’s different for everyone. Start there and work backwards.


Founder conflict: “it is OK to talk about it, put it on the table, and deal with it.”


On Holocaust Memorial Day

2 min read

It’s easier than ever to understand how the Holocaust happened. As many have said, it doesn’t start with concentration camps and gas chambers; it starts with scapegoating, labeling certain groups as inferior, and reducing their rights. It starts with bigotry. And standing by in silence, which is a quiet bigotry all of its own.

The Holocaust was legal. It’s the clearest example of how justice isn’t the same as the law, and how standing up for what’s right is not the same as upholding what is legally allowed.

The people who were silent were patriots. They believed in their country. They believed they were putting Germany first. They didn’t question their leaders because they believed in the greatness of their nation. Or, worse, they just didn’t want to care about “politics”.

After the war, principles were established. If you don’t question authority - even as a soldier - you are complicit. If you’re asked to be part of a war crime, or if a war crime is the path of least resistance, you must refuse. Everyone has agency and you don’t get to hide behind superiors. Soldiers have commanding officers; civilians have peer pressure and social norms.

It’s worth asking, in 2018, what you would do if you saw any group marginalized in the way people of Jewish descent were in Germany in 1933. What would you say? Where would you march? To what lengths would you go to preserve democracy and equality?

Because it’s up to all of us. It always is.


It's two minutes to midnight, for real

2 min read

For 71 years, the Bulletin of the Atomic Scientists has used a clock to represent how close we are to armageddon. Midnight represents the end of human civilization. And today they brought it closer to midnight than it's ever been: just two minutes away.

In the United States, Russia, and elsewhere around the world, plans for nuclear force modernization and development continue apace. The Trump administration’s Nuclear Posture Review appears likely to increase the types and roles of nuclear weapons in US defense plans and lower the threshold to nuclear use. In South Asia, emphasis on nuclear and missile capabilities grows. Conventional force imbalances and destabilizing plans for nuclear weapons use early in any conflict continue to plague the subcontinent.

This is closer than during the Cuban Missile Crisis or any moment in the Cold War since 1947 - which sounds surreal, or melodramatic, even. But here we are.

Somehow we've moved away from global peace and diplomacy to a world full of posturing, inequality, and isolationism. The only way to turn back the clock is to bridge divides and create a more inclusive, empathetic society once again: one where everybody has the ability to prosper and the emphasis is on the global human experience, not the exceptionalism of just one nation.

Turning back the clock is all of our jobs.



Eventually, every app builds for the web. Here's why.

2 min read

Snapchat is letting users share some stories to the web:

By opening up Stories to the web, Snap envisions a way for content on its platform to go more broadly viral — the way Twitter and Instagram posts have captured real-time news and cultural events. News organizations, for example, could link to Snapchat Stories on the web, while celebrities will be able to share their personal Snaps outside of the app.

This is exactly why every social app will eventually allow users to share to the web. A crucial part of every user journey is discovery: that touch point where someone discovers your service for the first time. Building something slick and assuming users will just show up is a massive mistake: they simply won't.

In the name of growth, and because it's a genuinely useful feature, every social service eventually allows you to share content with people who haven't signed up yet. And when you do share to someone who doesn't have the app installed yet, there are really two possibilities:

1. They get a page telling them to install the app.

2. They get a preview of the content that they would experience using the app.

Speaking for myself, I would never randomly install an app from a share - or at least, the barrier is much, much higher. Most people carefully guard what they install on their phones. But if I click through and see some great, personal content without needing to install the app - and then I see more and more of it over time, perhaps via Twitter or Facebook, but potentially sent to me via IM or email - I'm much more likely to install the app and sign up myself. That's the growth story for Instagram. A version of it was the growth story for YouTube. And even Twitter, back in the old days, had amazing web embeds that started to show up on peoples' blogs.

Sharing an experience without asking you to install software is something only the web can do.

It's a sign that Snapchat wants to grow faster and build a much larger audience. It's also a sign that it's growing up beyond what was an exclusive, and slightly obtuse, social network into something it wants everybody to use. Such is the path of every social network.



Could stagnant wages be because we - and particularly non-coastal workers - have a smaller total number of employers to choose from?


Decentralized paid subscriptions for independent publishers

3 min read

A real problem that needs to be solved is making it easier to subscribe to independent publishers putting out great, regular content. Online magazines, blogs, podcasts, etc. Independence and autonomy are important, but discovery and ease of use are too.

RSS is a pretty ancient technology, but it's in far more use than you'd think. For example, every podcast runs on RSS. There are a lot of sites that use MRSS behind the scenes, to power portals like AOL News, and to ingest multimedia content in back-end systems. Readers are largely gone, but not the backbone technology.

What RSS is missing is authentication. Knowing who the user is would allow for more personalized experiences, and it would also allow publishers to add business models to monetize their distributed content.

So what if we added OAuth 2.0 as a really simple auth layer, so that content providers could accurately assess who was requesting a feed, podcast, etc?

Add three new tags to the RSS feed:

  • The URI of the OAuth endpoint
  • A human-readable URI where an authenticated user can pay to subscribe or manage their account
  • Whether this feed contains premium content or not (maybe a label for the content level - "free" / "subscribed")

This way, a compatible feed reader / podcast client could tell a user if it's possible to subscribe to get premium content. They could auth the user (possibly allowing them to register with the publisher) and point to a subscription page.

From then on, the reader makes a signed request whenever it looks for the feed. The publisher is responsible for figuring out whether to serve premium content or not based on the user's identity.

The publisher gets to decide which CMS to use, which payment provider to use, how much to charge, etc etc - they retain full autonomy. If they want to use Stripe; fine. Bitcoin; whatever. The only major standardization point is authentication itself.

The market is then open to anyone who wants to create a hub for finding content. Publishers might pay the hub to promote their sites - or lots of business models are possible. But paid subscriptions are baked into apps and readers, and are totally under the publisher's control.

Everyone gets to have their own website and content model. Everyone gets to have a standard way of pointing to a built-in revenue model, and decide what that is.

Imagine if Apple News, Flipboard, Medium, and maybe even the Facebook news feed, as well as hundreds of independent apps, could all feed directly into independent publisher revenue streams.

Anyway, just a thought I've been having. Thought I'd share.


This piece was originally a tweetstorm.


The news industry needs to wake up and join the web

4 min read

Emily Bell has a timely opinion piece in The Guardian today about Facebook's ethical responsibility with respect to news:

Facebook’s retreat from news, and the complexities of taking responsibility for the type of content circulating on its platform, has many implications for press organisations in the US and Europe, but at least in rich, western democracies, its actions can be mitigated by other strategies. In countries such as the Philippines, Myanmar and South Sudan and emerging democracies such Bolivia and Serbia, it is not ethical to plead platform neutrality or to set up the promise of a functioning news ecosystem and then simply withdraw at a whim.

Yes, Facebook needs to recalibrate itself and understand the responsibilities that go alongside its position. But in so much news commentary there's a subtext that megasilos like Facebook, and the internet as a whole alongside them, are some unmovable force of nature that require a reactive response.

The internet is an open platform evolving through collaborative means. The web is open source. All of the paradigms we've come to use across software have evolved over time, one set of developers iterating on ideas created by another, iterated upon by another set, and so on. Standards on the web are open source. New movements and innovations are typically created by very small groups of people, failing fast and prioritizing running code over consensus, which are then codified by working groups that themselves are made of loose federations of people.

Yes, Facebook et al deeply need to understand their responsibility to democratic society and adjust their objectives in that light. But the news industry need to deeply grok that it isn't subject to the whims of the internet. If organizations lean in, they can materially help shape the platforms that have disrupted their businesses. They're not doomed to be outsiders; they are welcome to join.

At the beginning of Emily's piece, she notes:

The homepage is back, and not just for those chronically old people over 40, but for every news organisation that wants to survive falling off the great Facebook cliff of 2018.

The homepage's return is a very good thing. Any information business needs to have control over its platform. Returning to the feed economy and innovation around new ways to subscribe to information will also be good; let a thousand reader services bloom. I'm still waiting for the first decentralized reader with integrated subscription or per-item content payments, but those are the kinds of developments we need. And they're the kinds of developments that need to have publisher voices included - or even to be driven by publisher organizations.

Why were news organizations so dependent on one company's algorithmic policies to begin with? Yes, they capitulated to insane supplier power, and yes, it looks like a horrible decision in hindsight (as well as to those of us who worked in open technology at the time). But their business models were collapsing, and it was an easy answer. Most of us would probably have made similar decisions under similar pressures. But it's time to move on.

Publishers need to be supporting and collaborating with teams building products, perhaps through organizations like Matter (selfish plug, but also, the partner program really works). They need to be supporting the evolution of technology platforms by joining organizations like the W3C and participating in groups like WHATWG.

And finally, they need to start collaborating by building the software they want to see in the world, under an open source license, in a way that allows all of them to benefit. It's not about building something that draws a direct profit; instead, they can help create an ecosystem that better supports their current businesses, and provides a clearer framework for supporting them as their businesses evolve into the future. They need to hire teams to build an ecosystem that holistically supports them, and in turn, democracy.

Because honestly, Facebook has put journalism in peril. And there's no such thing as democracy, or freedom, without it.



Very sad to hear that Jon Oberlander died. He was responsible for a lot of important change in Edinburgh informatics and beyond - and more importantly, was a kind, thoughtful man.


America’s attitude to immigration right now is disgusting.


The letter from a Birmingham jail is an important piece to reflect on today, in a world where many still hold the law to be more important than justice.


Bad UX and the Hawaii missile scare

2 min read

It turns out that the origin of the missile scare that terrified Hawaii the other day was a poor UX choice. Instead of triggering a test alert, a civil defense employee accidentally triggered a real one - which then wasn't rescinded for 38 minutes:

Around 8:05 a.m., the Hawaii emergency employee initiated the internal test, according to a timeline released by the state. From a drop-down menu on a computer program, he saw two options: “Test missile alert” and “Missile alert.” He was supposed to choose the former; as much of the world now knows, he chose the latter, an initiation of a real-life missile alert.

He "feels terrible", according to reports. I bet.

It reminded me of a UX story that was relayed to me by a university advisor.

My computer science dissertation was on accident reporting: I built a system that allowed you to log events that led to an accident, and then draw weighted causal links between them (using primitive JavaScript, because this was 2001). You could then determine the root causes of an incident using graph theory, often revealing issues that might not have been obvious.

This particular advisor had been involved in assessing the incident at Three Mile Island. The controls to shut down the reactor were, as you might imagine, in a prominent, protected place in the operations room, to avoid accidental shutdowns. It wasn't quite a big red button, but it was close. Which was all fine, until a maintenance officer had to change the overhead bulb: they were observed climbing onto the console and placing their feet on either side of the shutdown control.

I don't know how true the story is, but it's good even if it's just a parable. Observing your user and understanding their context is vital if you're going to design something for them. That's particularly important if the consequences could be life or death, when surprising insights can mean the difference between war and peace.

I feel really awful for the people of Hawaii. I'm also not at all keen on the Cold-War-esque atmosphere that seems to be ratcheting up. May cooler heads prevail.


Facebook is deprioritizing news posts, and that's great!

3 min read

A lot has been said about Facebook's upcoming changes to its news feed, which will downgrade posts from Facebook Pages and news publishers in favor of people you actually know. Facebook stock fell 4.5% in response: not a lot, but enough to be felt.

It's easy to see why they're changing their strategy here, even though it will result in shorter visits to Facebook and fewer ad dollars spent in the short-term. In addition to having been instrumental in the Brexit referendum and the instrument for foreign actors hoping to sway the US election (not to mention a propaganda weapon for the likes of Duterte), passively reading your Facebook feed makes you feel bad. Over time, that can only result in fewer people using the service. (It's also worth noting that linking itself so tightly to journalism may cause it difficulties in China.)

Publishers are variously up in arms. Digiday's post was particularly alarmist:

The end is nigh. Facebook is planning a major change to its news feed, starting as early as next week, that will decisively favor user content and effectively deprioritize publishers’ content, according to three publishers that have been briefed by the platform ahead of the move.

The end is nigh. Later on in the piece, one anonymous publishing executive is quoted as saying, "we're losing hope".

But I don't think any of this fear is warranted. This is the web, and Facebook isn't the only game in town. Publishers are already diversifying away from it in order to acquire readers, strengthening their businesses in the process. Facebook's monopolistic supplier power has been overwhelming for the last few years, and the result has not just been felt in the publishing businesses themselves, but in democratic society. A change is long overdue.

Some good thought experiments for web technologists in publishing houses are: what does it look like to retake control of our distribution? How can we work with other publishers, as well as startups and technology companies, to make reading the news easy and fun? We've been hacking the monolithic social network model to be a news distributor for the last decade, but what else is already out there, and who can we work with?

There's a lot out there, from new kinds of technologies explicitly designed for distribution that gives publishers more control, to new ways to pay for content, to interesting new platforms for discovery. And this is before we consider new paradigms like ambient computing (Alexa etc), AR and VR, which are all on the up.

Overall, a lot is possible on the web, if you speak to experts, understand your audience (and your potential audience) deeply, and approach distribution with an innovation mindset.

And to think, not so long ago, publishers were contemplating moving themselves wholesale onto Facebook itself. What a disaster that would have been.