Brad Feld

Month: July 2018

We’ve been soaking (and investing) in the world of 3D printing since our investment in MakerBot in 2011. Since then, we’ve made three other investments in the world of 3D printing – Formlabs, Glowforge, and Looking Glass. While Looking Glass is a holographic display and is the inverse of a 3D printer, you’ll see how it fits into this in a moment.

While I continue to be impressed by Desktop Metal, the incredible press that they get, and am a big fan of the writing of Jason Pontin, I think Jason’s story in Wired – 3-D Printing Is The Future of Factories (For Real This Time) – misses several key points that take the idea that 3D printing is the future in a different direction.

I’ve decided that 3D printing is the REPL for Hardware. Dan Shapiro, the CEO of Glowforge, coined this and he’s completely nailed it. If you don’t know what a REPL is, it’s a programming concept called the Read-eval-print loop. Following is an example of a Python REPL running in a browser.

In the middle window is the Python code for a simple factorial. You hit the “run” button and the REPL reads the Python code, evaluates it, and prints the answer (120) in the right window. Hang on to that idea – Read, Evaluate, Print – we’ll be back to it later.

My first thoughts around 3D printing started at the beginning of 2010, when I read Chris Anderson’s Wired essay In The Next Industrial Revolution, Atoms Are The New Bits. His 2012 book Makers: The New Industrial Revolution helped me understand this better.

At first, I was obsessed with distributed 3D printing. On the desktop. For professionals. Each of MakerBot, Formlabs, and Glowforge took expensive industrial products that cost $50,000 to $500,000 and put them on a desktop for $2,000 to $5,000. The metaphor we used for this was that of the evolution of the laser printer market. If that’s elusive to you, the path from mainframe to PC works also.

We started with our investment in MakerBot, which used a technology called FDM (fused deposition modeling), which is a cousin of FFF (fused filament fabrication). This is a fancy phrase for “heating up plastic, extruding it, and building a 3D thing with the heated plastic”. It was magical, but limited on many dimensions based on the constraints of the materials and the process. As a result, MakerBot (and FDM in general) is primarily a hobbyist and DIY product.

FDM is additive manufacturing. So, when Glowforge came along, we immediately recognized it as the analogous subtractive manufacturing technology. Glowforge is a laser cutter (basically the addition of a laser to a CNC machine). You shine a laser at a material – any material – over and over again to subtract from the material to make your 3D print. As magical as MakerBot was, what Glowforge could do was mind-bending and took 3D printing to an entirely new dimension for me. You could work with paper, granite, sushi, cardboard, chocolate, wood, and basically any other material. Plus, well, LASERS!

We knew Formlabs from our experience at MakerBot. If you haven’t seen the Netflix documentary Print the Legend and you are interested in this stuff, go watch it. It’s the early story of both MakerBot and Formlabs, has endless cringe-worthy moments in it, stars some of your friends, and shows how incredibly challenging a startup is.

Since MakerBot had been acquired, it cleared the way for us to invest in Formlabs, which we did in 2016. Formlabs first product was based on SLA (stereolithography). In this technology, you shine a laser at a vat of resin and it builds up the 3D print (again – additive manufacturing). The quality and fidelity of the 3D prints are much higher with SLA, there are a wide variety of resins, and, as a result, Formlabs has had great success in the prosumer market. Next year, Formlabs will be shipping the Fuse, an SLS (selective laser sintering) 3D printer, which is another, even more advanced technology, at a desktop price point.

With Formlabs and Glowforge, we now have high-end desktop 3D printers at a price point under $10,000 (or at least 1/10th the price of similar industrial products). They are WiFi connected (just like today’s laser printer), have contemporary software, are integrated with everything 3D software related, and work extremely well.

Let’s go back to REPL. You start with a 3D image, which you can either design, get from the web, or get from Thingiverse or Pinshape, Formlabs and Glowforge printers then provide the REPL – it reads the 3D code, evaluates it, and prints it. On your desktop. Next to you. In high fidelity. Right away. Inexpensively.

So – how does Looking Glass fit into this? It still takes some time for Formlabs and Glowforge to print the 3D object, so the output from the REPL isn’t immediate. What if you could visualize the 3D object, in 3D, as an intermediate step? Voila – a holographic display (also known by a variety of other names like lightfield or volumetric display.) Looking Glass is 3D visualization on the desktop, which makes the desktop 3D experience for the professional even more powerful.

In 1984, HP shipped their first HP Laserjet. a 300-dpi, 8 ppm printer that sold for $3,495. A decade later, HP shipped its 10-millionth LaserJet printer. By the end of 2000, they had shipped 50 million of them. Over the weekend, I installed an HP LaserJet Pro M277dw Wireless All-in-One Color Printer, a 600-dpi, 19 ppm WiFi connected color printer / scanner / fax that sells $484.

When someone asks me what they can do with a 3D printer, I wish I could shove them into my time travel machine and send them back to early 1985 to ponder the question “Why do I need a laser printer on my desktop – I’ve already got an Epson MX-80.”

If you make anything in 3D, you now can have a 3D REPL on your desk.


Google Boulder recently did a phenomenal thing. They recently gave a gift of over $2 million to CU Boulder, which included free office space for NCWIT for the next six years (valued at $1.3 million.) As of a few weeks ago, NCWIT now has a great long-term home in an older Google office on 26th Street in Boulder off of the CU Campus.

The head of Google Boulder (I think his official title in Googlespeak is “Engineering Site Director”) is Scott Green. I’ve known Scott since shortly after I moved to Boulder in 1995. He was an early employee at Email Publishing (which became MessageMedia), my very first Boulder-based angel investment. After MessageMedia, he spent some time working at Return Path (where I’ve been an investors since 2000) early in its life before moving to @Last (which we were not investors in, but were fans of since some of our friends, including Brian Makare (the co-founder of Email Publishing) and Mark Solon (then of Highway 12, now at Techstars) were investors.) While Scott and I don’t spend a lot of time together, we’ve both been part of the evolution of the Boulder startup community going back to the late 1990s.

In 2006, Google bought @Last (makers of SketchUp). That was the beginning of Google’s presence in Boulder, which is now around 1,000 people on a new, very nice, and well-integrated campus in the middle of town. Scott and the Google team have always been great corporate citizens of the startup community, offering up their larger event space on a regular basis, participating in, and sponsoring, many of the local startup events over the years, and generally just being a constructive and healthy part of the mix. Google’s continued expanded presence in Boulder is a positive reflection on the overall startup community and their new campus is a really nice addition to our little city in the mountains.

NCWIT (National Center for Women & Information Technology) has long been a hidden gem of Boulder. I got involved shortly after it was founded in 2004 and became the board chair in 2005 (which I served as until I resigned all my non-profit board positions at the end of last year.) I’m still deeply involved and it is a major initiative of the Anchor Point Foundation (the foundation that Amy and I run.)

Physical office space at CU Boulder has always been a struggle for NCWIT. When the organization was small, it fit nicely in a corner of the CU Roser ATLAS Center on the second floor. Amy and I were appreciative of this and sponsored the bathrooms on this floor of ATLAS. As NCWIT grew, they crammed into a small space, then overflowed it, expanded a little, but then lost it in a mysterious space shuffle that I’ve never really understood. Eventually, NCWIT moved over to some old space in the engineering building, but the space was poorly configured, had no cell signal, and wasn’t secure.

At the beginning of 2017, Lucy Sanders (NCWIT CEO) and I started looking for other space in Boulder. We tried to get different space on CU’s campus but were unsuccessful. We had a few near misses with commercial space, but either the economics didn’t work out or the space wasn’t right. Last summer, Google Boulder engaged as their new campus was opening up. A few weeks ago, NCWIT moved into their new, long-term home.

I’m incredibly appreciative for what Google Boulder has done here for NCWIT. It makes me extremely happy to see a #GiveFirst approach from Google in our startup community, along with the extensive support for NCWIT. It’s always nice to be part of an organization that is on the receiving end of this kind of generosity, especially one as deserving as NCWIT.

Scott, Google, and the rest of your team at Google Boulder – THANK YOU!


A few weeks ago I read Jaron Lanier’s Ten Arguments for Deleting Your Social Media Accounts Right Now. It helped consolidate some thinking on my part and I sent a few copies out to friends who I knew would have thoughtful and interesting responses. One that came back is very worth reading as it has a healthy critique as well as some personal reflections. The note from my friend after reading Lanier’s book follows.

He makes a reasonable case (obviously with a lot of room to dispute individual points) that social media is “bad” in general and a source of concern. Some of it is old hat but the way he puts it together is certainly helpful. It seems like it would be good if a lot of people read it.

I had two major concerns with it structurally. First, he positions the book as making arguments as to why *the reader* should delete his or her accounts. But as is common these days, it conflates reasons that are self-interested with reasons that might justify a “boycott.” Many of the arguments are not about how the use of social media affects the reader directly as an individual, but rather its systemic effects. Even the economic argument doesn’t work individually – even if I’m a gig economy person, it does not hurt my prospects to use social media, it’s that the BUMMER business model exists at all that causes the problem. It’s all the rage of course to talk about boycotting anything that has any secondary effects we don’t like, but it rarely works, especially as we realize everything affects everything else, which is why people in Boulder who are concerned about CO2 still drive up to the mountains constantly just for fun. So I thought this really weakened the argument that he does not separate the two things. It’s really Three Arguments why you should delete your social media accounts and Seven Arguments why you should Boycott them.

The second concern is that he conflates Google with social media. Last I checked, no one uses Google Plus. Yes, Google has an advertising and manipulation-oriented business model, but it’s extremely different from Facebook and Twitter. I find the ads Google gives me generally useful, and I don’t see Google making me more of an asshole than I already am. It certainly does not make me sad. Yes, search does have the effect of causing SEO and content-poaching and all that stuff, so this distinction connects to my first point. I think the book would have been better if he had made a more clear compare/contrast with Facebook. I do worry that he is a Microsoft employee and he has a Google-is-the-enemy bias. I’d be very open to hearing how Google is bad for me because I have thought about this and I don’t see it (other than the same things that happen when I pass a billboard on the highway or whatever). I also like Chrome Mobile’s news feed – it’s very much tuned to things I find interesting (cosmology, AI, poetry, etc.) in a way that a news site like the NY Times, which thinks that POLITICS is what is important (just like the MSM) – he talks about religion but does not connect the dots that the MSM have elevated politics-is-the-most-important-thing into a form of religion.

From a personal perspective, in the past year, I went through a couple of transformations regarding Facebook (I don’t use Facebook and never really have). The first was after the election I realized I had gotten caught up in the politics-is-important cycle and was posting frequently on it. At some point, I realized I had been sucked in, and mostly stopped posting on current politics. That took a month or two. Then I had a run-in with a particular individual on something controversial I had posted, and it made me realize I too had been sucked into making controversy and drama there. My approach now is only to post things I think my friends will find funny (NOT political satire) or that offer an update on my life. Yes, I mostly post positive things, but generally not competitively. Instead of commenting I just Like posts, or just read them and move on. I mostly ignore the politics or I just smirk at how absorbed and overconfident everyone is. I probably waste a little more time on Facebook than I would like, but I do find that scrolling through stupid dog and cat and political posts and all that sometimes leads me to a post I am really glad I saw. So, noise to signal is high but really what isn’t?


Over the past 25 years, I’ve invested in many startups that sell products to large enterprises. Many of these companies end up either creating or helping to create a new category. As the startups (or the category) become visible, they inevitably attract the attention of industry analysts, who write reports on the categories and the startups as part of the industry analysts’ business.

Engaging with analysts can result in significant investments of time, effort, and capital on the part of the startup. The choice is a complicated one since startups are often challenging the status quo and industry analysts, while well-intentioned, don’t necessarily have a full grasp of the underlying industry changes taking place until well past the point that changes – and resulting trends – become obvious.

One of our portfolio companies recently engaged with an industry analyst for the first time with a very disappointing outcome. In this case, the company has created and is leading a new category. As a result of growing their customer base quickly, they were invited to participate in an analyst evaluation. Having invested little in relationships with this specific industry analyst, the company was hesitant, but the study was directly relevant and the industry analyst conducting the evaluation was prestigious, so the company decided to participate.

Unfortunately, it quickly became clear that the analyst conducting the review simply didn’t understand the problem being solved or why this company’s solution was so disruptive to the other vendors. The outcome was a deeply flawed report. In retrospect, the company would have been better off if they had never gotten involved.

While it’s easy to say “oops” and move on, this company will now have to deal with this report for a while in competitive situations. Rather than be pissed off about it, our feedback was to use this as a learning moment in the development of the company, figure out why this happened, and determine what could be done differently in the future.

Several of the issues were exogenous to the company, but one big one was under the startup’s control. And, in all cases, the startup should have been much more forceful about their perspective on each issue. The specific issues follow:

The terminology was loosely defined by the analyst. Big shifts in technology are often interpreted at first as evolutionary, not revolutionary. It was notable that several of the “leading” companies in the report introduced their products over a decade ago, well before the category being addressed in the report was even invented. As a result, the younger companies approaching the problem in a completely new way were ranked poorly because the analyst missed the real value to the customer.

The analyst didn’t behave like a customer. In this product category, most customers perform an in-depth analysis of vendor capabilities through a thorough review based on their customer’s buying criteria before deciding on the solution. This analyst didn’t feel like the study warranted a deep look and used vendor demos instead. This eliminated the opportunity for the analyst to understand the customer’s perspective and to compare and contrast the different solutions being evaluated. All decisions and scoring were left to vendor claims (also known as “marketing”) while operational aspects of the customer, and how the various products addressed them, were ignored.

The analyst went wide instead of deep. The magic of this company’s product and the new category they have helped pioneer is a result of focusing on a very specific, yet critically important aspect of a broader problem. The analyst either didn’t understand this or didn’t focus on it and included a wide range of product capabilities, many of them irrelevant to the problem being addressed, in the evaluation. As a result, the study favored broad tools that covered more surface area (mainly from very large, established technology vendors), but had less specific capabilities, especially in the new product category being addressed.

The company failed to fully engage the analyst. Since the company didn’t have a fee-based relationship with the industry analyst firm, there was no long-term relationship. To young companies, paying analyst fees can feel like extortion, but it’s an essential part of engaging with and helping the analysts to better understand your product and how it’s different, especially when you are leading the creation of a new category. In this case (as in many others), the established vendors of broad products had spent years shaping analyst opinions. Even though these broad products didn’t compete effectively in the new category, their relationship with the analyst, who in this case relied on marketing information rather than real product engagement, won the day.

If you sell a product to large enterprises, neglect analyst relations at your peril. I generally categorize this activity in the same bucket as PR, even though they are different functions and often driven by different leaders in the company. Don’t assume that the industry analysts are all-knowing. Instead, start early and feed them regularly or risk having large, established companies win at this game.


Recently, I was talking to a CEO of a company I’m on the board of. We were discussing a problem in the category of something new Is fucked up in my world every day

He gave me a great idea. He apparently plays a game with his young (I think around 10 years old) daughter. When they are sitting around in the evening, she occasionally says “Daddy, give me a CEO problem.” He does, she thinks about it a little, and then gives him a solution. He suggested to me that this often helps break him out of whatever thought rut he is in given how wacky and creative the answers typically are.

Unfortunately, I don’t have a daughter (or a son). While I have two golden retrievers and I’m a practitioner of rubber duck debugging, I don’t think this works as well as what my friend is doing. Oh – and I’m not a CEO, although the list of CEO problems that I’m exposed to is pretty long.

The next time you have a CEO problem (which will likely be in the next seven minutes if you are a CEO and awake), try to think about it through the lens of a 10-year-old and see if that gives you any new ideas.


Yesterday’s post Relentlessly Turning Input Knobs To 0 generated a bunch of interesting private comments. It also generated a few public ones, including the link to the article What is the problem with social media? by Jordan Greenhall which was extraordinary.

Jordan asserts that the problem with social media can be broken down into four foundation problems.

  1. Supernormal stimuli;
  2. Replacing strong link community relationships with weak link affinity relationships;
  3. Training people on complicated rather than complex environments; and
  4. The asymmetry of Human / AI relationships

He then has an essay on each one. The concept of supernormal stimuli is straightforward and well understood already, yet Jordan has a nice set of analogies to explain it. Tristan Harris and his team at the Center for Humane Technology have gone deep on this one – both problems and solutions.

I found the second essay – replacing strong link community relationships with weak link affinity relationships – to resonate with something I’ve been experiencing in real time. As my weak link affinity relationship activity diminishes (through lack of engagement on Facebook and Twitter), all the time I spent on that has shifted to strong link community relationships. Some of these are in person, some by video, some by phone, and some by email, but they are all substantive, rather than shallow (or weak.) I also find that I’m having a wider and deeper range of interesting interactions, rather than a continuous reinforcement of the same self-affirming messages. And, I’m more settled, as I’m not reacting to endless shallow stimuli or interacting with lightweight intention. And, my brain feels like it has more space to roam.

The third essay – training people on complicated rather than complex environments – totally nailed it for me. Ian Hathaway, my co-author on Startup Communities 2, has been working deeply on how startup communities are complex (rather than complicated) systems. This is a central theme of our upcoming book and the contrast between a complicated system (having a finite and bounded (unchanging) set of possible dynamic states) and a complex system (having an infinite and unbounded (growing, evolving) set of possible dynamic states) is a really important one. I loved Greenhall’s conclusion:

“In the case of complexity, the optimal choice goes in a very different direction: to become responsive. Because complex systems change, and by definition change unexpectedly, the only “best” approach is to seek to maximize your agentic capacity in general. In complication, one specializes. In complexity, one becomes more generally capable.”

He then goes on to define social media as training humans to navigate a complicated system, taking time away from us “training our sense making systems to explore an open complex space.” His examples of how this works in the context of Facebook are excellent.

While the asymmetry of Human / AI relationships is nothing new, the Ke Ji / AlphaGo / AlphaGo Zero story is a reminder of what we are contending with. I loved:

“The Facebook AI is Alpha Go. The equivalent of Alpha Go Zero is a few minutes in the future. We need to get our heads around the fact that this kind of relationship, a relationship between humans and AI, is simply novel in our experience and that we cannot rely on any of our instincts, habits, traditions or laws to effectively navigate this new kind of relationship. At a minimum, we need to find a way to be absolutely dead certain that in every interaction, these gods of social media have our individual best interests in mind.”

I didn’t expect this treat to come out of my blog post yesterday, but it’s part of why I blog. And I doubt I would have found it scanning my social media feeds.


I’ve got a lot on my plate. I always do. Presumably, I like it this way because I’d change things if I didn’t. And yes, that’s continuous fodder for conversations with my therapist and with Amy.

I have always tried to ignore the macro, especially short-term dynamics, in the context of my work. I collect a lot of data and like to be well informed. I get this data from lots of different inputs. I regularly play around with the volume on the inputs as well as try different inputs.

One of my key inputs is reading books. I read 50 to 100 books a year (the number seems to be steadily increasing as I get older.) It’s a great joy of mine to sit and read, especially stuff friends recommend to me. I read across all categories and am game to try anything. And I’m willing to quit something I’m not enjoying.

A week ago I read Jaron Lanier’s Ten Arguments for Deleting Your Social Media Accounts Right Now. While it had a few annoying characteristics (I didn’t love his forced acronym for the BUMMER machine), the insights from it were right on the money. I let it roll around in my head the past week as I considered my own behavior over the last six months.

Basically, I’ve turned down the input knobs on almost all real-time social media inputs to 0. I no longer look at Facebook or Twitter. I never really got Facebook, so I was a Twitter guy, but since mid-2016 engaging with Twitter has simply made me anxious, upset, jangly, and distracted. By the beginning of this year, I was broadcast only – sending out links via Buffer when I saw something I found interesting – but that’s about it.

Until a few months ago, I still had a bunch of inputs turned on. I had a Daily folder, which I’ve opened first thing in the morning for over a decade. The contents would periodically change, but it was always something like RSS Reader, some daily reads, Hacker News, my LinkedIn messages, or Google News.

I deleted my Daily folder a few months ago from my browser bar. The inputs were distracting me instead of informing me.

I’ve been using Sanebox for two years to filter out all the noise from my email. I’ve effectively unsubscribed (or – in Sanebox terms, blackholed) thousands of email newsletters. The ones I want to read each day go into my SaneNews folder, so I don’t read them once a day. The number in that folder is now very small and don’t include anything beyond stuff from the tech industry anymore.

While I haven’t deleted my social media accounts, I have turned all the inputs way down. For work, I’m very focused on my existing portfolio, Foundry Group business, and my writing. Beyond that, I’m spending my time with books and with people.

I feel different than I did six months ago. It’ll be interesting to see how I feel in six more months.


My friend Katherine pointed me to the Number Gossip site. If you like numbers, you’ll quickly lose the next hour playing around. Since 49% of the US is taking today off, it seemed like a relevant thing to spend some time on.

For example, did you know that 67 is the only number such that the common alphabetical value of its Roman representation is equal to its reversal (LXVII – 12+24+22+9+9=76)?

Or, did you know that 111 is the smallest palindromic number such that the sum of its digits is one of its prime factors? It’s also the age at which Bilbo Baggins leaves the Shire.

I had forgotten that evil numbers are a number that has an even number of 1’s in its binary expansion. But, I didn’t know what an odious number is (it has an odd number of 1’s in its binary expansion.) Apparently, being evil is related, but the opposite, of being odious.

While we all know that 42 is the answer to the ultimate question of life, the universe and everything as calculated by Deep Thought, did you know it is also the number spots on a pair of dice? It’s also the smallest abundant odious number.

Have fun. Don’t forget to come up for air once in a while.


As I’m already getting lots of out of office messages for people taking this week off, I thought I’d revisit an approach to how to deal with email after a vacation.

In 2011, Josh Kopelman of First Round Capital came up with what, at the time, was what I thought was the best email vacation auto-responder in the history of email. Now, I have no idea if Josh invented this, but I’m going to give him credit for it.

I evolved this in 2014 when I took a one-month sabbatical. If you ever send me an email when I’m on my quarterly one week off the grid vacation, you now get this.

I’m checking out for a vacation until [date]. I’ll be completely off the grid.

When I return, I’m going to archive my inbox so I’ll never see this email. If you’d like me to read it, please resend it after [date]+1.

If you need something urgently, please email [my_assistants_email_address] and she’ll either help you or get you to the right person at Foundry Group to give you a hand.

On [date]+1, I usually get around 50 emails (in addition to my usual email flow of 500 daily emails) that are resent to me. That’s only 50 to respond to, instead of the roughly 3,000 emails I get each week.

It appears that The Atlantic has caught up with this thinking in The Most Honest Out-of-Office Message. I thought the article was fascinating, both in how the writer addressed the issue, but also in the intellectual and emotional tension around it.

Does it make you nervous to think about “Selecting All” on your existing inbox and archiving (if Gmail) or deleting (if Outlook)? While some of my friends do it periodically – as a result of pain or just to get a fresh start on a new year, I like to do the equivalent every quarter when I get back from a full week of an off the grid reset.

Josh – it was a while ago, but thanks for the inspiration. And, for those of you on vacation this week, I hope you aren’t reading this.