Issue #1
Meta Meets Microsoft | What to Watch in AI | Venture Capital Is Ripe for Disruption | Free Resources- Desklamp, Journalist's Toolbox, OEC
Bridal Veil Falls, Yosemite Valley, California (1871-1873) by Albert Bierstadt
Source: Artvee
Tech 🤖
Meta Meets Microsoft
By Stratechery
Annotations
After testing Horizon Workrooms Ben is quite convinced about it given its spatial audio feature, ability to convey hand gestures and viewing directions.
Ben believes in the enterprise use cases for VR stating that companies are always on the lookout for ways to increase employee productivity and that VR will follow the path of PCs- embraced by companies at first and will soon find their way into homes when consumers are convinced about its usefulness.
The partnership between Microsoft and Meta seems to be a massive win for Microsoft not just because it gets privileged access to one of the most advanced VR headsets in the market but also because it opens up a new market for Microsoft’s suite of enterprise tools.
The social experience use cases such as gaming are not so clear yet given the significant barrier to entry in terms of cost (Quest Pro costs US$1500). Despite there being a strong enterprise case for the Quest, Meta seems to be concentrating on social experiences.
Other downsides of the Quest Pro include the weight (heavy), battery life (1-2hrs), and resolution (too low).
Ben sums it up best: “What is clear is that Zuckerberg in particular seems more committed to VR than ever. It may be the case that he is seen as the founding father of the Metaverse, even as Meta is a potential casualty.”
Reflections
I personally feel the enterprise vision for VR could become a reality in the age of remote work and propel Meta given its first-mover advantage. But first movers are not always the ones who eventually become the leaders.
Contrary to popular opinion, Meta is still a solid business owning three (FB, WhatsApp, and Instagram) of the top four social media websites in terms of monthly active users (MAU). In its 2022 Q3 results, it reported US$4.4B in net income with 20% operating margins.
Yes, Apple’s privacy measures are an existential threat to Meta’s ad-driven business model but I believe it’s sitting on a gold mine-WhatsApp. There are two relatively unexplored solutions where I would like to see more innovation- Digital Storefronts on WhatsApp and Productivity Tools for WhatsApp (advanced search for messages, separate work and personal accounts, etc). While these solutions might already exist, my point here is that Meta should concentrate more on bringing new productivity features to WhatsApp.
What to Watch in AI
By The Generalist
Annotations
GitHub Copilot
Helps developers in writing, testing, and refining code in popular languages.
Built using OpenAI’s Codex, Copilot assists in completing lines of code and can also generate them from a plaintext description, thereby saving countless hours in the process.
Copilot need not be restricted to devs alone, it could potentially serve other professions as well such as attorneys (help draft contracts) and architects (model, and optimize buildings).
Digital twins in clinical trials
The pharmaceutical and biotech industries are extremely complex and the process of bringing a drug to the market is a lengthy and highly risky affair. The average drug needs a decade and US$2B to bring to market.
Typical failure rates for clinical trials can range from 70-90%. Around 80% of clinical trials suffer from enrollment-related delays which results in trial sponsors losing up to US$8M/day in potential revenue when a trial is delayed.
Digital twins could be a solution. Generative machine learning (ML) models can simulate placebo outcomes for patients in clinical trials. By creating a digital twin for each participant in the experimental group we can model how that patient would’ve reacted had they been in the control group (experimental vs control group).
This helps pharma and biotech firms to reduce the number of participants that they need to recruit for the control group.
Industrial automation
Each year, there’s a 7% decrease in skilled human welders, while the demand for them increases by 4%. By 2024, it is estimated that there’ll be 400K human welder vacancies in the US alone.
Path Robotics could help bridge the gap. Founded in 2018, Path is an Ohio, US-based startup that enables manufacturers to use off-the-shelf robots to autonomously weld unique parts without the need for time-consuming and expensive reprogramming. Its robots can learn to weld visually instead of being taught with code.
Reflections
I think we’ve finally reached an inflection point in AI, where working solutions have started to appear and we’ve exited the gimmicky and experimental phase.
According to Gartner’s Hype Cycle for AI, many of the innovations are within reach (2-5yrs away) with computer vision already nearing the plateau of productivity. Other interesting observations:
2-5yrs away: Generative AI, Edge AI, Deep Learning
5-10yrs away: Smart Robots, Natural Language Processing
10yrs+ away: Artificial General Intelligence, Autonomous Vehicles
VC 🚀
Venture Capital Is Ripe for Disruption
By Napkin Math
Source: Napkin Math
Annotations
What’s wrong with the current VC model?
On average, to hit a 3-5x return for a fund, VCs look for US$50B+ exits which force founders to take on more and more risks.
From limited partners (LPs) to founders to VC firms, the Power Law is deeply entrenched among all the industry participants forcing them to excessively focus on outsized outcomes and shun smaller and more practical opportunities.
VC used to be about investing in breakthrough technologies and backing mavericks but now with more than 1000 VCs (in the US alone) chasing a handful of startups that can deliver US$50B+ outcomes coupled with “pattern recognition” has resulted in only a fraction of the ideas getting funded.
What does the ideal VC model look like?
Smaller fund sizes of US$150M or less.
Offering equity buybacks and revenue sharing options for investors in the startup in case going public is not viable.
In this scenario, a larger portion of the portfolio would at least return the capital with a few exceptional exits (US$500M to US$1B would be enough).
Reality check
Similar to how investing in blue chip firms is the norm in public equity investing, investing in blue chip VCs such as Sequoia and Benchmark is the norm for the LP’s fund managers.
LPs (especially university endowments, sovereign wealth funds, hedge funds, private equity firms, etc.) looking for large investment opportunities to deploy capital will not entertain fund sizes of <US$150M.
Reflections
Evan’s ideal VC model honestly sounds attractive and is more practical than the current one. I was never a fan of growth at all costs which is precisely what led to the ideology of move fast and break things producing companies such as Facebook, Uber, and Airbnb to name a few.
VCs chasing a handful of startups with the promise of multi-billion dollar exits is probably why due diligence takes a back seat while investor FOMO ends up being a key driver.
To fight climate change we desperately need breakthrough technologies and for that, we need more patient capital which does not always have to be the government.
Free Resources 💡
Desklamp: The free and amazing PDF note-making tool that lets you highlight, collaborate, take notes and search intelligently is finally here.
Journalist’s Toolbox: A great website to find helpful resources for almost anything related to journalism, from data journalism to transcription tools.
The Observatory of Economic Complexity: Visualize, understand, and interact with the latest international trade data. An absolute treasure trove of trade data.
For more free resources (150+ websites and tools) please check out Searching (it’s a Notion database that I’ve created).
This is a great read!