TAO makes TAO waste! Messari researcher dissects Bittensor's network effects
Original article: "Breaking Down Bittensor's Network Effects" Author: Seth Bloomberg (Messari Researcher) Compiled: Zombie Messari Researcher Seth Bloomberg recently published an article on X, detailing the network effects of Bittensor and potential bottlenecks for future development.
The following is the original compilation:
Due to the recent discussions about Bittensor, I would like to share my views and some concerns about it.
Thanks to Saypien (former Messari executive) for the extensive discussion on this topic, and also thanks to Sami Kassab (former Messari researcher), I know he will come to point out my mistakes here.
A basic layer network must establish network effects to succeed, in order to attract and retain users and developers. For smart contract platforms like Ethereum, network effects come from on chain liquidity and stable execution capabilities. For Bittensor, its current network effects come from:
1) Familiar with AI/ML developers who mine on the internet (i.e. developers who create AI models and services);
2) A deep liquidity TAO market that can absorb the selling pressure of tokens. Usually, less mature networks are more susceptible to token selling pressure.
Due to the widespread applicability of (2) to most encrypted networks, I will primarily focus on (1). Specifically, the network effect or flywheel effect generated by (1) is as follows: the new subnet benefits from experienced AI/ML Bittensor developers who continuously produce high-quality results → AI/ML engineers benefit from continuous TAO token issuances. It can be expected that the revenue from TAO tokens, combined with Bittensor's AI/ML talent, will continue to attract more subnet builders, thereby attracting more AI/ML talent and forming a cycle.
Whether you are bearish or bullish on TAO, it is worth exploring how this network effect may collapse. So, let's get started.
Scenario 1: TAO token revenue<operating expenses
AI/ML talents are attracted to the internet because the TAO token returns they receive exceed their operating costs on Bittensor. Mathematics is simple: there is no need to raise a large amount of funds, acquire customers, or generate income. Just integrate the model into Bittensor's subnet and earn TAO token issuance revenue. The assumption of this model is that your earnings (based on current or expected TAO value) should be able to cover your expenses denominated in US dollars. If this account is not established, AI/ML talents will choose to leave Bittensor and seek other opportunities.
Scenario 2: Opportunity Cost Calculation
Imagine you are a skilled AI/ML developer on Bittensor, or you are a small team of two or three AI/ML developers building projects on Bittensor. At some point, you have to ask yourself: 'What is the opportunity cost of continuing to be a miner here?' Well, you know the upside benefits of staying on Bittensor only come from the issuance of TAO tokens and the rise in TAO token prices. With the benefits on this side, you also need to consider the other side: whether creating a new protocol without Bittensor can bring greater benefits? I know... this may be considered 'blasphemy' for some people, but if you don't consider this, then you're not taking responsibility for yourself.
Bittensor has realized the need to provide additional incentive benefits for subnets and their miners. 'Dynamic TAO' seems to be their approved solution.
It is currently running on the testing network, so it is likely to be launched by the end of this year or early next year. Through dynamic TAO, each subnet will have a fund pool similar to Uniswap V2, which will pair and price subnet tokens with TAO. The 'dynamic' aspect of this upgrade lies in the new dynamic TAO token issuance model: the more TAOs pledged in the fund pool, the more TAO tokens allocated to subnets, and the higher the price of subnet tokens. It sounds very cost-effective. AI developers can gain appreciation of subnet tokens while retaining their advantage in the Bittensor ecosystem (i.e. TAO token issuance).
However, pairing subnet tokens with TAO for pricing will set an "artificial upper limit" on the valuation of individual subnet tokens. Ethereum and its L2 or Helium and its subDAOs also have similar token dynamic relationships. This is where opportunity cost calculation comes into play. If you are one of these powerful AI/ML developers or subnet owners, why limit yourself in the way mentioned above? Why not just establish your own protocol or network? I dare say that some venture capitalists may be eager to fund you after seeing a bunch of mediocre AI and encryption projects, and this valuation may be higher than a rigorously designed subnet. Or even better, you have accumulated a large amount of TAO during mining, and after calculation, you believe that you can self fund to establish your own protocol, attract some applications, users, and revenue, and then consider external funding (which is rare today).
In my opinion, this is the long-term challenge Bittensor needs to face - designing mechanisms, incentives, and ecosystems to retain the best AI/ML talent.
I think a recent tweet from mrink0 (Delphi Digital researcher) also hinted at this issue (if not, please forgive me for taking it out of context; if there is an error, please correct me). In my opinion, people like Nous (open-source AI research organization Nous Research) leaving Bittensor to build their own networks indicate that for some Bittensor developers, the incentive to stay in Bittensor is currently not enough to compete with building and publishing their own networks.
Addendum:
After mrink0, in Seth Bloomberg's comment section, he pointed out that he agrees with most of Seth Bloomberg's views and stated that the key challenge will be whether Bittensor can attract a model that can solve the problem of people being willing to pay. Otherwise, it will only be (temporary) token emissions.
Seth Bloomberg gave the following response to this:
Even if the model solves the problem of people being willing to pay, miners may not necessarily benefit from it in any way. People will pay for applications or products that use these models, but the protocol does not define any mechanism for miners to share these benefits. The worst-case scenario is that the demand for miners increases (inference costs rise), but they are unable to generate any income from these applications or products. The income of Bittensor miners cannot increase with the demand or revenue of the applications they serve.
(责任编辑:财经专题)
-
而正是从五月份开始,《王者荣耀》里的版本更新内容疯狂的加入各种各样的社交功能,战队、恋人、师徒,直到最新的LBS玩法,完全消除用户在游戏内的社交障碍,力推用户把《王者荣耀》变成他们现实社交的一部分...[详细]
-
但在电视剧领域,拥有优秀操盘手的小IP则更容易成功。...[详细]
-
2,搜索结果数量:一个关键词的优化难度很多时候可以由关键词的检索结果而决定。...[详细]
-
看看淘品牌里在坚果、服装、家居上的表现,这两年多少已经被淘汰出局了,流量都在向大卖家转移,虽然大卖家日子也不好过,广告费hold不住啊,竞争压力也太大,都狠劲的在活动上熬着,就跟双十一一样,参加了...[详细]
-
创业12年,罗江春的实战经验丰富,但是百度长江学堂的老师们讲授的是系统的理论知识,配合不同商业形态的学员实战分享,罗江春自认收获很大。...[详细]
-
如今微信指数也出来子,也自是闲不住的在微信群里与众好友一起研究了一下微信指数的算法,群里有位大神得出的微信指数算法是: 采用数据:总阅读数R、总点赞数Z、发布文章数N、该帐号当前最高阅读数Rmax、...[详细]
-
” 大学毕业之后,郑志刚又去了日本研究了一年东亚文化,当时是标准的文艺青年范儿,在学校旁的寺庙里听悠扬的钟声,一坐就是大半天,“每次钟声敲响,我就感觉有水在心中流动...[详细]
-
10月31日宏利改革动力混合A净值下跌0.48%,今年来累计下跌8.47%
比如全美在线(835079.OC),全美在线是一家做证券、基金等考试测评服务的公司。...[详细]
-
而关键词的优化难易程度至少有如下四大类型,由难到易。...[详细]
-
10月31日浙商智选价值混合C净值下跌0.80%,近1个月累计下跌6.68%
在广告系列中,将所有其它广告系列的关键词添加为广泛匹配后进行效果监测,通俗的来说就是排除法。...[详细]