Hackers, War and Venture Capital

In my previous post of this series, I discussed the role of military funding in the formation of a ‘genealogy’ of university laboratories, their projects, and staff which produced the conditions for hacking during the 1960s and 70s. As I drafted that post, I found myself drifting into a discussion around the role of venture capital but I have split that discussion into this final post below so as to highlight another important aspect in the study of the role of the university in the development of hacker culture.

Levy (1985) points to the arrival in 1959 of the TX-0 computer as a seminal moment in the history of hacking. The computer had been donated by the Lincoln Laboratory to MIT’s Research Laboratory of Electronics (RLE), the original successor of the Rad Lab and today, “MIT’s leading entrepreneurial interdisciplinary research organization.” Similarly, Eric Raymond points to the arrival at the RLE of the PDP-1 computer in 1961 as the moment that defined the beginning of ‘hackerdom’. Notably, at that time the RLE shared the same building as the Tech Model Railroad Club (TMRC), the legendary home of the first hackers. The history of hacking is understandably tied to the introduction of machines like the TX-0 and PDP-1 just as Richard Stallman refers to the demise of the PDP-10 as “the last nail in the coffin” for 15 years of work at MIT. Given the crucial significance of these machines, a history of hacking should include a history of key technologies which excited and enabled those students and researchers to hack at MIT in the early 1960s. To some extent, Levy’s book achieves this. However, in undertaking a history of machines, we necessarily undertake a social history of technology and the institutions and conditions which reproduced its development and in doing so we reveal the social relations of the university, the state and industry (Noble, 1977, 1984).

The birth of Digital Equipment Corporation

In 1947, the US Navy funded MIT’s Servomechanisms Lab to run Project Whirlwind to develop a computer that tracked live radar data. The Whirlwind project was led by Jay Forrester, leading systems theorist and principle inventor of magnetic core memory (the patenting of which was marked by a dispute between MIT and the Research Corporation resulting in the cancellation of MIT’s contract with the Corporation).

MIT’s Lincoln Lab was set up in 1951 to develop the SAGE air defence system for the US Air Force, which expanded on the earlier research of Project Whirlwind.  The TMRC hackers’ first computer was a TX-0 from the Lincoln Lab with its use of a cathode-ray display borrowed from the SAGE project’s research into radar. Though large by today’s standards, the TX-0 was smaller than Whirlwind and was one of the first transistor-run computers, designed and built at MIT’s Lincoln Lab between 1956-7 (Ceruzzi, 2003, 127). Much of the innovation found in the TX-0 was soon copied in the design of the PDP-1, developed in 1959 by the Digital Equipment Corporation (DEC).

DEC was founded by Ken Olson and Harlan Anderson, two engineers from the Lincoln lab who had also worked on the earlier Whirlwind computer. Watching students at MIT, Olsen had noticed the appeal of the interactive, real time nature of the TX-0 compared to the more powerful but batch operated computers available and saw a commercial opportunity for the TX-0. Soon after they established their firm, they employed Ben Gurley, who had worked with them at the Lincoln Lab and designed the interactive display of the TX-0 which used a cathode-ray tube and light pen. It was Gurley who was largely responsible for the design of the PDP-1. DEC is notable for many technical and organisational innovations, not least that it permitted and encouraged its clients to modify their computers, unlike its competitor, IBM, which still operated on a locked-down leasing model. DEC’s approach was to encourage the use of its machines for innovation, providing “tutorial information on how to hook them up to each other and to external industrial or laboratory equipment.” (Ceruzzi, 2003, 129) This not only appealed to the original TMRC hackers but appealed to many of its customers, too, and led to DEC becoming one of the most successful companies funded by the venture capital company, American Research and Development Corporation (ARD).

The birth of venture capitalism in the university

ARD, established in 1947, is regarded as the first venture capital firm and was “formed out of a coalition between two academic institutions.” (Etzkowitz, 2002, 90). It was founded by the “father of venture capital”, Georges Doriot, then Dean of Harvard Business School, Ralph Flanders, an Engineer and head of the Federal Reserve Bank in Boston, and Karl Compton, President of MIT. ARD employed administrators, teachers and graduate students from both MIT and Harvard. The motivation for setting up this new type of company was a belief by its founders that America’s future economic growth rested on the country’s ability to generate new ideas which could be developed into manufactured goods and therefore generate employment and prosperity. This echoed the argument put forward by Vannevar Bush that following the war, “basic research” should be the basis for the country’s economic growth and both views confirm the idea/ideology that innovation follows a linear process, from basic research which is then applied, developed and later taken into production. However, whereas government was funding large amounts of R&D in universities, the founders of ARD complained of a lack of capital (or rather a model of issuing capital) that could continue this linear process of transferring science to society.

ARD funded DEC after Olsen and Anderson were recommended by Jay Forrester. This led to an investment of $100,000 in equity and $200,000 available in loans and within just a few years DEC was worth $400m. This allowed ARD to take greater risks with its investments: “The huge value of the Digital Equipment stock in ARD’s portfolio meant that the relatively modest profits and losses on most new ventures would have virtually no effect on the venture capital firm’s worth.” (Etzkowitz, 2002, 98). ARD’s success marked the beginning of a venture capital industry that has its origins in the post-war university and a mission to see federally-funded research exploited in the ‘endless frontier’ of scientific progress. It led to the development of a model that many other universities copied by providing “seed” capital investment to technology firms and the establishing of ‘startup’ funds within universities. Most recently, we can observe a variation of this method by the ‘angel investment’ firm, Y-Combinator, which specifically sought to fund recent graduates and undergraduate students during their summer breaks.

Y-Combinator and the valorisation of student hackers

A proper analysis of Y-Combinator in the context of the history of hacking, the university and venture capital is something I hope to pursue at a later date. In this current series of posts discussing the role of the university in the ‘pre-history’ of hacker culture I want to flag up that Y-Combinator can be understood within the context of the university’s role in the venture capital industry. Just as academic staff have been encouraged to commercialise their research through consultancy, patents and seed capital, in its early stage, Y-Combinator sought to valorise the work of students by offering its ‘summer founders programme‘. Similarly its founder, Paul Graham, has often addressed students in his writing and discussed the role of the university experience in bootstrapping a successful start-up company. Graham’s on-going articles provide a fascinating and revealing body of work for understanding the contemporary relationship between students, the university, hacking and venture capital. In this way Y-Combinator represents a lineage of hacking and venture capital that grew out of the university but never truly left because despite recent claims that we are witnessing the demise of higher education as we know it, the university as a knowledge factory remains a fertile source of value through the investment of public money and the production of immaterial labour, something that Vannevar Bush would be proud of.

Series conclusion

This is the last of a series of six posts on the role of the university in the development of hacker culture. These posts are my notes for a journal article I hope to have published soon which will argue, as I have done here, that the pre-history of hacking (pre-1960) is poorly documented and that much of it can be found in an examination of the history of American higher education, especially MIT.

As an academic who works in a ‘Centre for Educational Research and Development’, and who runs various technology projects and works with young developers, I am interested in understanding this work in the context of the trend over the last decade or so, towards ‘openness’ in higher education. Ideas and practices such as ‘open education‘, ‘open access‘, ‘open educational resources‘ (OER) and most recently ‘Massive Open Online Courses’ (MOOCs) and ‘open data‘, are already having a real impact on the form of higher education and its institutions and will continue to do so. My work is part of that trajectory and I recognise that the history of openness in higher education goes back further than the documented last 10-15 years. It is well known that the early efforts around OER, OpenCourseWare and the concurrent development of Creative Commons licenses owes a great deal to the ‘open source’ licensing model developed by early hackers such as Richard Stallman. I hope that in these posts I have shown that in turn, the free and open source software movement(s) was, in its early formation, a product of the political, economic and ultimately institutional conditions of the university. Richard Stallman felt compelled to leave the academy in 1984 as he found that “communism”, a foundational ethos of science as famously described by Merton (1973), was by that time little more than an ideal that had barely existed at MIT since the Great Depression.

This points towards a history of openness in higher education that is rooted in hacker culture and therefore in the commercialisation of scientific research, military funding regimes and the academy’s efforts to promote a positive ideology of science to the public. Stallman’s genius was the development of ‘copyleft‘, in the form of the GPL, which was very influential in the later development of Creative Commons licenses used (and partially developed) in higher education. Through the growth of the free and open source software movements in the last 25 years, the academy has been reminded (and as participants, reminded itself), that the ideal of communism in science forms the basis of a contract with society that can still be achieved through the promotion of openness in all its forms. However, in hindsight, we should be cautious and critical of efforts to yet again valorise this new agenda in science through calls to adopt permissive licenses (e.g. CC-BY, MIT, ODC-by) rather than Stallman’s weapon of scientific communism: Copyleft.

Seminar: Hacking and the University

Hacking and the University

The role of the university in the development of hacker culture

The standard history of hacking begins with the Tech Model Railroad Club at MIT in 1961 and has continued to be closely associated with academic culture. Why is this so and what intellectual and institutional culture led to the development of a ‘hacker ethic’?

This seminar will propose a history of hacking in universities from the early 20th century, taking into consideration the role of military sponsored research, the emergence of the ‘triple helix’ of academic, commercial and government enterprise, the influence of WWII cybernetic theory, and how the meritocracy of academia gave rise to Y-Combinator, the most successful Internet angel investment fund there is today.

Part of the Centre for Educational Research and Development’s Thinking Aloud seminar series.

November 27th, 1-2pm, MB1012

The cost of developing a good idea

How much does a student hacker need to develop a good idea to the point that it attracts further investment?

I’ve been thinking about this recently for a couple of reasons. I was reading the early Y Combinator site, via the Wayback Machine, about how they reckoned on $6,000 per person for their first Summer Founders Program. Each new startup could expect to receive less than $20K (the average is $17,000 / £10,000), with two or three friends being the ideal number of founders per company. The Summer Founders Program was aimed at undergraduate or graduating students.

I’ve also been looking at JISC’s Elevator funding programme, where people working in UK universities and colleges (with a *.ac.uk email address), are able to pitch an idea to receive up to £10,000 funding from JISC.  That’s the same amount of money Y Combinator seeds their successful applicants with. I think the JISC Elevator is a great idea, but looking at the proposals that have been submitted so far, I’m surprised and disappointed that there aren’t any proposals where the money goes directly to students to develop ideas of their own.  Maybe students haven’t been told? I’ll admit I’ve not publicised it at Lincoln, having been busy bidding for other JISC funds (where graduating 3rd year students are the main contributors to the projects) and awarding funds to projects of our own (where students receive most of the money). Still, I feel bad about not supporting JISC Elevator more. I have voted for one proposal.

I asked Alex, an undergrad and co-worker, how much a student who is hacking on an idea all day, every day, needs to live on in Lincoln, and he reckons about £600/month. That sounds harsh to me, so let’s assume they need £800/month and that there are three of them, because after all, if you can’t persuade a couple of friends that an idea is worth working on, then it probably isn’t a very good idea (or so says Y Combinator). On a related note, Google’s Summer of Code provides students with a $5000/£3000 stipend for the summer.

When I first heard about the JISC Elevator, my immediate thought was that the £10K maximum per project isn’t very much to attract FEC costed projects involving staff, but is perfect for offering to students as bursaries. A bursary, as I understand it, is supposed to cover the costs of living, rather than being seen as a wage, so they’re similar in purpose to the GSoC and Y Combinator funds. On our DIVERSE project, almost all of the money received went to paying the fees and bursaries of two MRes students. We are also prepared to contribute a larger percentage of the overall cost. Our recently funded beBOP project is an example of this, with a recent graduate being employed on grade 4, and the funding from JISC covering only 65% of the overall cost, compared to the maximum 80%.

I’ll admit, I don’t really understand how FEC works and where a lot of the money actually goes, but for the kinds of projects that the JISC Elevator is trying to attract, as well as JISC’s Rapid Innovation calls, I do wonder whether the GSoc or Y Combinator model of funding is a more cost-effective one. Pay students to hack over the summer, with a member of staff overseeing their work and call that the institutional contribution. £10K will pay for three students to hack over the summer, travel to a conference to talk about their work and pay for some servers on Rackspace for a few months. The tools to develop software in the early stages are cheap (a basic Linux stack on Rackspace is £7/month and there are enough open source tools available to explore ideas and develop prototypes, even if the ideal tool happens to be a proprietary one.

At Lincoln, we recognise that, given the opportunity and mentorship, undergraduate students have much to contribute. They’re not simply consumers of education. Like other universities, we’ve been running funding programmes each year that fund students to work on a research project with a member of staff over the summer. At Lincoln, it’s called UROS, the Undergraduate Research Opportunity Scheme. The Student as Producer UROS call was announced a few days ago. The LNCD group, which I co-ordinate awarded five projects £1000 each last week, which focus on the use of technology for education (more info on those projects soon). For the UROS and LNCD funded projects, almost all of the £1000 goes on undergraduate student bursaries. In my experience, undergraduate hackers can produce good work. Work that’s worth funding. Y Combinator thought so, too, and they’re now the most admired angel fund among young hackers. Each Y Combinator funded start-up is now guaranteed $150,000 as follow on funding by another investor. If you go Wayback to the first Summer Founders Program FAQ, you’ll see this:

Why are you doing this?

Partly because we feel guilty that we all got rich almost seven years ago, and still haven’t yet given seed money to new startups; partly because we think it is an interesting hack; and partly because we think it may actually make money.

We suspect that students, and particularly undergrads, are undervalued. Twenty years ago the idea of grad students starting companies would have seemed odd. Not after Yahoo and Google. And if grad students can do it, why not undergrads too?

I agree. Undergraduates can do it and I think institutions, together with JISC, should be thinking about our own Hatchery for Hackers.

The university as a hackerspace

Developers at Dev8D

Developers at Dev8D

I spend most of my time working with students and recent graduates. At first Alex, then Nick and Jamie, and in a week or two, Dale and Harry will join us. Dale and Alex are finishing up their final year in Computer Science and work as part-time Developers in ICT Services. Nick, Jamie and Harry graduated last year from studying Computer Science and work with me on JISC-funded projects. They’re all in their early to mid twenties. I learn a lot from them. Sometimes they make me feel old. I’m only 38.

In a year or so, they’ll probably move on to other things. Alex and Nick already have their own company and a business plan. Jamie wants to do a PhD. If I can secure us interesting and useful work, maybe Dale and Harry will stay on for a while longer, I don’t know. I hope they will stay but it needs to be for a good reason otherwise I would encourage them to look for new challenges.

These days I wonder how we (‘the university’) can support young hackers like the bunch I work with. Career progression for  developers working in universities is not great. Paul Walk recognised this as does the JISC-funded DevCSI project, run by Mahendra Mahey. Without their work, which highlights the importance of local developers, I think the HE sector would be a pretty barren place for hackers to commune. No Dev8D, no DevXS, fewer hack days and developer workshops. Along with several other people, I was invited to be on the Steering Group for DevCSI today, which I am very pleased about, and I look forward to working with the project more closely in the future.

There are a few things I’d like to focus on with DevCSI, based on my experience working with young hackers: the first is about how hackers learn. As I see it, this requires research into the history of hacking in universities, the role of undergraduate and graduate students in funded research, from ARPANET to Total ReCal, hacking as a cognitive craft that has its own learning communities and learning environments, that creates tools which help improve the effectiveness of learning and how those tools eventually shape the tools the rest of us use to learn.

The second thing I’d like to focus on, is how universities can learn from what we see happening in hacker culture: new reputational models, fablabs and hackerspaces, peer-production, and new methods of funding. On this last point, I’d like to develop an academic programme that attracted hackers and graduated start-ups. I think the Y Combinator model is a good example of this already happening, where students and recent graduates are receiving a little funding and lots of support, while asking for just a very small cut of the company.

Finally, I’d like to think about how we can break down the distinction between developers working in professional services and developers working on research projects and remind ourselves that we all work in universities: autonomous institutions for research, teaching and learning, and that when companies want to instill a culture of innovation, they often emulate the research culture of universities. I’d like to work towards developers in HE knowing there was the opportunity to be paid the same as professors when they clearly add similar value to the institution, which I think is easily possible. I see no reason why  great hackers  – experienced software craftsmen and women – working in universities shouldn’t be paid £100K or more, just as professors and other senior staff can be. Of course, working in a university, they would teach other hackers, run software development projects, contribute to the strategic direction of the university and produce superb software for the institution, too. Would they be on an academic or a non-academic contract? I don’t know. At that level, I don’t think it matters, but at the junior and middle periods of their careers it remains a divisive distinction that affects people’s aspirations.

Being great hackers, they would attract students that aspire to be great hackers, too. Just as I decided which graduate school to study at based on the reputation of a single professor working there, so young hackers would want to work with and learn from great hackers in universities, even more so if their programme of study included a Y Combinator style opportunity for angel investment.

I’d like to see an academic programme led by experienced software craftsmen with reputations to match, where students from different disciplines spend their degree in a university space that resembles a hackerspace or dojo, working together on ideas of their own under the guidance of more experienced staff, leading to potential angel investment at any point in their degree. Those that don’t get funded, leave with a degree, a valuable experience and a network of alumni contacts. Those that do get funded are given the support they need to develop their work into a real product or service. Sometimes, it might be one that the university would use itself, but not always. Over time, successful alumni would help attract more students to the programme, developing a culture of hackers and successful startups attached to the degree programme.

What excites me about this is that it’s a mixture of what universities always say they are about: research, teaching, learning and enterprise, but it recognises that those processes are changing and that hackers are already developing a model that is replacing these functions of the university: the opportunities for learning, collaboration, reputation building/accreditation and access to cheap hardware and software for prototyping ideas, can and are taking place outside universities, and so they should. However, I think that university culture is still a place where the hacker ethic (respect for good ideas, meritocracy, autonomy, curiosity, fixing things, against technological determinism, peer review, perpetual learning, etc.) remains relevant and respected. A university is a place where people come to learn from each other and we should be creating these new spaces and programmes that recognise the value of developers in universities.

Like any programme of study, work and investment, it needs careful thinking about how to set it up right, but from where I stand, it feels like a gaping hole in higher education that needs to be filled. Do you know of any examples that are already running? A ‘MIT Media Lab lite’ could be close to what I have in mind, but I have no experience of how it’s run and whether it breaks down the distinction between academic and non-academic staff to the extent I have in mind.