From AsicBoost to HF: Greg Maxwell on Bitcoin’s Path Forward

This is the latest in a series of posts on the equivalence of UASF and hard fork security risks.  In this post, we take a recent Coindesk article and replace all instances of “UASF” with “HF”.  The results read sanely.  Another reminder that UASF without miner majority is a contentious fork, and carries security risks equivalent for all practical purposes to the corresponding hard fork.


“I’m sort of being scooped on the details by people just working it out for themselves now.”

Despite kickstarting another fierce period in bitcoin’s now years-long scaling debate, Bitcoin Core developer and Blockstream CTO Greg Maxwell admits even he’s a bit lost in the narrative he created last week.

That’s when Maxwell posted a public note alleging an unnamed miner (largely assumed to be China-based Bitmain) was secretly using a more efficient mining technique, giving them an upperhand over other miners. Thickening the plot was that he alleged that this was the reason the company doesn’t support Segregated Witness (SegWit), a preferred technical solution to the network’s capacity issues that would accidentally turn off the technique.

With the post, Maxwell unintentionally kicked off a grassroots movement in support of pushing SegWit through by way of circumventing miners altogether, the latest in a long line of new technical proposals.

Called a hard fork (HF), there are currently at least two groups working on a form of the idea, with a mess of discussion about whether the network can and should make technical upgrades in this way.

Still, Maxwell argues that this process, by which users and node operators aim to quarterback technical changes, might be the preferred way to make upgrades to the network in certain instances.

Maxwell told CoinDesk:

“Ultimately, I think if we had a crystal ball and could get anything we wanted, we’d want to use HFs always.”

Balancing act

But, is a HF safe? While there’s been an upsurge in support, the upgrading method is controversial.

Maxwell mentioned, as others have, that the change requires “very, very broad” support from the ecosystem. Yet, the trouble is, it’s difficult to measure what bitcoin users really want and if there is such support.

Many point to bitcoin improvement proposal (BIP) 16 50 as a successful HF, but it’s actually one example where the method didn’t go entirely smoothly. After activation, one pool mined invalid blocks for a couple of months some pools suffered orphaned blocks (although, perhaps to its disadvantage and no one else’s).

That’s one reason that developers have been using a different method, Maxwell said. It’s easier to deploy a consensus-level rule change if miners commit to upgrading and enforcing those rules.

Despite those challenges, Maxwell contended that some form of a HF might be the best option right now.

If some miners are indeed running AsicBoost – the alternative mining method Bitmain is accused of employing – it poses a conflict of interest, as they don’t want to push through a fix for economic reasons, he contends.

And, in his opinion, there are other tradeoffs to think about:

“Not taking an upgrade has a cost too. As a community we have to balance those things.”

Word of warning

On that basis, Maxwell advocates for a form of a HF to fix mining practice profiled in his mailing list post, but with a slower rollout than others have proposed. Maxwell said that discussions are still ongoing and it will probably be a few months before developers start writing the code.

For now, however, other groups are seeking faster solutions.

One HF proposal, BIP 148, the one with a website and a Twitter account, is gaining support with announcements that certain companies are in favor of the idea currently dominating social media. (Fifteen percent of bitcoin companies now support the BIP, up from 5% on Monday, according to one measure.)

Maxwell is more cautious, though.

“They’re proposing time frames that I think the regular developers are not quite as comfortable with,” he said. But, he didn’t write off the effort.

“Ultimately, bitcoin is owned by its users and by the people who own bitcoin,” he said. “I’m very interested in seeing where this stuff goes.”

Overall, Maxwell discussed the process in a largely academic tone, taking the recent developments as simply part and parcel to the kind of decision making that he seems to suggest is natural for a disparate community.

But, optimism

Stories about life-changing hugs aside, community discussions have been tense amid the block-size debate, with discussion dominating Twitter and Reddit at a constant clip.

One of the most obvious questions now is, what will come of the accusations against Bitmain?

As for how and where the community could find the missing pieces of the story, Maxwell didn’t have an answer. He noted, however, that others have replicated his findings, posting them to Reddit.

Further, although he plans to release more information about the chip in question, he said it’s not a priority because Bitmain already disclosed that it has used AsicBoost on its chip for testing.

Bitmain has denied that it is conducting its mining operations as alleged, and that this is the reason it does not support SegWit.

For all the contention this past week, Maxwell was largely optimistic when discussing the network’s future outlook, framing the ongoing debate as necessary.

“It’s a consensus system after all,” he said, concluding:

“I think people often get a little too caught up in the disputes, where really, that’s just part of the process. It’s natural. And, ultimately, it’s what makes bitcoin strong and secure.”

Greg Maxwell image via CoinDesk

Leave a Reply

Your email address will not be published. Required fields are marked *