Computer scientist & VC. I invest @Lunar_VC in 🇪🇺 pre-seed CompSci moonshots. Autism Spectrum visibility🏆. Wrong in interesting ways.📩 elad@lunarventures.eu
1 subscribed
Oct 13, 2020 • 34 tweets • 8 min read
DAOs & Token-based Governance is a fantasy.
Stop emphasizing token-based voting; start focusing on assigning executive power. Even countries and big corporations fail to ensure wide participation by their "token holders" - instead they maintain strong executive bodies
Thread 👇
The crypto world has a fantasy about token-based governance. It goes like this: you want to have a commercial organization with a mission (e.g. a DAO). So you mint tokens that represent both shares in the profit and governance rights in the organization. Exactly like stock.
Jul 28, 2020 • 12 tweets • 3 min read
This is a funnel problem: there's no onramp from being a research scientist to being an investor.
We're working on building such a funnel @lunar_vc: writing the playbook for transitioning from a research scientist to VC (I'm the guinea pig :).
This is a future-defining problem, IMO. The most interesting and important startups are increasingly based on hardcore STEM. The VC industry needs partners with similarly hardcore R&D background to do pre-seed/seed, otherwise these investments just fizzle.
Mar 1, 2020 • 15 tweets • 3 min read
The Game Theory of Coronavirus Intervention -- why Coronavirus creates a tragedy of the commons.
Or: "Should I cancel my conference trip next week?"
No. And Yes. And No. But really, Yes.
A thread 👇
TL;DR: Coronavirus creates a tragedy of the commons: The correct societal response is to cancel all conferences & events, but selfish individuals rationally have no reason to stop attending conferences.
Let me explain:
2/
Apr 27, 2019 • 19 tweets • 8 min read
@thellimist 1/ There are many examples from the history of algorithms and data structures. Here's one that comes to mind:
Computer Science has a phenomenon called "The Curse of Dimensionality" en.wikipedia.org/wiki/Curse_of_…@thellimist 2/ The curse of dimensionality states that for geometric problems (such as nearest neighbor search), the complexity of the problem is exponential in the number of dimensions. This means that problems like nearest neighbor search become intractable as the dimension increases.
Jan 31, 2019 • 7 tweets • 2 min read
These dedicated networks will still use the Ethereum mainnet to perform any major adjudication. The mainnet will be a "fallback of last resort", like a supreme court. This will allow the app-dedicated chains to enjoy the security of eth, but will barely consume any tx volume
1/ Insight: Decentralization is a tool, not a goal. The true goal is reducing friction. Reliance on trusted central entities creates a lot of friction: compliance friction, accounting friction, legal friction, negotiation friction. Decentralization often reduces these frictions.
2/ But decentralization introduces other types of friction: UX friction; adoption friction; InfoSec friction. To minimize friction, we must pragmatically use technological tools depending on the use case. Decentralization is only one such tool.
Jan 25, 2019 • 4 tweets • 2 min read
Fascinating thread on the decline of newspapers. @JeremyLittau points out that newspapers perform an incredibly important public good, as the public watchdog, the fourth estate. And then he suggests ways you can help: by subscribing to local publications, by investing. >>
>> But there's an anomaly here. Since journalism is a critical public good, why aren't we funding it collectively with tax dollars? Because we're afraid of it getting politicized? Court systems are funded with tax dollars, and we found a way to keep them separated from politics>>
Jan 16, 2019 • 7 tweets • 2 min read
Berliners: I'll be speaking tonight at Nature 2.0 Meetup about "Free Range Algortihms": about the coming wave of self-sovereign algorithms, running on Blockchain systems, unstoppable, self-mutating, paying tokens to humans to help them grow and evolve.
@ryan_caldbeck Whow, I haven't read that thread yet and I'm so happy you linked to it, it's exceptionally interesting. And very convincing.
I think it does have one big weakness: Suppose there is nearly infinite data on the CPG companies results until they get to $10M revenue. >>
@ryan_caldbeck Now you need to build a predictor that takes in that historical data and learns to predict which companies will become unicorns and which will fail. My question is how do you build such a predictor, given that it takes 3+ years to actually get the correct "label"/result? >>
1 subscribed
