Math is harder to MITM than science.

Mathematical claims can be independently reproduced on anyone's computer, while hard-to-reproduce science has to be taken on faith.

As such we should reduce as much as possible to hashed, timestamped, verifiable math.
Normally we don't think of "math" and "science" as being in tension. But it is true that math > science > religion in terms of rigor...and reproducibility.

Billions of phones means anyone can now independently reproduce any mathematical statement. But not all scientific ones.
This suggests that we want to reduce as many assertions as possible to mathematical rather than merely scientific statements.

There's actually much that can be done here using NLP, public datasets, reproducible research, and the ledger of record concept.
Science is how we understand the *natural* world. But math is perhaps even more useful in the constructed digital worlds that we spend more and more of our time in.
With >50% of waking hours spent in front of a screen, much more of people's daily experiences become amenable to cryptographic verification.

Everything that passes in front of you on a screen was at least stored on a computer, if not generated there.

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @balajis

22 Sep
Farming was the 1800s. Manufacturing was the 1900s. Counterintuitively, could investing become the most common "job" of the 2000s?

Reason: crypto and fintech are turning everyone into an investor, just like the internet turned everyone into publishers. How far does that go?
One way of measuring this: what percentage of income do people derive from investments vs wage labor?

We think of wage labor for a salary as the default today, but the transition from farming to manufacturing was a big change. An improvement in some respects, and not in others.
Many bemoan the financialization of the economy. But perhaps we lean into that. Perhaps investing is what people do in a robotic economy if they don't want to be founders.

Investing is similar to consumption. Just click to buy. Picking is hard. But you can join a rolling fund...
Read 6 tweets
22 Sep
Someone should set up a site like Scott Aaronson's Complexity Zoo, but with all the proof-of-X papers and code.
If you combine reproducible research, cryptographic proof-of-X, and federated analytics, you can envision a world where every paper isn't just reproducible, but where every row of data is logged on-chain with all relevant proof-of-Xs to demonstrate data integrity.
- Many papers are based on datasets
- Those datasets are rollups of individual data points, like measurements on an individual
- The datasets (and sometimes the raw data points) are currently in silos like NCBI or Bloomberg
- But you could put them on-chain
Read 8 tweets
21 Sep
Ignoring the competition is just level one. The real task for founders today is ignoring the distraction.
It's hard because these are 95% rules, not 100% rules. You usually want to be aware of what the competition is doing, but run your own game. Yet sometimes that isn't possible. Similarly with distraction.
With competition you might have some rule like "wait before three customers mention this competitor before taking them seriously". Though that may mean you react too late, it means you only react to likely real things. In many ways it's like angel investing with your time.
Read 4 tweets
20 Sep
With respect to proof of location, you can’t naively use GPS because it can be spoofed. You might go with trusted towers of known location, but that’s centralized.

Another approach may be to assume K of N nodes are honest reporters of GPS, and use speed of light as a constraint.
Specifically, suppose you have N nodes, each of which is reporting a time-varying position [x(t), y(t)] over a time interval with M time steps.

Any node which purports to be near another node can be challenged by an equation that it must solve & return as quickly as possible.
For a pair of nodes with reported locations, the speed of light gives a lower bound on how quickly info can go back & forth.

Obviously network topology can dramatically change ping times. But a node should be able to solve more challenges more quickly from nodes it’s nearer to.
Read 11 tweets
20 Sep
An important corollary is that blockchains can be part of a protocol rather than the whole thing.

Many protocol designs become easier if you can assume that all peers read and write to the same global database.

From MVC to CBC, client-blockchain-client?
MVC: model-view-controller, dominant paradigm for web apps. You have an ORM on top of a relational database (the model) that processes commands from users (the controller) to render requests (the view).…
CBC: client-blockchain-client

An alternative paradigm where every client has a local crypto wallet, some coins, and a local relational DB. It leans on a central blockchain to find other clients and route messages to them. The local relational DB is used to cache & index state.
Read 6 tweets
16 Sep
Synchronous remote is indeed much easier than asynchronous remote. Though you can hire night owls or early birds in other locations, remote teams are partitioned by time zone.
This page lists all the countries of the world, grouped by time zone. Image
Some interesting spreadsheets to make:

- What is the distribution of world population across time zones?
- What about in +/- 3 hours from that time zone?
- What is the distribution of tech clusters and unicorns?
- How does time zone overlap with other criteria, like rule of law?
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!