On November 4, 2017, some of the brightest minds in cryptocurrency research gathered at Stanford University for Day 1 of the Scaling Bitcoin conference, an event that would prove pivotal in shaping the technical roadmap for Bitcoin and blockchain networks for years to come. The conference, which brought together students, experts, and researchers from around the world, was explicitly focused on engineering solutions rather than the political debates that had dominated Bitcoin discourse for months.
TL;DR
- Scaling Bitcoin Conference Day 1 held at Stanford on November 4, 2017
- Ian Miers from Johns Hopkins presented “Bolt,” a system for anonymous payment channels
- Stanford researcher Benedikt Bunz unveiled “FlyClient,” shrinking Ethereum block headers from 2.2GB to just 3MB
- Bitcoin Unlimited demonstrated Gigablock Testnet results showing software — not hardware — was the real bottleneck
- Proposals for Bitcoin script modifications aimed at strengthening payment channels were also presented
Anonymous Payment Channels: The Bolt Proposal
One of the most intriguing presentations came from Ian Miers of Johns Hopkins University, who detailed a research paper titled “Bolt: Anonymous Payment Channels for Decentralized Currencies.” The techniques described in the paper allowed for the construction of anonymous payment channels designed specifically to reduce the storage burden on payment networks.
Miers, who is also one of the key scientists behind Zcash, acknowledged that while the research was applicable to Bitcoin development, it was more likely to be implemented in privacy-focused networks like Zcash. However, he noted that even for Zcash, this particular improvement was not a top priority at the time. The significance of the Bolt proposal lay in its demonstration that payment channel privacy was an achievable goal — a concept that would influence subsequent Lightning Network development and privacy protocol design.
FlyClient: Super Light Clients for Cryptocurrencies
Perhaps the most immediately practical presentation of the day came from Stanford University’s own Benedikt Bunz, who proposed a new concept called “FlyClient.” The idea addressed a fundamental challenge in cryptocurrency: as blockchains grow, the burden on light clients increases. While Simplified Payment Verification (SPV) clients don’t grow linearly with the number of transactions, they do grow with the downloading of block headers — and over time, this data accumulates significantly.
FlyClient employed a clever statistical method that eliminated the need to download every single block header. When tested on the Ethereum network, the results were striking: 2.2GB of header files were compressed to a mere 3MB. This dramatic reduction had profound implications for mobile wallets, IoT devices, and any application requiring lightweight blockchain verification. The technique also showed potential applicability for Bitcoin Cash and other networks seeking to maintain on-chain scalability.
Gigablock Testnet: Proving Hardware Isn’t the Limit
For proponents of on-chain scalability, the standout presentation was undoubtedly Bitcoin Unlimited’s Gigablock Testnet results. Titled “Measuring Maximum Sustained Transaction Throughput on a Global Network of Bitcoin Nodes,” the research demonstrated that the limitations facing Bitcoin’s transaction throughput were not hardware constraints at all.
Instead, the bottlenecks identified were entirely software-related. Among the optimizations proposed were tweaks to Satoshi Nakamoto’s original code and the introduction of a “try later queue” mechanism. The finding was significant because it challenged the prevailing narrative that Bitcoin’s block size limit was a necessary constraint imposed by hardware realities. If software optimization could unlock dramatically higher throughput, the debate over on-chain versus off-chain scaling would need to be reconsidered on different terms.
Bitcoin Script Modifications for Stronger Payment Channels
Johnson Lau and Olaoluwa Osuntokun also presented proposals for modifications to the Bitcoin scripting language, designed to enable further functionality and strengthen payment channels. These script improvements were part of a broader effort to make Bitcoin more programmable and better suited for the complex conditional transactions that payment channels require.
The proposals reflected a growing recognition that Bitcoin’s scripting capabilities, while deliberately limited for security reasons, needed strategic enhancements to support the second-layer ecosystem that was rapidly taking shape around the network.
Why This Matters
The Scaling Bitcoin Conference on November 4, 2017, took place at a critical inflection point for the cryptocurrency ecosystem. Bitcoin had just crossed $7,300 and was surging toward its eventual December 2017 peak near $20,000. The scaling debate had reached fever pitch, with the SegWit2x fork controversy still unresolved and the Bitcoin Cash hard fork having occurred just weeks earlier in August.
The research presented that day laid foundational groundwork for technologies we now take for granted. FlyClient’s concepts evolved into the light client protocols used across modern blockchain networks. The payment channel research directly informed the Lightning Network’s development trajectory. And the Gigablock Testnet’s findings continue to inform the ongoing debate about blockchain scalability limits.
In an era where the crypto community was deeply divided between on-chain and off-chain scaling philosophies, this conference demonstrated that both approaches had legitimate, well-researched paths forward — and that the ultimate solution would likely involve elements of both.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Past performance is not indicative of future results.
2.2GB to 3MB block headers with FlyClient was insane. Bunz rarely gets enough credit for this work
FlyClient was the precursor to modern light client proofs. Bunz basically laid the groundwork for what zkSync and StarkNet use for state proofs
flyClient to modern zk proofs is a direct lineage. bunz should be mentioned in every history of ethereum scaling but somehow gets skipped
zk_proof_ FlyClient to modern zk proofs is the most underrated lineage in crypto. bunz should be in every scaling history but gets skipped for Vitalik tweets
Benedikt Bunz shrank ethereum headers from 2.2GB to 3MB and somehow gets zero credit in mainstream crypto media. every zk L2 traces back to that paper
2.2GB to 3MB was basically magic in 2017. half the zk infrastructure today traces back to that one paper
Petra M. Ian Miers presenting Bolt at Stanford while the block size war was raging was peak cypherpunk. anonymous payment channels proposed in 2017 and Lightning privacy in 2026 is still barely usable
FlyClient shrinking Ethereum block headers from 2.2GB to 3MB was the foundation for every ZK light client that came after. Bunz rarely gets enough credit outside academic circles
bunz_credit FlyClient to modern zk light clients is a direct lineage. every zkSync and StarkNet proof traces back to that 2.2GB to 3MB compression
Ian Miers presenting Bolt at Stanford while the block size war was raging was peak timing. technical solutions while everyone else was fighting politics
ian miers bolt presentation at stanford in 2017 was peak timing for anonymous channels
Stanford hosting Scaling Bitcoin while the block size war was tearing the community apart was a nice contrast. science over politics for once
Ian Miers presenting Bolt while the block size war was tearing BTC apart was peak academia vs politics. anonymous payment channels proposed in 2017 and LN privacy still sucks in 2026
flyclient took block headers from 2.2gb down to 3mb and zk proofs followed right after
Bolt anonymous payment channels in 2017 and people are still debating privacy on lightning in 2026. the research was way ahead of the implementation
9 years later and lightning privacy is still barely usable for normal people. the research was there, the UX was not
Kai R. lightning privacy in 2026 is still basically non-existent for normal users. Bolt proposed anonymous payment channels in 2017 and the research just… sat there for 9 years
lamport_sigs the academic-to-production pipeline in crypto is broken. FlyClient, Bolt, all brilliant papers that took a decade to get anywhere near implementation. meanwhile meme coins ship in a weekend
Ravi Chandra the academic to production pipeline is brutal. Bolt took 9 years to get nowhere while meme coins ship in a weekend. incentives are completely broken
Ian Miers proposed Bolt anonymous payment channels in 2017 and lightning still leaks metadata 9 years later. the gap between academic crypto research and actual shipped code is painful
Bolt proposed anonymous payment channels in 2017 and Lightning still leaks metadata in 2026. the gap between crypto research and shipped privacy is embarrassing
Gigablock Testnet proving software not hardware was the bottleneck was the most important result from Stanford and nobody talks about it
Elina J. exactly. Bitcoin Unlimited ran Gigablock and still couldnt break 30 tx/s because the propagation code was the bottleneck not the block size itself
ian miers bolt presentation at stanford in 2017 was peak timing for anonymous channels. block size war raging and he was solving actual problems
the academic to production pipeline in crypto is brutal. Bolt took 9 years to get nowhere while meme coins ship in a weekend. incentives are completely broken
Ravi Chandra Bolt took 9 years to get nowhere while meme coins ship in a weekend. incentives are completely broken. FlyClient had a similar fate honestly