1\ Today we uncover a mystery deeply embedded in smart contracts: patterns that enable us to predict the nature of a smart contract with the help of AI.

Our paper, "Reputation Oracles" was published today demonstrating how you can use this to identify scammers with 94% accuracy. 2\ We see reputation assessment as one of the most promising areas at the intersection of crypto and AI.

Why?

Users have little to lose if the model hallucinates, but a lot to gain if it prevents them from interacting with a contract trying to steal their funds.

How much does it cost to 51% attack Bitcoin and Ethereum?

To find out, we simulated what an attack would look like.

Our paper, Breaking BFT, was published today with some interesting results ⬇️

papers.ssrn.com/sol3/papers.cf… 1/ The mere possibility of these types of attacks has caused significant anxiety.

You often hear questions like:

Will Bitcoin remain resilient against 51% attacks after the 21M cap is reached??

Can Lido takeover Ethereum if it controls 34% of validators??

1/ The size of the Bitcoin blockchain is growing at an unprecedented rate.

But how concerning is this trend, really? Will it prevent regular folks from running nodes?

A thread on Ordinals and post-Taproot Bitcoin 👇 2/ Taproot has completely (and inadvertently) changed how data gets stored on the Bitcoin blockchain.

While there is no doubt this increase in state growth has been driven by Ordinals, this is really just the tip of the iceberg.

The block space floodgates have been opened.

1/ Over the past few weeks, my team at @coinmetrics has been obsessed with answering one question:

How did Alameda manage to lose billions of dollars of FTX user funds?

I think we've found some answers 👇🧵 2/ First, let's get something straight:

The line between Alameda and FTX was immensely blurred from the very beginning.

Our analysis showed thousands of large transactions from FTX to Alameda, so user funds were likely siphoned over the course of many months.

1/ FTX might have minted Serum (SRM) off thin air to prop up its balance sheet:

Serum's total supply increased by 60% this year via 2 huge mints. These were not previously disclosed based on anything I could find.

1st mint: Feb 19, 50M SRM
2nd mint: May 25, 50M SRM 2/ Earlier today, it was reported that FTX valued its SRM position at $2.2bn USD: the largest position on its balance sheet.

Serum's market cap is less than $88mn, so they likely valued their holdings using a much higher price for SRM, which is probably illegal.

Here's how you know FTX is done. Let's read between the lines.

Why would they have to "swap" these tokens instead of withdrawing them? Simple: FTX is no longer in possession of them.

This is not a "credit facility". It's not an "injection" of liquidity. Allow me to explain 👇

https://twitter.com/FTX_Official/status/1590763245488713729

The loan that FTX made to save Alameda was likely in crypto.

They lent out their customer's crypto and that's why they weren't able to meet all of their withdraws, a large chunk of their balance sheet went to Alameda.

Today, 2 FTX whistleblowers confirmed it on Reuters:

1/ I found evidence that FTX might have provided a massive bailout for Alameda in Q2 which now came back to haunt them.

40 days ago, 173 million FTT tokens worth over 4B USD became active on-chain.

A rabbit hole appeared 🧵👇 2/ That day, September 28, over 8.6 Billion USD worth of FTT was moved on-chain.

That was by far the largest daily move of FTT in the token's existence and one of the largest ERC20 daily moves we ever recorded at Coin Metrics.

Yes, the "Mock" Merge last night was eventually a success... but we did see some hiccups that might impact users when the real Merge takes place.

We've been working on it for nearly a year at @coinmetrics and thought I'd share some notes 👇 Nearly all clients (nodes) struggled to determine what was the final block of the PoW chain last night.

That's because of the high rate of reorgs (changes in block ordering) witnessed in Goerli/Prater.

22Q2 is the first time in the history of stablecoins where Total Supply decreased.

Even if we exclude UST, over 10B has been redeemed *directly from the treasuries* of major issuers

Some @coinmetrics data 👇 Tether has seen the most redemptions out of all centralized issuers, with a decrease of ~7B in total supply (on ERC, OMNI and TRX).

The sharpness of that decrease suggests that a single entity, or small cohort, was behind it.

Celsius calls itself a "network" or a "lender"

But in reality, they operate more like a highly-leveraged hedge fund.

Their business model consists of deploying user deposits across DeFi protocols with the goal of maximizing yield.

1\ A thread on what went wrong 🧵 2\ Fundamentally, when the ultimate goal of your product is to maximize yield, you need to have a solid risk framework.

You need to be able to determine when a risk is not worth the marginal yield, especially when user funds are at stake.

Pretty reasonable, right?

1\ Yesterday we learned that Chainalysis was able to deanonymize the BTC held by TheDAO hacker, even after it went through the Wasabi mixer.

Have we reached the end of privacy mixers?

🧵 👇 2\ Quick context: mixers are intended to increase user privacy by aggregating several senders and receivers within the same transaction.

Most mixers use a technique called CoinJoin to make it harder for those observing the blockchain to pinpoint specific users.

1/ It's time to have a conversation about #Dogecoin.

If you're invested in $DOGE after @elonmusk's much endorsement (such wow), there are a couple of things that you should probably know👇 2/ First, let's talk about network security (I'll try to keep it simple).

Unlike BTC, ETH, or really any other major network, Dogecoin does not have its own miners -- it currently relies on another network, Litecoin, to survive.

So how did this happen and what does it mean?

1\ There is an alarming amount of misinformation (fueled by the media) on what exactly happened to Bitcoin yesterday, and whether funds were "double spent"

Here's everything you need to know 👇 2\ On the 18th, a user broadcast a transaction with very low fees.

When users underpay fees, their transactions gets stuck because miners have more profitable opportunities.

Users are left with 2 options:

a) wait until fee levels drop
b) tell miners they will increase fees

In previous Bull Markets, the MVRV ratio was one of the most reliable indicators of market tops.

It reached 3.96 in 2017 when Bitcoin flirted with $20k

MVRV-FF is currently at 1.96

If you believe history rhymes, this suggests we're not even close to the "euphoria" phase. This is not investment advice. MVRV simply tracks how the overall market valuation (market cap) compares against everyone's "cost basis" (realized cap).

It spikes when market value is disproportionately higher than the value being moved on-chain.

What is the main driver of this Bitcoin bull run?

Some speculate that Chinese miners haven't been able to sell their BTC because of a regulatory crackdown, and that has led to a "liquidity crunch"

Lots of anecdotes, but here is some empirical evidence ⬇️ 1\ First, let's look at supply held by Mining Pools (red line) and individual Miners (green line).

As you can see, Pools (red) are not selling, but that's part of a long-term trend.

Individual miners (green) are selling, which goes against the narrative of a liquidity crunch.

Auditing $ETH's supply is not the problem.

The problem is a json file in the Ethereum genesis block listing accounts that, from day 1, held a big portion of the monetary base.

That json file created long-lasting power structures.

It introduced governance.

It created leaders. Bitcoin's equivalent to that .json file is a pseudonym that left the project after it launched, did not take a dime from it, and never returned.

That's beyond "fair launch". There were no precedents. No investors. Bitcoin was treated as a monopoly money in its early history.

Bitcoin's Epoch III has ended and it's truly remarkable to see how far it has come along since 2016.

Here are some @coinmetrics charts showcasing some of these accomplishments, taken from my presentation at #ConsensusDistributed 👇 Bitcoin settled the equivalent of *2.3 trillion USD* in Epoch III

For context, Visa+Mastercard Credit Cards settled 2.6T USD in the entire United States in 2017.

It's insane that a bootstrapped, 10yr old network settles values in the same magnitude of the largest CCs in the US.

Bitcoin mining is evolving and it is -quite literally- a beautiful thing to see.

In this week’s SOTN, @karimhelpme showcases a fascinating type of #Bitcoin data, a distribution of nonces; the magical numbers miners compete to find.

Can you guess what these lines represent? 👇 These streaks have to do with the way S7 and S9 family of ASICs sample nonces; how they approach the cryptographic puzzle in mining.

Patterns began in late 2015, which coincides with the release of the S7. In 2016, when the S9 was released, they became narrower.

How to destroy a Proof-of-Stake network in 5 steps: 1) Increase intermediation between users and the PoS network via Staking-as-a-Service third parties -- it's happening already.

The custody of the majority of staked assets centralizes around a handful of custodians.

Challenge accepted 🤜🤛

Here's a decade in Bitcoin.

The most important decade of the 21st century? 🤔

Let's see 👇

https://twitter.com/soonaorlater/status/1202658019676319744

2010 - Satoshi decentralizes Bitcoin by leaving his leadership role as its creator, and never again commenting on its development.

The Bitcoin community self-organizes, and begins to grow into new type of global institution.

Are you tired of the "Bitcoin is old technology" meme?

I just published "A Look at Innovation in Bitcoin’s Technology Stack" where I map 40+ edgy protocols being built atop Bitcoin.

medium.com/@LucasNuzzi/a-… In a lot of ways, this is an attempt to expand upon @CremeDeLaCrypto's excellent post on the emerging Bitcoin technology stack:

medium.com/blockchain-cap…