Funding the commons and closing the innovation chasm
How are we going to fund the steps we need to take to promote #DigitalEquity and build the circular economy around Community Owned Internet Networks? The Broadband Institute Foundation believes that the best way to advance #DigitalEquity is to take steps towards remaking the Internet as a public commons. In another post we talked about open access fiber. Companies such as Net Equity see fiber as a secure, long-term investment.
Following the electrical grid to deploy fiber, financing fiber like a utility and distributing bandwidth through an open access network can modernize the electrical grid and bridge the digital divide with one investment….
Treat fiber like a public utility
- Key principle: Separate infrastructure from services through contractual obligations.
- Financing construct: Utilize a well-established project finance construct (e.g. build, operate, transfer) under ~20 your contract term.
- Source of equity: Obtain capital commitments from infrastructure funds seeking a repeatable, predictable and cash yielding business model.
Net Equity
A Bit of Internet History
Before the privatization of the Internet during the 1990s, as discussed by Ben Tarnoff, there was NSFnet, which existed as a public commons. The following quote is by David Bollier, author of The Wealth of the Commons, a World Beyond Market & State.
The end-to-end principles of the Internet and its shared protocols constitute a vital infrastructure for creating countless online commons. This lecture gives a brief overview of this history, with readings by Lawrence Lessig, Richard Stallman, Eben Moglen, David Bollier, Elinor Ostrom and Charlotte Hess.
Listen to this podcast by David Bollier
Decentralized investment structures for public goods and Commons
Transcript
The transcript has been copied below to enable website search ability.
0:02really excited to see all of the funds
0:03that are coming together
0:04to talk about what they’re doing to talk
0:06about how to build structures for
0:09better funding mechanisms and not just
0:10funding but
0:11optimization mechanisms so this is a
0:14crucial difference that i think is
0:16tremendously overlooked in our macro
0:18system worldwide nation-state
0:21based
0:22grant funding oriented structures which
0:24is primarily a funding system with no
0:27good feedback mechanism and no good
0:29feedback loop and therefore it cannot
0:31optimize well so what i hope to do in
0:33this talk is to talk about
0:36why we need optimization to be much
0:37better what do the feedback loops
0:39represent i’ll use venture capital as an
0:41example to kind of
0:43learn lessons from and then i’m going to
0:45talk about a bunch of the work that
0:46we’re doing to build these things that
0:47we’re calling network funds or blue
0:49funds um
0:51the blue and green dichotomy you’ll
0:52you’ll get a sense that whole like
0:54gradient that you see at the top that i
0:56use in a bunch of talks has a meaning
0:58we’re going from the science-oriented
1:00part the early
1:01long-term
1:03development of knowledge to translate
1:05all of that science into technology that
1:06we can put in products and we can use
1:08and so on so green funds work really
1:10well venture capital funds work really
1:12well
1:12we need better structures for blue funds
1:14or like these teal in-between funds
1:18so the structure of talk is this uh
1:19where we’ll go through
1:21capital allocation as a is a distributed
1:23optimization problem and and why um
1:26capitalism works really well and why
1:27venture capital specifically works well
1:29um we’ll dive into the detail of uh what
1:33makes venture capital work well and then
1:34we’ll try to back into
1:36what must network capital look like how
1:39must it operate what are the components
1:40that we need to
1:42build into these systems to make them
1:43work
1:45so
1:46what’s going on in the in the
1:48broader landscape is that there is a
1:52infinitely dimensional um but maybe
1:54sparsely populated
1:56space of variables to optimize at any
1:58given moment in time and maybe what
1:59humans care about is a much smaller
2:01space than that but still extremely
2:03dynamic extremely large extremely
2:06difficult to measure um and extremely
2:08difficult to make predictions about uh
2:10and so what’s going on in our broader
2:12coordination problems and broader
2:13coordination systems
2:15and this affects us every
2:16from the
2:18local scale of like how you’re operating
2:20right now as a human to how the entirety
2:22of our species is operating at a grand
2:24scale um it’s all about optimization
2:26systems
2:27all right
2:28disconnecting great specific flickering
2:31thank you
2:32um
2:33yeah seem like i think the new macbooks
2:35have not yet optimized their
2:37uh their new hdmi ports so the dongles
2:40probably are still good um
2:42so the the optimization process that we
2:44um
2:45that we have to like get to is how to
2:47allocate and organize our resources to
2:50move through this these these surfaces
2:52that are extremely large extremely
2:54dynamic um and how to do so as
2:56effectively as possible
2:57um in reality like no group gets to look
3:00at the whole surface you’re usually only
3:02looking at a patch you’re looking at a
3:03small area and you can maybe
3:05reason about locally about what’s going
3:07on and you can maybe summarize that
3:09information and propagate it to other
3:10people and i’m sure that all of you have
3:12had trouble convincing somebody else of
3:14something you know to be true based on
3:15all the observations you have seen so
3:18you know how hard it is to then have
3:19chains of reasoning and chains of
3:21of descriptions so imagine how hard it
3:23is to get our systems to
3:25coordinate on some local information and
3:27aggregate that to then make decisions
3:28somewhere else about about reality
3:31this is actually not quite true in
3:32reality like you’re looking at that most
3:34of the time and um the reason that
3:36capitalism has worked so well over the
3:38last 100 few hundred couple hundred
3:40years is that it is a structure that
3:43works well at the at various different
3:45scales in this optimization surface um
3:47it’s similar to to biological evolution
3:50where it enables a run time where new
3:53organisms can emerge
3:54they need resources so they are forced
3:57to optimize if they don’t optimize they
3:59will die
4:01and if they find a good sweet spot they
4:04can grow
4:05potentially reproduce potentially
4:07continue operating and so on and if at
4:09any point they lose track of reality
4:12they’ll weather and die and that’s
4:13extremely crucial to this optimization
4:15process working
4:16the other thing that’s crucial it is
4:18that is that it’s open and
4:19permissionless that at any point a new
4:21entity can can start anywhere in the
4:23optimization surface and try to optimize
4:24from there um using local information
4:27ideally
4:28so
4:29putting this together like imagine how
4:31difficult it is to do central planning
4:32about around a surface of this magnitude
4:35when all the parts are only looking at a
4:36local patch and then trying to convince
4:38each other to create one set of policies
4:41and apply them everywhere even with
4:42supercomputers now even um you know
4:45massive scale ai systems will have
4:47distributed algorithms that to reason
4:49about
4:50all of the complexity in a local way and
4:51apply local policies
4:54so
4:55part of what varies in in the capital
4:57structure is that what looks like this
4:58to some some group might look very
4:59different in another area of the market
5:01or another area of the world or another
5:03industry and so you have all of these
5:04different entities looking at some small
5:06patches and tons of competition in that
5:08area
5:09as the surface is changing all the time
5:11as this surface is encountering new
5:13technologies that totally change the
5:15structure
5:16now
5:17what makes all of this work is that
5:19there is a system of message passing and
5:22value flows between all of these parties
5:24to help organize and
5:26help optimize this large large structure
5:29uh cool so um
5:32one of the things that i’m extremely
5:34interested in helping optimize is this
5:36innovation chasm it’s it’s about um what
5:38happens between coming up with good
5:40ideas and trans translating those good
5:42ideas into technology that we can put
5:44into products and broadly diffuse
5:46because i think that our super the
5:47superpowers that technology grants us um
5:51are uh or and the superpowers that
5:53science maybe illuminates as potential
5:55uh only get realized when you do an
5:57enormous amount of r d um development uh
6:00to get it get it out there
6:02and there’s a lot of like lacking
6:03optimization in that area there’s no
6:06good structure for optimization in that
6:07in that part
6:09um
6:10the
6:11incentive structure on the right is kind
6:12of like the capitalist infrastructure
6:13and that tends to work well for whatever
6:15the optimization service is doing but
6:17it’s very brittle when it comes to what
6:19capital values so if the capital
6:21structure’s value is something that
6:23humanity doesn’t really value then you
6:25get bad results and you can probably
6:26look at
6:27the state of the world today as a good
6:28example of like what happens when you
6:30have capital systems misaligned with
6:32human values
6:33so maybe you know rewriting the
6:34structure there and orienting capital to
6:37maximize for what humanity does value
6:39might yield to much better outcomes
6:41on the flip side on the academic side
6:44we have all kinds of
6:45incentive structure problems as well
6:47where it there’s
6:49academic credit um and that works as a
6:51robust incentive structure
6:53but it tends to be to create an
6:54environment that is very difficult to
6:57experiment in where
6:59you
7:00are punished for trying things that may
7:02not work out you’re um punished for um
7:05uh publishing things that or or not just
7:07publishing but but trying things that
7:09will um that are potentially promising
7:11bad directions and you’re certainly not
7:12rewarded for translating
7:14any of your ideas and thoughts and and
7:16um and good uh publications into
7:19technology that that creates some broad
7:21right utility
7:23uh so let’s look at venture capital for
7:24a moment
7:25uh these like so-called green funds
7:29i like um
7:30should i unplug and replug
7:32or is that too dangerous
7:37yeah i’ll try it
7:42it might be
7:44it’s worked fine in other times i think
7:46it might be cable um
7:48can you live with that is that okay all
7:50right
7:51um so let’s talk about venture capital
7:52um the
7:54basic structure of how venture the
7:56venture capital industry works is this
7:58where there’s you know instead of there
8:00isn’t it’s really a kind of service
8:01providing industry that is helping solve
8:03an optimization problem within within
8:05the broader structure there’s um a lot
8:08of um money in the world um in large in
8:11managing large massive massive scale
8:13funds that don’t have the visibility and
8:16ability to look at some local area in
8:18the optimization surface and then on the
8:19other side there’s entrepreneurs and
8:21companies and builders who need capital
8:23to be able to flesh out their ideas and
8:25get things out there
8:26venture capitalists
8:28venture capital structures are different
8:30funds
8:31that find a spot in the in the
8:33optimization surface
8:34raise capital from other investors and
8:36deploy it into startups and so they
8:38serve as like this really useful signal
8:40processing node in between to use local
8:42information to try and deploy deploy
8:44capital and in reality it’s not just one
8:46entity there there’s a whole set of
8:47networks in there
8:49um
8:50in terms of i won’t kind of like recap
8:52recapitulate the success of venture
8:53capital but you know just look at some
8:55graphs of you know most of the
8:56technologies that we use today
8:58um have been you know in some way shape
9:00or form influenced by venture capital um
9:02and so on
9:04the
9:05the structure has this kind of what i
9:07like to call a ladder
9:09or staircase structure where you have
9:11companies being able to raise different
9:12amounts of funding at different stages
9:15so that you can try out things at
9:17smaller scale and if they’re working if
9:19you get a good signal then you can start
9:21scaling and so this entire structure
9:23this ability of of the the market to
9:25optimize for these pieces
9:27is part of what enables a venture
9:28capital
9:29success story that you can have many
9:31different funds operating at these
9:33different stages with different
9:34perspectives and thesis about how to do
9:36it right um and and in reality you have
9:39thousands to tens of thousands of
9:40entities doing this um to to make this
9:42work really well
9:44so again it’s a signal processing
9:46problem it’s about how do you aggregate
9:48information in that massive scale
9:49optimization surface bring it together
9:51into a resource allocation
9:53distributed system
9:55uh there’s enormous amount of money
9:57flowing through it it’s been very
9:58successful in producing a lot of
9:59corporations and a lot of technology
10:01and there’s an enormous amount of of
10:03capital in in the system
10:05and there’s tons of funds um there
10:07really are um it’s a
10:09super vibrant ecosystem that enables new
10:11fund managers to emerge to it to build
10:13new structures deploy capital and if
10:14they’re successful
10:16and this is the key part if they’re
10:17successful
10:18they tend to be able to raise larger
10:19funds and larger funds and larger funds
10:21and the moment they stop being
10:22successful they raise smaller and
10:23smaller and smaller funds or die and
10:26that’s a critical part of making this
10:28whole optimization process work
10:30um
10:31so i kind of like how this model for for
10:32how the picture works which is you have
10:34a startup or a company of some sort and
10:36selling some products into a market and
10:38creating cash flow now it could try and
10:40sell its stock in the broader public
10:42stock market um but the broader public
10:44stock market doesn’t have enough
10:45information gathering about that
10:47particular company to be able to make
10:49really good decisions um about how to
10:51run it how to help it and so on
10:54uh so you end up in a structure where
10:55instead um that company will
10:58sell part of its cash flow or potential
11:00future cash flow um to investors that
11:02can help it in that stage of its life
11:05and so that means think of like
11:06fractionalizing the potential future
11:08outcome it’s not the outcome today it’s
11:09not the cash flow now it’s a potential
11:11future cash flow as sold to another
11:13market of investors the public markets
11:16um
11:17and they sell that potential prospective
11:20measure
11:22at that moment in time and so this is
11:24really a predict what this is doing is
11:26creating a prediction market on the
11:28um future cash flows of these
11:30corporations
11:32and you know there’s of course many ways
11:33to like realize that cash flow through
11:35dividends liquidations acquisitions and
11:37so on but usually it ends up being
11:39selling the stock in a public market at
11:41some point
11:42which by the way is kind of a bug in for
11:44other reasons but i won’t go into that
11:46now let’s look at a vc fund itself so
11:47this vc fund works by
11:50funneling cash flow
11:51funneling money into companies taking
11:53some pieces from them and they
11:55themselves have
11:57a set of investors and vcs raised from
12:00other investors using the same exact
12:01composable structure so they can
12:04float
12:05this trend this
12:06the transactions of a vc fund are
12:09composable in a super neat way they um
12:12you have a transaction of cash flow
12:13going out in one direction and kind of
12:15getting stock back and you have these
12:16same exact transactions on the other
12:18side to raise capital uh the vc fund
12:20gives stock and uh gets cash on the
12:22other side so you have this structure
12:23where you can feed forward cache and
12:25back propagate
12:26stock and you have this signal
12:28processing unit
12:30and the ease compose super nicely and
12:31super neatly so you you have many uh fun
12:34defund structures
12:36or vcs with with uh angel scout funds
12:38and these you know stock and stagger and
12:40so on there probably aren’t that many
12:41layers but but what i want to sort of
12:44convey to you is that there’s something
12:45else that we’re building a lot of that
12:47looks a lot like this
12:48um that i’ll get to in a moment
12:51but anyway part of what’s doing this is
12:53that the fun the funds upstream are much
12:56larger in size and in capital
12:58range and
12:59these these structures enable um
13:02more fine-grained allocation of
13:04resources and capital across across the
13:05board
13:06and so this is what helps build this
13:08massive staircase with tons of you know
13:10thousands of thousands of of vc funds
13:12that are allocating value this way
13:15so there’s a lot another thing that
13:16looks a lot like this
13:17and these are neural networks um they’re
13:20i don’t think this is an accident i
13:21think this is uh something fundamental
13:23about these composable units that can
13:25over time develop and and help bring
13:27optimization they’re not doing proper
13:28you know feed forward and back properly
13:30in the way that neural networks do but
13:31these are kind of
13:32bayesian reasoning type
13:34entities with
13:35an evolutionary structure encoded into
13:37it which is precisely what you need to
13:39optimize the surface
13:41now let’s talk about uh network capital
13:43i’m running out of time so um the the
13:45key thing that i wanted to to kind of
13:47impart upon is that decomposition of the
13:49venture capital system as an integrated
13:51system with that composability of the
13:53structure so we’ll look at um blue funds
13:55and network capital
13:56and the pieces of those and to see how
13:58how to compose something like that in um
14:01to create funding structures for public
14:03goods
14:04so going back to this problem we we what
14:07we want to do is create incentive
14:08structures and funding things that can
14:09be successful in that area of the of the
14:12market now there isn’t good cash flow
14:14optimization there or this would have
14:15been a solve problem through vc so we
14:17need something else that can that we can
14:18use as a way to
14:20derive value here
14:22the good news is that there’s enormous
14:23amount of funding that is deployed
14:25through to science and
14:27in charity and so on but the bad news is
14:29there aren’t that many funds
14:31it’s it’s highly centralized
14:33the further up you go the fewer entities
14:35there are so it is not a good
14:37optimization structure and and worst of
14:39all is that it’s not composable you
14:41don’t have a good way to stagger funds
14:43and to propagate funding across
14:45different different
14:47to build a network of funding structures
14:49so going back to this we want something
14:50like this we want to have a version of
14:53this that works to do
14:55this kind of public good funding not
14:57just funding
14:58the creation of a private cash flow as
15:02the stock market values but a funding
15:04structure that can enable the broad
15:06production of public goods
15:08and ideally with a feedback mechanism
15:10that that enables a prediction market to
15:12form in to optimize that surface
15:16so let’s look in detail at public good
15:17organizations then they tend to kind of
15:20create a public good and they sort of
15:21give it to to the broad public
15:23and usually what comes back in the other
15:25direction is kind of like some sense of
15:27mission accomplished some goal
15:29has been done that is often one of the
15:31primary reasons why people do this kind
15:32of stuff the other parts is thanks or
15:35broadly um
15:36broadly recognized credit uh for doing
15:38this kind of a good impact
15:40some other kind of measures and so on
15:41but it’s probably you know these kinds
15:43of these kinds of structures and there
15:44is some amount of fundraising that
15:46happens where those organizations can
15:48turn to a set of uh donors or you know
15:50investors in the in the outcome of
15:52public goods
15:53and propagate the credit and the utility
15:55or the sense that the mission has been
15:56accomplished and so these other you know
15:58kind of things are getting propagated
15:59back but it’s not very fungible and
16:01there’s no good way of scaling it
16:04so what we ideally want is we want the
16:06structure kind of like this where you
16:08can you can develop a large scale
16:10optimization surface where anybody can
16:12go and create a new fund and you and you
16:14enable these signal processing units
16:16that can operate at very different
16:17scales you need
16:19fund the funds to to exist here and
16:21crucially you need performance signals
16:23to impact the funding scales many grant
16:26programs out there including some of the
16:28largest scale grant programs in the
16:29world do not really uh have a a
16:33strong feedback loop between the capital
16:35deployed and the capital invested in a
16:37process how well that turned out
16:40and feed that back into their ability to
16:41allocate more and more capital to a
16:43field you also get a cases where you
16:45have very large funds trying to allocate
16:47capital in areas where in many different
16:50fields without a lot of knowledge and a
16:51lot of signal processing and you get
16:52into like these very really bad feedback
16:54loops where massive amounts of money
16:55just gets deployed in poor ways so what
16:58we need is some some composable
16:59structure the ability to create a fund
17:02structure similar to a vc fund that has
17:04composability so you can a fund itself
17:06can raise capital from other funds like
17:09it
17:10that it can deploy capital and can get
17:12feedback directly in that transaction we
17:14need to make that transaction itself
17:16composable and this is what part of the
17:18problem here we don’t have a good unit
17:20to sort of propagate back
17:22um however we might be able to create a
17:24unit like this um so through things like
17:27you know the public funding mechanism uh
17:28exploration that’s happening in the in
17:30the grip space and the dial space um we
17:32might be able to get to
17:34a structure here let me switch
17:38so so one idea and it’s not clear
17:40whether this will work but this kind of
17:41what i think might work
17:44is that we can what we can do is we can
17:46fractionalize the
17:48the value flow of one of these sort of
17:50public public good funding structures
17:53through the use of impact certificates
17:55so if we can use impact certificates to
17:56quantify the value of a particular fund
17:59as it deploys capital in an ecosystem
18:02and use that as a version of a stock
18:04certificate um in a corporation so if we
18:07can use input certificates to create a
18:08much more um
18:11aggregateable and and evaluatable and
18:15legible and so on structure for the and
18:17tradable structure to deal with um uh
18:20impact this kind of a broad-based public
18:22goods impact then we have a component
18:24and a tool that we can then use in this
18:26transaction to trade back and forth
18:28through these through these entities so
18:29that’s kind of what we need in these
18:32structures over here
18:34so if we if we can turn this
18:36so what what in the vc world is is a
18:38stock certificate in the public goods
18:40world we can use an impact certificate
18:42and
18:43part of what’s required here is a broad
18:45market like the stock market that is
18:46going to value these impact certificates
18:48that is going to value the utility of
18:50those components and is going to pay
18:52large amounts of money we’re talking
18:53billions of dollars worth of capital
18:55deployed against these impact
18:56certificates if we can get that to work
18:58then we can build an optimization
18:59surface like this an optimization system
19:02with a bunch of little composable
19:04units
19:05that can build uh you know this broad
19:07larger scale network of blue funds uh
19:10blue here is like the this kind of
19:12public uh good oriented oriented thing
19:14one of the key things here is you can do
19:16carry structures with impact
19:17certificates you can you can have the
19:19traditional vc style
19:21performance indicator of carry and not
19:24have and really couple the the
19:26performance of the fund associated with
19:27the impact that you care about doing and
19:29what’s really cool about impact
19:30certificates is that you can do um
19:32impact certificates associated with
19:33specific fields you can value those
19:35impact certificates with a group of
19:37people that understand that field and
19:39can know after the fact five years later
19:4110 years later and so on what that
19:43impact uh was really worth
19:45so my sense is that we can do this kind
19:47of thing for open source and we can use
19:48impact certificates in these kinds of
19:50kind of structures there’s a great talk
19:53from puja in the last event that goes
19:55into into a bunch of structures i think
19:56that impact certificates can be weaved
19:58through most of these
19:59we can do this in the science and r d
20:01funding we can create
20:03as we have the traditional fund grant
20:05systems we can start flowing back impact
20:06certificates and start accumulating them
20:08we can just start doing it and start
20:09issuing them and just accumulate all
20:11these impact certificates once we have a
20:12bunch of them
20:14then we can start
20:15putting prices for them something would
20:16be really useful that i’ve talked to
20:18people about doing is committing to buy
20:20a number of impact certificates at a
20:21particular moment in time so if a large
20:23funder says we will buy five million
20:26dollars of impact certificates on this
20:27field every year
20:29then there’s a very strong
20:31encouragement for participants all over
20:32the world to do a lot of work to create
20:35those impact certificates and then
20:36compete for that for that funding at the
20:38end
20:40we can do this for decarbonization and
20:42so on and one of the good news is that
20:43carbon credits and
20:45renewable energy certificates and a lot
20:47of other financializing instruments are
20:49exactly the kind of impact certificates
20:52that we’re talking about those are just
20:53a special case uh
20:55example in a specific field but what
20:57we’re talking about is something much
20:58more general something much more
20:59composable and something much more
21:01legible by the capital structures that
21:02exist today like we basically want to
21:04come up with something that we can just
21:05hook into the capital structures and
21:07make them all work
21:09uh cool so with that um last plug for
21:12um sorry this bitcoin it auto corrected
21:14that’s so annoying uh just like
21:18how rude
21:22it’s the most the most detent thing ever
21:25instead of turning into the most regen
21:27thing ever
21:30yes exactly if it only were that easy
21:32right um cool so with that uh thank you
21:35very much uh sorry for going a bit over
21:37over time and i apologize for having a
21:38new macbook and uh filling us all with
21:41problems thank you
21:43[Applause]
Responses