Zuck's Metaverse, Neobank Subscriptions, Computer-Driven Funds, GPT-3 and Zerodha
The Tech Tok Weekly #23
Last Monday, the sun rose in the East, the sky was blue, and Messi was at Barcelona. This Monday, the sun has risen in the East, the sky is still blue, but Messi has had to leave Barcelona in a development that is arguably more shocking than the sun rising in the West or a green sky.
Messi couldnβt hold back the tears, and neither should we.
Monday Musings
The path towards getting out of the current predicament we are in with climate change starts with effective government regulation. Left to their own devices without any incentives to change, the relatively short-term thinking of private companies (which is, to a certain extent, understandable) will not lead to the kind of significant change that is needed to reach a maximum increase of 1.5C in the earthβs temperature. The reason is simple: even though a record number of companies are making climate pledges, these fall far short of the required pace. Indeed, at this level, even if all Paris climate agreement pledges are met, we will see catastrophic temperature rises of about 2.4C by the end of the century. Many pledges are several decades out, have few interim targets to meet, and are difficult to track. Moreover, incentives are severely misaligned. ESG is being seen as more of a buzzword and a PR commitment to be met rather than something to engender real change, and according to former sustainability executives at firms like BlackRock, there is a βsustaina-bubbleβ, where companies are more focused on marketing sustainable funds rather than creating investments that have actual climate impact. Unless governments give these companies incentives and punishments for not meeting targets, they will have no economic motivation to make real change. It is therefore critical for governments to create frameworks and step-by-step plans that chart out exactly how they can reach the 1.5C pledge, and bring private companies into the fold.
Meanwhile, this weekβ¦
Zuck explores the metaverse, neobanks need to dive headfirst into the subscription model and offer new, innovative services, and computer-driven funds are ready to enter markets that have previously been inhospitable to them. Also, Zerodhaβs CTO shares some lessons on how the company was able to scale so quickly and efficiently over the last year, and a16z breaks down the promise of GPT-3.
#1 Zuckerberg is Turning Trillion-Dollar Facebook into a βMetaverseβ Company, He Tells Investors
The concept of the metaverse has quickly reached a level of public prominence that would have been unimaginable even a year ago β to such an extent that Mark Zuckerberg mentioned it 20 times during Facebookβs earnings call. According to Zuck, the metaverse is a βvirtual environment where you can be present with people in digital spaces.β From a high level, though, Facebookβs strategy seems hollow for a couple of reasons. Firstly, it seems as if the way Facebook aims to make its presence in the metaverse felt is via AR/VR, through Oculus and the acquisition of VR-focused game studios. VR devices, though impressive, still have a lot of friction in the consumer experience and lack the fervour from users that other (rapidly growing) gaming ecosystems have. It is difficult to imagine critical uptake of a metaverse where consumers need a specialised piece of hardware to access it. Secondly, Facebook doesnβt seem like the right company to espouse the ideals of a metaverse β namely, according to Foundation, an βemerging shared digital space that can be navigated without corporate interference.β A metaverse, according to them, empowers us to own our own data, decentralises power, and helps free the internet, none of which seem like ideals that mesh with how Facebook operates or how it is seen in the public sphere. Being cynical, Facebookβs interest in the metaverse seems like it stems from an opportunity to capture the βnext frontierβ for purely commercial reasons, rather than from any truly innovative or altruistic motivation.
The metaverse offers Facebook an opportunity to draw a line between its moonshot efforts and its core business, building a wide-reaching hub that shines on augmented reality and virtual reality platforms but feels just as friendly on mobile and desktop.
#2 The (Neo) Bank Bundle & Transition to Subscription Revenue
Every aspect of our life is now dictated by the subscriptions we have. If we want to order food, we can get Deliveroo Plus and eliminate delivery fees. We can even subscribe to get craft beer from different countries every month β the list is endless. One obvious industry where subscriptions have not caught on meaningfully, however, is financial services, and it is not for lack of trying. Neobanks like Revolut apparently generate 30% of their revenue via different subscription tiers. Even so, the suite of services offered as part of these subscription packages is lacking, especially compared to what could be offered. Revolut, for example, offers higher ATM withdrawal limits, 0.50% on savings, virtual cards, premium designs, free SWIFT transfers, etc. It would be an understatement to say that these offerings are not enticing. There are a multitude of services that these neobanks could (and need to) offer, especially in light of long-term profitability concerns. An βoptimalβ neobank bundle would contain some offerings that are the bare minimum (ATM withdrawals, brokerage, card issuance, crypto, loans, PFM tools and P2P payments) and some that could catapult them into profitability. Examples of these βInnovativeβ offerings are alternative assets, art, collectibles, private company shares, DeFi (yield farming, staking, etc.), bill pay, liability optimisation, BNPL, financial advice, options, insurance, etc. The range of possibilities also means that neobanks can differentiate themselves and appeal to various types of consumers, while reducing their dependency on revenue from fees and net interest margin.
According to Northwestern there are Eight Dimensions of Wellness including (i) Physical (ii) Emotional (iii) Social (iv) Intellectual (v) Environmental (vi) Spiritual (vii) Vocational & (viii) Financial. While thereβs a subscription for almost everything else, financial wellness remains a significant whitespace opportunity.
#3 Computer-Driven Funds to Muscle into New Markets, Says Man Group CIO
According to Sandy Rattray, CIO of UK hedge fund Man Group, the coming decade will see quants which use data and computer-driven models make significant headway in areas of finance like fund management where they have previously not fared well. He says this with no lack of experience in the space β since Rattray was made CIO in 2013, the AUM of Man Groupβs systematic, computer-powered AHL unit grew from $11.9 billion at the end of 2013 to $50.7 billion this year, an increase of 326%. As he retires at the end of this year, Rattray also says that he wishes that he could have been a quant in private equity, since the business is βabout as un-tech savvy as you can imagineβ. Currently, PE managers spend a lot of time on manual tasks and the lifecycle of a traditional PE investment has seen little change over the past few decades. Data aggregation is poorly done by PE firms, many of whom have no way to efficiently aggregate their data; valuation models are primarily created using Excel; and risk assessment and due diligence for most transactions is still done manually, adding costs, inefficiencies and financial strain in narrow investment windows. Algorithms can take over many of these tasks β they can model companies by using peer data or analysing trends, they can map business-to-business relationships around the world, and as algorithms get better at analysing textual data (e.g., GPT-3), they can sift it for profitable trading patterns.
Machines canβt read bond prospectuses as well as a human. But they might eventually get there, and when they do, the good news is that they never get tired or bored of reading them.
#4 Scaling with Common Sense #2: Being Future Ready
Zerodha has been one of the success stories of the Indian FinTech ecosystem, and a primary reason for its growing influence is its technological prowess. Over the last year, Zerodha managed to scale with a 31-person tech team (up from 30 pre-COVID) in the midst of a tripling of its customer base and a 6-fold increase in customer traffic. Kailash Nadh, Zerodhaβs CTO, imparts some priceless insights from their experience. Some of the best ones are:
In constantly changing business environments, it is necessary to be pragmatic rather than search for a perfect state.
A stable API maintaining compatibility with internal and external systems is critical to respond to unexpected events and change.
Clear, generic and well-documented APIs can be transformative, and even lead to potentially new revenue streams β e.g., Kiteβs API enabled Zerodha to build a βBrokerage-as-a-Platformβ service.
Many times, it is important to not continuously add features but rather improve the quality and performance of existing software itself. At Zerodha, they consistently assess all dimensions of code that may have performance problems and decide when to refactor or rewrite it, all at the discretion of the tech team. Β
A fundamental principle of future readiness is that non-technical teams should generally communicate problem statements instead of specific implementation details.
Removing and swapping out entire systems with minimal disruption can be done by self-hosting and managing as many systems as possible; this gave Zerodha complete control, insight, and liberty to change them.
Over the last eight years, we have established that self-hosting FOSS (Free and Open Source Software) systems is an extremely cost efficient and common sense way of building and scaling a technology organisation even with a tiny tech teamβour stack on our terms under our control.
There are few things as exciting in tech as GPT-3, OpenAIβs natural language processing-based text editor that could be the first step towards general artificial intelligence. GPT-3 is a deep neural network that can do concurrent computations on large sentences; i.e., to disambiguate sentences, GPT-3 considers large chunks of text all at once, in context. OpenAI has been training the neural network on huge datasets composed of half a trillion words, two orders of magnitude higher than previous attempts. The GPT-3 model is trained once, and is given only one example, priming the system to answer one type of question without adjusting the modelβs weights β it is, essentially, trained to give a right or wrong answer, similar to how children learn language. The possibilities are endless β for example, companies that want to solve a text problem (e.g., chatbots or customer support) can build cheaply on top of such infrastructure using an API. It has the potential to significantly reduce the AI implementation timescale β gathering and cleaning the data and building the data model. While some current jobs may become irrelevant in a theoretical GPT-3-ruled world, new jobs can replace these, namely those that prime the AI system with the right examples at the beginning, and those that debug it at the end. A key consideration is also ethical; text can contain many social norms we do not want the model to learn, so we need to identify how to teach it the social norms we do want to emit.
The OpenAI product team is going to have conversations with dozens and dozens of startups that are using their technology. And then they successfully refine the API and improve the performance, and set up the security rules and all of that, so that it becomes something as easy to use as say, Stripe or Twilio.
To join our journey into the worlds of tech and business, subscribe below!
Disclaimer
This is a personal blog. Any views or opinions represented in this blog are personal and belong solely to the article authors and do not represent those of people, institutions or organizations that those authors may or may not be associated with in professional or personal capacity, unless explicitly stated. All content provided on this blog is for informational purposes only. The owner of this blog makes no representations as to the accuracy or completeness of any information on this site or found by following any link on this site. The owner will not be liable for any errors or omissions in this information nor for the availability of this information. The owner will not be liable for any losses, injuries, or damages from the display or use of this information. Any views or opinions are not intended to malign any religion, ethnic group, club, organization, company, or individual.