Basic Attention 백서

Basic Attention

원문

bitpro가 원문에서 글자만 모은 내용이에요. 표·그림·수식은 빠지거나 깨질 수 있어요. 전체 4쪽 중 3쪽

-founded Kerf Software. Machine Learning, Forecasting & Quantitative Finance. • Bradley Richter, Head of Design, Previously: EFI/Fiery, Co-creator: eBeam & Luidia, Percipo. Advising Circullio. • Catherine Corre, Head of Communications, Previously: AOL, Netscape. • Marshall T. Rose, Senior Engineer, PhD from UC Irvine, co-creator of SNMP and was with the Internet Engineering Task Force. • Brian Johnson, Senior Engineer, was previously at JD Power and Korrelate. • Luke Mulks, Senior Ad-tech Specialist, for technical incident response, investi- gation, support & issue resolution for ad tech and the Brave Browser. Devel- oping/advising on ad tech and tracking threats that Brave shields users from (pr/blog). • Aubrey Keus, Senior Engineer, Previously: Pulse360. • Yan Zhu, Senior Engineer, EFF Fellow. Previously: Yahoo, Tor Project, HTTPS Everywhere, Privacy Badger. 21 5 Token Launch 5.1 Token Launch summary Our goal is to raise a maximum of $24 million USD and a minimum of $5 million USD. Some of the numbers may change with ETH/USD exchange rates and volatility, but the following numbers are best effort estimates as of May 28, 2017. • Maximum financing: 156,250 ETH -this may change with exchange rates. • Minimum financing: 27,343.8 ETH. • Exchange rate: 1 ETH = 6,400 Basic Attention Tokens (BAT) -this may change with ETH exchange rates. • Token contract address: TBD (Published through various channels 48hrs be- fore crowdsale launch date). • Launch date and time : 8AM PST May 31, 2017 block number 3,798,640 • Token launch time-frame: 30 days (based on Blocknumber 3,963,480). • Token launch completion: Token launch will end when either the maximum number of ETH are raised or block number 3,963,480 is reached. If less than the minimum ETH are raised, ETH can be retrieved by holders of BAT. 5.2 Token Distribution • Brave: 13.3% of max; 200 million BAT. • User growth pool: 300 million BAT. • Token available to public at launch: 1 billion (corresponding to the ETH raised at token launch). 5.3 User Growth Pool User growth fund is used to incentivize users to participate in the BAT ecosystem. • A 300 million endowment is for early adopters of Brave and the BAT at up to 5 BAT/user. • BAT received as a reward can only be used within the BAT ecosystem for value added services. • Unused BAT after 6 months will be sent back to the user growth fund which can then be used for new users. • Existing Brave users can get tokens by updating their app and verifying phone number. • No new tokens will be created once the user growth pool is exhausted. 22 Figure 12: Budget allocation of ETH raised 5.4 Budget Allocation • BAT Team: 58% of budget The team consists of just over 20 engineers. This financing allows for the rollout of the BAT solution, including the necessary ad- justments to and development of the existing Brave browser technology. • Administration: 10% of budget Consists of BAT legal, security, accounting and other associated administration costs. • Marketing: 12% of budget Marketing will focus on expanding awareness and adoption of the Brave browser and the BAT solution among users, publishers and advertisers. This also includes the growth and maintenance of the world-wide community. • Contractors: 13% of budget These funds will be directed at third-party providers offering engineering, marketing, growth-hacking, PR, partnerships, af- filiate programs and more. • Contingency: 7% of budget This is a set-aside for unforseen costs. 23 6 BAT FAQs What does BAT stand for and what is it? Basic Attention Token. The BAT, a token based on the Ethereum technology, is a unit of exhange in a new Blockchain based digital advertising system. User attention is anonymously monitored in the Brave browser and publishers are rewarded accordingly with BATs. Users also get a share of BATs for participating. What do BATs represent? BATs are tokens in a new Blockchain and attention-based digital advertising platform. They are not refundable, nor are they securities or for speculation. There is no promise of future performance. There is no suggestion or promise that BAT has or will hold a particular value. BATs give no rights in the company and do not represent participation in the company. BATs are sold as a functional good. Any value received by company may be spent without conditions. BATs are meant only for experts in cryptographic tokens and blockchain-based software systems. What amount is being raised? Whats the cap of tokens? Will there be a follow-on offering? We are targeting a raise of as much as $24 million USD and a cap of 1.5 billion tokens. We do not plan to have a follow-on offering. What crypto-currencies are accepted in the crowdsale? ETH will be accepted in the crowdsale. You will be required to have an Ethereum wallet pointed at the token/crowdsale address to participate in the crowdsale. BAT are Ethereum derived tokens. If you hold BTC or some other crypto-currency it can be exchanged for ETH and used to participate in the crowdsale. When will the Crowdsale happen? We’re working with security auditors to finalize the contract. When they have completed their analysis we will announce the date. Note that the BAT crowdsale parameters will be tied to blocknumber, so times will depend on Ethereum mining rates. The contract will be pushed to Ethereum mainnet 3 days before the crowdsale starts. We’ll also give people a week to interact with the contract on Ropsten/testnet. What is the price of BAT? BAT will be a fixed ratio to ETH. This may vary slightly with ETH volatility as we get closer to the contract deployment date. The exchange rate will be 6400 BAT per ETH. 24 How will Brave use ETH raised during token launch? The ETH received in the crowdsale will by used by Brave Software to build out the Blockchain-based digital advertising system, which uses BATs as a unit of exchange. How will Brave store ETH? Brave will use the standard Ethereum multisig wallet to store ETH. Are BAT tokens transferable? Crowdsale tokens are immediately transferable. Tokens used in the Browser may only be donated or used to pay publishers for premium content or for other services. Tokens may also be used by publishers for promotions. 25 7 Appendix 7.1 A More Efficient Market: Coase Theorem Problems involving social and transactions costs have been studied by economists. Ronald H. Coase was awarded the Nobel Prize in Economics in 1991 for his work on the allocation of radio frequency resources.[25] Modern problems in ad-tech are addressable using the work of Coase and subsequent commenters on his idea. At present, the effects of today’s overcomplicated advertising ecosystem is a negative externality or “social cost” for the user. The user’s privacy is invaded, the browsing experience compromised, and even the limited supply of internet bandwidth on mobile devices is depleted by the present state of this ecosystem. Effectively, the market for user attention has become inefficient; the transaction costs of advertisers purchasing attention have become too high. The widespread adoption of ad blocking technology adds a negative externality on the publishers as well. If everyone blocked advertisements, there would be little content left to exchange for user attention, as publishers go out of business. An efficient market for attention would remove these negative externalities, or compensate all parties to the transaction in an efficient way. The Coase theorem states that trade in an externality or “social cost” is possible. If there are sufficiently low transaction costs, information symmetry, and well defined property rights, bargaining will lead to a Pareto-efficient outcome regardless of the initial allocation of property. The standard textbook example of the Coase theorem consists of a factory which produces pollution as a side-effect of the manufacturing process, and a neighboring landowner who suffers from the pollution. In the case where the neighbor owns the pollution rights; Q = 1− (P +c) c is marginal cost of production, P is price for pollution permit, Q is marginal cost function for the manufacturer in the case. Neighbor has valuation ν for clean environment, and the sale of Q pollution permits entails a loss of νQ =ν(1− (P +c)), so the neighbor finds the price of pollution permits by maximizing net benefit max P {(1− (P +c))P−ν(1− (P +c))} The benefit maximization is 1− 2P−c +ν = 0 Giving the price P = 1−c +ν2 and the units bought by the factory Q = 1−c−ν2 26 If the factory has the entire property right, the neighbor effectively purchases some share of the pollution right from the factory which it doesn’t use. The neighbor wants to buy Q = 1− (P−ν) units. The factory maximizes its net benefit with max P {(1− (P−ν))P−c(1− (P−nu))} The factory’s profit maximization is 1− 2P +ν−c = 0 So the price is still P = 1−c +ν2 For Coase’s theorem to hold symmetrically, it requires well-defined property rights. By definition, the user’s attention is the valued quantity. The user can make the decision to block ads from a given publisher, or choose to forgo interacting with a publisher altogether. This makes it obvious that attention belongs to users de facto and notwithstanding the efforts of some publishers and advertising firms to assert ownership of user attention de jure. Even in commonplace situations where user attention is de jure required, de facto, users still own their own attention. For example, attention is required while the safety demonstration is given on an airline flight, but people often ignore it anyway. Another requirement for validity of the symmetric version of Coase’s theorem is information symmetry. Information asymmetry between publishers, advertisers and users has kept the existing advertising ecosystem in place for some time, but as we can see from the growing use of ad-blockers, the information asymmetries on the user side are crumbling. At present, advertisers and publishers have a severe information asymmetry in that most of the metrics they use to assess campaign effectiveness are indirect and admin- istered by middlemen whose interests are not aligned with the interests of one or both parties. Complex “viewability” metrics create unnecessary conflict between advertisers and publishers. There is no technical reason for this information asymmetry; it can be mitigated with better technology, in particular browser technology at the endpoint where all the data can be measured privately and confirmed anonymously. The final requirement, which is only a soft requirement for Coasean analysis in the case of well-defined property rights, is that of low transaction costs. The Coasean transaction cost refers to the cost of negotiating a deal which can suit all parties to a dispute involving social costs. With the existing ecosystem, the transaction costs are impossibly high, with advertisers, publishers and users unable to come to terms. In our example of present-day ad networks, we have a potential Coasean bargain between publishers and users, with a better outcome for advertisers as well. A Coasean solution to the attention economy inefficiencies for publishers and users is for advertisers to pay publishers by actual attention given to the publisher by the user. Advertisers will pay the publisher for a share of the valuable attention the user pays to the publisher. Readers also will be directly compensated for their valued attention. 27 Figure 13: BAT Digital Ad Flow The “pollution” of privacy invasiveness, slow browsing and data costs can be almost completely mitigated. Advertisers will know if their messages are delivered without hav- ing to resort to complex arguments about “viewability.” Publishers will not experience the negative externalities from the growing problem of ad blocker adoption. Various proxies have been developed by advertisers and publishers to attempt to measure user attention using indirect techniques of “viewability,” but the advent of ad- blocking technologies and the increasing problem of fraud from non-human entities have cast doubt on such methods. A more direct technique would be to pay publishers via cryptographically secure methods, and serve the ad directly in the browser. Since the browser ultimately measures how the user interacts with the website better than any indirect meddling by intermediaries, involving the browser software itself in the process provides accurate measures of user attention bestowed on the publisher and advertiser. The browser also provides a much richer data set for understanding what the in- dividual user is interested in. The Brave browser will contain opt-in and transparent machine learning algorithms for assessing user interests. While an ad campaign targeted to a financial publisher may have value to the broad interests of the overall readership of the publisher, individual readers can be given ads tailored to their individual and even private preferences. For example, sending an ad for discount bond brokers to people who are following the markets in municipal bond issues. The user who is reading about tech stocks and who has no interest in municipal bonds won’t receive the ad. The advertiser will effectively target the precise microsegment they are interested in reaching. The user receives more relevant ads while interest lasts, and private interests are not revealed to publishers or advertisers. The idea that user attention should have monetary value is familiar to both publish- 28 ers and advertisers. The idea of publishers and particularly users being paid directly for attention bestowed on the publisher is novel. Since the valuable commodity is user attention, it makes economic sense that the user be compensated for their attention. One could justify this as a compensation for the externalities imposed on users by the advertising ecosystem. One could also justify this by the fact that one is more likely to perform an action if one is compensated for it. There is also confirmation that the actual user attention is bestowed on the publisher via the addition of cryptographic contracts built on blockchain to this advertising stack. The code is open source and can be reviewed by researchers and interested parties on the advertiser and publisher sides. Since the transactions for the first deployment of BAT will happen through the Brave Ledger, which has privacy and deterministic user anonymity by design, full transparency can be achieved while user privacy is maintained. While this centralized solution should fulfill economic and technical goals, for further iterations, a decentralized solution could be developed to allow for trustless auditable transactions. While paying a user to look at a publisher content may seem heretical to advertisers, the reality is the advertiser is paying someone. Removing the vast field of middlemen who add no value to the user/publisher relationship allows for a situation where the user may be compensated for valuable attention (made more valuable and relevant by measures of user interest at the browser) with no impact to advertiser costs and positive impact to publisher revenues. From a financial point of view, this could be seen as a variation on some other kind of short term promotion: advertisers regularly provide coupons and rebates on products. Promotions do not solve the problem of informing the user of the advertiser’s product in the first place. Promotions also don’t induce user loyalty or engagement. Most CMOs agree that short term sales can be improved with promotions, but sustainable competitive advantage can’t be achieved using promotions, hence the use of advertisements. 7.2 A Three-Way Coasean Bargain The three-way Coase theorem is a source of much research interest among economists. The existence of “empty cores” in some situations have called into question the appli- cability of the Coase theorem to real world examples involving multiple distinct play- ers[26]. While there are many more than three participants in the online ad market, we can idealize them as consisting of three participants: the advertiser, the publisher and the user. This analysis is useful for understanding the game theoretic considerations, for addressing any “empty core” arguments against the proposed Coasean bargain, as well as for illustrating the dire state of the publishing industry. We propose the Basic Attention Token (BAT), a cryptographically-secure token, as the medium of exchange for facilitating this Coasean bargain while protecting the privacy of the user. The advertiser wants to purchase user attention. This is broadly analogous to the “cost of production” in the exposition of the Coase theorem above, whose notation we follow. The advertiser values the user attention with price Caa. The publisher wishes to 29 monetize the attention Cpa paid to the website. The user who views the website values the content of the website with attention Cca. Advertisers and publishers in the present ecosystem have transaction costs associated with monetization of attention. Publishers are paid by advertisers to provide user attention. The intermediaries of the present system create costs therefore Cpa <C aa. Note, when we talk about “transaction costs” apropos the Coase theorem, we refer to the transaction costs for negotiating a deal between the players of the Coasean game, therefore, rather awkwardly, the monetary costs of getting the ad to the publisher is not considered a “transaction cost” per se. The present advertising ecosystem produces “social costs” or attention pollution as we have discussed above. These social costs are known to be large. For some large fraction of users (22% Lumascape state of the ad industry), the social costs are larger than the attention cost. We will label the pollution cost following the above example as Pca. In the present situation, the user will view the publisher and advertisers content so long as Cca > Pca. Every user is different, and of course, the publishers and advertisers vary as well, but the existence and growth of a large population of users for whom Cca < Pca indicates that we are approaching the time where this inequality is always violated. The consequences of this are that Cpa = 0 ⇐⇒ (Cca<P ca) SinceCca is proportional to Publisher profit (and advertiser profit in “attention”), any value which keepsCca>P ca is advantageous to the Publisher and Advertiser. Effectively the advertiser and the publisher combined are the factory in this argument, and the user owns the pollution rights. However, the user also values the product of the publisher. In the degenerate case where Cca<P ca the user is also eventually harmed as the attention economy collapses, and the user takes up other hobbies. The social cost should be decomposed into its constituent parts. We have identified the primary components of the social cost in our exposition of the advertising industry above. Security risk is one component, Ps. Hacker networks can place ads in irrespon- sible ad exchanges, which could have very large costs for individual users as well as the publisher who displays those ads. Privacy loss is a very important social cost associated with the advertising landscape as it presently exists,Pp. Privacy invasions are presently required by advertisers to make sure the advertisement is actually viewed by a relevant user. In effect, the advertisers are paying for something which adds value to the attention. Data costs are also a significant part of the social cost of the present day advertising ecosystem Pd. These costs are often borne by the user as a result of the activities of the middlemen who serve the advertiser and publisher. These costs seem most trivial, but for many users, they are among the top causes driving ad blocker adoption. For all viewers of online ad funded content, considerable time is taken in dealing with the cost of downloading and executing all the privacy-violating code. In addition to this cost, for those users who are using mobile devices, the monetary charges can be significant. It has
프로젝트로 돌아가기