AAcquired
← All episodes
18 July 2021

Special: Solana (with CEO Anatoly Yakovenko)

5Frameworks
14Insights

Listen

Frameworks in this episode

Innovation6 steps

Backtrack From the Hardware Ceiling

Derive the performance target from the hardware's physics, then write software up to it.

Read →
Strategy6 steps

Commit to the Weaker Assumption, Then Optimize to the Limit

Once you've paid the price of a compromise, stop hedging and exploit it fully.

Read →
Strategy6 steps

Design Backwards From the Adversary

Pick the one number an attacker must beat and let it set every requirement.

Read →
Innovation6 steps

Proof of History: Scheduled Turns Instead of Races

Replace the race to produce with a verifiable clock that pre-schedules every turn.

Read →
Entrepreneurship7 steps

Recruit the Supply Side First

Win the operators who'll suffer for something rare before you chase mainstream users.

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster

Benchmarking is a game — how to read a throughput claim

Yakovenko is explicit that his own headline number is a benchmark, not a promise. Benchmarks flood the system with non-overlapping, easily parallelized transactions under ideal conditions, in the same way a chip foundry's process node names its smallest feature rather than a typical one. Real workloads like a central limit order book carry heavier reads and writes and behave differently.

  • Benchmarks use ideal, non-overlapping, easily parallelized transactions
  • Compared to a foundry quoting its smallest achievable process node
  • Serum's order books use megabyte reads and writes per order
  • Hitting a limit is treated as a three-to-four-week optimization, not a wall
  • Honest reading: the benchmark bounds the theory, not the experience

“there's a lot of caveats there because benchmarking is like I worked at a hardware company, benchmarking is a game”

“So when we benchmark, you flood the system with as many non-overlapping, easy parallelized transactions as you can.”

#benchmarks#measurement#engineering honesty

Hot Take· 2

Hot Take

You can't kill an open source community — Solana and Ethereum coexist

Drawing on watching Microsoft fail to kill Linux in the 1990s, Yakovenko argues neither Solana nor Ethereum can kill the other, because you cannot kill people who write code for fun at the weekend. He frames the overlap as healthy competition that speeds up the product-market-fit cycle, noting that kernel engineers from rival systems end up in the same bar anyway.

  • Analogy: Microsoft could not kill Linux
  • Competition is on overlapping features and a finite pool of users
  • Each community copies and improves the other's wins
  • Engineers across ecosystems are friendly; the ecosystems compete
  • Competition is treated as necessary to onboard the next billion users

“We can kill Ethereum just as easily as Ethereum can kill us, meaning it's impossible.”

“There's no killing that, right? Like it's just code that people want to write because they love to do that.”

#open source#competition#ethereum
Hot Take

Self-custody is the real adoption ceiling, not throughput

Yakovenko turns the adoption question back on the hosts, who admit they do not self-custody keys. His generous upper estimate is 30 million people worldwide who hold their own keys. Without self-custody you are still transacting through an intermediary, and he argues the layer-one performance debate is trivial next to that gap.

  • Upper estimate of self-custodying users: about 30 million, generously
  • Without your own keys you cannot connect directly to an artist or counterparty
  • DeFi summer involved 30,000-40,000 people; he imagines 30-40 million
  • The vision is flash communities forming worldwide in hours with no friction
  • Layer-one arguments compared to Intel versus AMD — engineer detail, not user reality

“I think my upper estimate for self-custody is maybe 30 million, but that's like a very generous estimate.”

“You can't directly connect to an artist and audience without your own self-custody, right?”

#self custody#adoption#crypto ux

Explainer· 5

Explainer

Why proof of work caps how fast a chain can grow

Proof of work lets anyone receive the full dataset and verify its security from the origin to the present in a black box, without talking to anyone. Yakovenko compares it to verifying gold as an element and measuring how much exists. That property only holds if a newcomer can still process the whole history, which puts a hard ceiling on how much data the chain can absorb.

  • Objective, from-scratch verifiability is unique to proof of work
  • It underpins Bitcoin's claim on store of value
  • Verification-from-scratch forces a limit on data growth
  • Holders of a trillion-dollar asset are cautious about raising throughput at all
  • The guarantee covers more than energy: no invalid state transitions since day zero

“systems based on proof of work have this wonderful property that when you receive all of the data, you can verify its thermodynamic security from…”

“And if you put that limit on it, it means that the underlying chain just can't grow that quickly.”

#proof of work#verification#bitcoin
Explainer

Settlement layer vs execution layer, and where Solana sits

Settlement is the final change of who owns what: slow, official, invisible enough that Yakovenko cannot name the company doing it, and commoditized to a thin fee. Execution is where the interesting activity lives — spot, futures, options, perpetuals. Solana targets execution, where the binding constraints are the speed of light and how much information can be compressed into a price.

  • Settlement is the bottom layer: final, secure, slow, heavily commoditized
  • Bitcoin behaves as settlement — an hour later the world considers it done
  • Execution and clearing are where derivatives and market activity happen
  • Execution limits are physical: speed of light and information theory
  • Ethereum attempts a middle position and pays for it in speed

“this is what like is the final state of the trade is when they look at like the list of who owns what and they…”

“And this is where we're talking about physics limits of speed of light and Shannon Hartley theorem”

#market structure#settlement#exchanges
Explainer

The actual throughput numbers: Bitcoin, Ethereum, Solana, Visa, Nasdaq

The episode puts hard numbers on networks people assume are enormous. Bitcoin's theoretical limit is about seven transactions per second; Ethereum runs roughly 11 to 20, around 1.2 to 1.6 million per day. Solana benchmarks near 50,000 with peaks over 65,000. Visa has measured up to 60,000 but runs nearer 5,000 at steady state, and Nasdaq handles a few hundred trades per second.

  • Bitcoin: about 7 transactions per second theoretical
  • Ethereum: roughly 11-20 per second, 1.2-1.6 million per day
  • Uniswap alone: 100,000-200,000 transactions per day
  • Serum on Solana: 20 million transactions per day, on far less capital
  • Visa: ~5,000 per second steady state; all of Twitter is ~5,000 messages per second
  • Nasdaq: two to three hundred trades per second, with far more orders and cancels

“seven transactions per second is what the Bitcoin theoretical limit is”

“And Serum, which is a central limit order book style exchange that runs on Solana, same kind of design as Nasdaq, that does 20 million…”

#throughput#benchmarks#scalability
Explainer

Why AMMs exist, and what they trade away

Automated market makers were an innovation forced by Ethereum's constraints: a hundred lines of code that let anyone create a market passively with a single signature, described as the New York Stock Exchange in code. But Yakovenko argues they are information-inefficient — an order book synchronizes far more messages and lands closer to real prices, even with less capital behind it.

  • AMMs did not meaningfully exist before Ethereum popularized them
  • Constraints drove the innovation: passive market creation with one signature
  • Uniswap has more users and more capital and is more transaction-efficient
  • Order books on Solana are message-heavier and closer to real prices
  • The trade-off is capital efficiency versus information synchronization

“It's it's a beautiful piece of math that allows people to passively create markets with capital in like a passive way with just a single…”

“But that has its own inefficiencies because it's not maximizing the amount of information that the world is synchronizing on.”

#defi#market design#amm
Explainer

Audius: why key-to-key beats a decentralized Spotify clone

Asked what makes a decentralized Spotify better rather than merely different, Yakovenko answers that you are not really using Audius at all — a fan's private key connects directly to the artist's, which signs the assets. Plays, fan relationships, and metadata sit in the open, so the platform can run off its token instead of a data-harvesting Web2 model.

  • Fans connect key-to-key with artists rather than through a platform
  • Plays and fan relationships are collected in the open on a decentralized platform
  • Removes the need for advertising or data-extraction business models
  • Roughly four million users, largely from a single EDM niche
  • Music historically leads new technology: no Napster, no Skype, no crypto

“You're not really using Audio, you are directly connecting to like Boys Noise or Dad Mouse, and it's their private key cryptography that's signing those…”

“They're four million people using Audius.”

#audius#creator economy#web3

Story· 4

Story

Solana's committed funds blew up mid-raise in the 2018 crash

Solana started at the end of 2017 into the peak of the last boom, then tried to raise straight into a fast, brutal crash. Dedicated crypto funds that had committed to the round collapsed in front of them. Yakovenko does not blame the investors, and frames the whole company as born out of that fire.

  • Project started end of 2017 / start of 2018, right at the peak
  • Crypto funds that had committed folded before the money arrived
  • Team never held more than roughly 18-24 months of runway
  • Shipping took two and a half years, not the one year they assumed

“And at that time, like the funds that were committed to us, like blew up in front of our eyes.”

“So yeah, we were born out of that fire.”

#fundraising#crypto winter#startup survival
Story

Qualcomm principal engineers said yes before he finished asking

With limited funds, Yakovenko recruited principal engineers and senior directors out of Qualcomm who had worked with him for over a decade. Their immediate willingness surprised him and served as a confidence signal about the design, which carried enormous technical risk because nobody else was working on proof of history.

  • Recruited long-time Qualcomm colleagues, not new hires
  • Instant yes read as validation of a high-technical-risk design
  • Only a few academics were working on verifiable delay functions at the time
  • Investors preferred pedigreed founders, making the unusual direction harder to fund

“I was able to pull a lot of like kind of principal engineer level folks, senior directors out of Qualcomm that were just itching to…”

“They're like as soon as I asked, they're like, Yeah, where do I sign up?”

#hiring#technical risk#team building
Story

The ASIC maker that mined with your hardware before shipping it

Yakovenko's first real crypto lesson came from a company building mining ASICs that delayed customer shipments by six months so it could mine with the units first. He treats it as a defining illustration of the space: a peer-to-peer wild west where anything goes, and where nothing you cannot cryptographically verify should be trusted.

  • Company built dedicated mining semiconductors and mined with them before delivery
  • Hosts call it front-running your own customers
  • Lesson: you cannot know who is on the other side of an internet connection
  • If you cannot verify the keys or the software, assume nothing

“but they built them and delayed the actual shipping them to customers for six months so they could mine with them first”

“Like if you can't verify their keys, right? If it's not software that you can verify, then you can't assume or trust anything.”

#trust#mining hardware#crypto history
Story

"Is it live?" — the demo that turned FTX into Serum

Nine months before launch, Solana pitched FTX on building something on chain and got asked whether it was live; it was not, and the conversation ended. Six months after launch they built break.solana.com, where you smash keys and watch smart-contract transactions fire and confirm on screen. Yakovenko demoed it to Sam, who immediately wanted his engineers to see it.

  • First pitch failed on a single question: is it live
  • break.solana.com made speed visceral rather than claimed
  • The demo was aimed at engineers, who could inspect the code
  • FTX's traders reframed it as decentralized finance potentially taking a large share of global finance
  • Solana had planned to build its own DEX but partnered instead

“And they're like, Is it live? Or like if we think in about six months, like, okay, two weeks, two weeks, yeah.”

“we built this little demo called break, like break.solana.com, where you smash your keys and you see transactions fire off”

#partnerships#demos#serum

Takeaway· 2

Takeaway

The pitch to app developers: a billion users, no shards, no layer twos

Asked directly why a well-resourced developer should build on Solana rather than wait for Ethereum's next version, Yakovenko gives a single-sentence answer about guaranteeing a connected world for a billion users without shards or intermediaries. His secondary argument is availability: the alternative is not live yet, and his team optimizes for user experience rather than protocol internals.

  • One promise: a billion connected users without shards or intermediaries
  • Explicitly declines to care about anything outside that goal
  • Availability argument: the competing design was still theoretical
  • Framed as caring about experience rather than the underlying hardware analogue
  • Relationship pitch: willing to eat glass for founders with a big vision

“Because we guarantee this super connected world for one billion users without shards, without intermediaries, this is where it's gonna happen first.”

“So you can go deal with shards, layer twos, all this other stuff if you want.”

#developer pitch#positioning#scalability
Takeaway

The internet was supposed to be decentralized — 1985 you would be shocked

A host observes that describing the internet to someone in 1985 would predict a decentralized world where anyone can connect, publish, and start a business, and that centralization into five big companies would have astonished them. Yakovenko agrees and uses it to explain why Solana ignores the sovereign-money narrative: the token exists mainly to prevent spam in a hyper-connected world.

  • The technology's natural shape looked decentralized from the start
  • Centralization into a handful of companies was the surprise, not the default
  • Solana's stated goal is guaranteeing anyone can self-publish without interference
  • The token is framed as spam prevention, not as money
  • Compared to a science-fiction city where everyone can talk to everyone

“if you were to tell me about the internet in 1985, I would believe that the natural way the internet would end up would be…”

“And it's this sort of surprising thing that happened that it led to centralization, where 1985 me would be shocked that there are five big…”

#decentralization#internet history#web3