Do You Want To Be Right, Or Do You Want To Be Wrong In Public?
A retroactive look at the first wave of predictions from A Crypto Future
“There’s no time to ask the meaning of things, no time to answer, and anyone who tries to find out is considered a great fool…” - Natsume Sōseki (The Miner)
More than 2 years ago now we wrote A Crypto Future, inspired by a LessWrong post about the future of AI. I’ve spent the better part of two years snarkily sharing headlines in a running thread whenever I see something related to one of those forecasts. Conveniently ignoring the things we got wrong of course.
But it’s time to pay the piper. So even though we’re only ~2 years into what was a 5+ year extrapolation, I thought it would be fun to rip through the 2023-2025 forecasts and do a bit of self-reflection. I also promised two new blogs before the end of July so it’s time to get moving1.
First, some context.
We published this May 25, 2023 when BTC was trading ~$25k, ETH ~$1800 & SOL ~$20. Silicon Valley Bank had failed just 8 weeks earlier. Gary Gensler was absolutely bucking his head. Hyperliquid’s mainnet closed alpha launched ~30 days prior. It’s easy to look back today and say all of what’s transpired was inevitable/obvious but this is mostly self-psyops in my opinion2. Part of the deal with writing in public is that you need to be comfortable looking stupid. But this is a small price to pay for the near infinite benefit it brings. Everyone should be writing on the internet more (and not just AI-slop pieces).
Where relevant I’ve added some color and in some cases a bit of perspective on how my view for the future has changed. I tried to grab anything that was specific enough to evaluate but feel free to re-read the entire original piece if you want to nitpick me :)
“DeFi quietly surpasses 30 million total users, with most of the new activity flowing into permissioned, KYC‑gated products. The discourse is mixed as some delight in watching their bags increase in value thanks to wider participation and a continued push into the space from traditional institutions. Others claim the essence of crypto is being undermined by these products”
Result: 1/2 right. I don’t particularly trust the data around active users and I think it’s fair to say the new activity is not necessarily flowing into permissioned, KYC-gated products (DAT’s aside). But the latter half of this has happened in some respects.
“OG” permissionless DeFi rebounds only modestly.”
Result: (Thankfully) Mostly wrong. You could argue about token price & FDV specifics but this wasn’t really the crux of what I was suggesting in the piece.
“US Treasuries have been brought on-chain en masse.”
Result: Mostly right. ~$475M at time of publishing, ~$7B today.
“Stablecoin volume crosses $500 billion”
Result: Lay-up.
“On-chain options markets are still not there yet”
Result: Correct. Perps are too good of a product still.
“The most exciting novel primitive is a mobile wallet application that has abstracted away all of the difficulties for non-native users and offers multi-chain, multi-asset yield generation.” incorrect
Result: Wrong. I’ve yet to see this in practice though there have been some teams pitching adjacent ideas to this. Still unsolved and at this point, it would not necessarily be considered a novel primitive imo.
“Games are not the killer use-case that brings new people into the space”
Result: Prepare for some aggressive self-fluffing…
“Hundreds of millions of dollars in funding for crypto games has very little to show for it by the end of 2024. Trailers continue to be released but no crypto-native game sees any meaningful sustained traction”
Result:
“The unlock in ‘user adoption’ comes from more intelligent, user-friendly, guided-search that makes it easier for normies to explore on-chain”
Result: Wrong. I think there have been some attempts at improving explorers and surfacing content for normies but it’s not as far along as I expected. There’s an argument that pump.fun unlocked some of the ease-of-use for a specific niche (finding & gambling on memes) but that’s not the spirit of what I was saying.
There are advanced on-chain search buddies that integrate seamlessly into on-chain data and wallets. ‘Show me the most popular NFTs over the last month’, ‘Buy $100 of ETH for me over the next 2 hours’”
Result: This is correct but admittedly not that interesting apparently? Cortex was the first version of this I used and I’ve seen a handful of other examples but perhaps users are not as interested in this type of product as I expected. I do wonder if pre-prompting specific details about tokens might make a next-gen version of this useful for trading bots and terminals.
“Saga sales aren’t astronomical (50k-100k units sold) but Solana establishes itself as the lead horse for mobile-focused non-native developers”
Result: Half right (sold 20k) and directionally correct I think, but the latter part of this is still unclear.
“Some social apps are temporarily intriguing but the depth of differentiation & new behavior enablement is mostly shallow. Though there is a sense by the end of 2024 that this vertical finally has enough talented crypto-product people to build novel applications normies will actually use”
Result: Mostly correct. We had FriendTech launch later that summer and then a couple other social apps (Fantasy Top maybe the next best example) but none have broken through in the sense of engraining a new behavior for users. I think the jury is still out on the second half of this prediction and tbh the [blank]dot[blank] conveyer belt is exactly the thing I was worried about.
“Electric Capital’s developer report for 2024 will reveal we’ve crossed the 75k monthly active developer mark & 25k full-time developers”
Result: Incorrect. 2024 report was ~23k developers & 7.7k full-time devs.
“Privacy moves to the backburner as animal spirits take hold and the bull market marches on”
Result: Mostly correct. Privacy continues to ripple below the surface in my view.
“Crypto becomes a partisan issue in Washington with Republicans calling for a looser regulatory framework while fear mongering about the US losing to China. On the other side, Democrats call for more oversight, regulation & freedom limitations, leaning hard into an anti-crypto position.”
Result: Unfortunately correct.
“The worries about a dystopian CBDC cool given the underwhelming launch of FedNow.” correct
Result: I legit cannot tell you the last time I heard someone talk about FedNow.
“When a Republican - as a slight underdog - wins the presidential election, many in crypto over-index on what this means from a regulatory perspective and kick-off an undeserved period of mania”
Result:
“Gary Gensler remains in his position and a thorn in the side of crypto, though his influence wanes”
Result: Wrong but directionally correct3, he actually got nuked quicker than even I expected. One of the worst to ever do it.
“One of the major ‘seemingly safe’ protocols (Uniswap, Aave, Curve) falls victim to the largest hack in crypto history. ‘Security’ as a narrative becomes en vogue for investors’”
Result: Half right as we did see a big hack on Curve just a few months later, but I’m not sure the security narrative ever really saw the type of attention I expected (yet).
“The crypto-to-AI VC pipeline turns out to be real. Lots of LP capital gets burned by investors who didn’t really understand crypto beyond a surface-level and definitely do not understand AI.”
Result: One of the easiest calls in the moment to fully shove.
“By the end of 2024, there’s still material skepticism from the outside world about ‘what has really been built’ but for those within crypto, this strawman argument no longer carries weight.”
Result: I think correct? But I’m biased
“The hype cycle is now in full swing.”
Result: The sitting President launched a memecoin for every one of his family members. It’s safe to say this was correct. Was it also an Easter egg for Hyperliquid bullishness? No. But will I claim that it was in 5 years? Perhaps.
“Gensler is a lame duck sitting SEC Chair and it’s clear he won’t be back when his term ends in 2026.”
Result: Already covered earlier. Worst to ever do it.
“While the CFTC hasn’t officially been deemed the regulator of choice for crypto, it’s evident that’s now just a formality.”
Result: This has largely played out as expected.
“All of the major banks and brokerages now offer crypto services of some kind.”
Result: We have already accelerated to this stage…
“Every sell-side desk now has a dedicated crypto-asset team”
Result: This is wrong though there are obviously crypto teams within banks. But this was predicated on the idea we would have more “analyst coverage” of crypto-specific names beyond the occasional mid take on BTC or ETH. Feels inevitable though. Especially since they’re all getting one-shotted by new Excel AI tools.
“JPMorgan unsuccessfully attempts to launch JPM-chain but it dies quietly”
Result: TIL Onyx rebranded to Kinexys in November of last year.
“Structured products markets on-chain are thriving.”
Result: Correct but we sometimes mislabel them as stablecoins.
“A lot of alternative lending activity has moved on-chain. There is now a robust market of on-chain entities who periodically issue bonds secured by this collateral”
Result: Wrong. There’s a handful of teams trying to solve this problem in different ways (Cycles, Wildcat, 3jane) but we haven’t yet seen credit return onchain. It does feel as if the market is still starved for it though.
“Ethereum’s burn is up to 15k ETH/day”
Result:
“An existential discussion becomes very divisive as one group looks to adjust the share of rewards given to stakers, while the other side argues this type of change is sacrilege…there are real concerns about long-term sustainability if nothing is done”
Result: I think this is mostly incorrect though there are the seedlings of a reconsideration of proof-of-stake. Perhaps the meme “economic security is a meme” has more truth to it than originally thought. PoW blockchains making a resurgence over the next 12-24 months would not surprise me.
“A steady-state ETH yield has become an easy meme for Tradfi to grok”
Result: It took Tom Lee shilling with all his heart but it seems to be resonating a bit more recently.
“There are 30+ protocols generating meaningful fee revenue ($100M annualized)”
Result: I kind of cheated here it seems. “Fee revenue” is a bit ambiguous. If we go by fees we’ve got 32. If we go by revenue we’ve got 15. I used Defillama, please don’t argue in the comments about this one.
“Some of the more mature protocols are now valued based on multiples of fee revenue but there is still discussion about whether flows are more relevant”
Result: I now suspect this debate will never go away. Increasingly it feels as if tradfi is moving directionally toward crypto (narratives are increasingly important relative to fundamentals) while crypto moves directionally toward tradfi (fundamentals in crypto are more important now than they were 3-4 years ago). Flows will always matter though and it’s one of the core reasons Bitcoin has been so strong the last 2+ years.
“BTC & ETH ETFs are finally approved, bringing further liquidity to the space and opening the door to wider retail participation”
Result: LFG. Imagine ever doubting.
“Stablecoins now account for the majority of transaction volume in a select number of LatAM & Asian countries”
Result: True in Argentina & Brazil but doesn’t appear to be the case anywhere in Asia at the moment.
“Digital assets are in the top 5 of most widely held assets across Asia, behind just equities, cash, fixed income & real estate”
Result: Damn this was close. Currently sits at 6th behind the 4 listed above and life insurance / pension reserves. Also proportionally more owned by wealthier classes within Asia.
“More & more people are wearing the next generation of AR devices, which are smaller, more comfortable and can be worn much longer”
Result: Technically correct but spiritually wrong as it hasn’t moved as quickly as I anticipated.
“Crypto projects have taken to AR advertising as a means to solidify mindshare among younger-generations”
Result: Complete whiff. I was far too optimistic here.
“V2 of the Saga is released and its markedly better than v1 (500k units sold)”
Result: TBD. I believe this ships next month and it pre-sold 150k units. This one also ties back to the Solana mobile dev question.
“Lots of liquidity & higher prices has led to renewed speculative fervor. This time it takes the form of crypto gambling adoption.”
Result:
“The collective crypto gambling industry now dwarfs the annual revenue of traditional players like Wynn & MGM (5x larger)”
Result: Wynn did ~$7b in revenue last year and MGM did ~$4b. So not quite.
“The first real scaled use-cases of zk tech become clear and it’s happening within the growing DeSci community. Health data protection & collection has been thrust to the forefront following continued leaks of sensitive health information”
Result: Admittedly this has moved slower than expected but the undercurrents continue bubbling up.
“Multimodal transformer sizes have become hilariously large (many now >5 trillion parameters) and are trained on images, video, audio and movement”
Result: I’m not 100p sure here but there are suggestions anywhere from 2 trillion to as high as 5-15 trillion for GPT 4.5. Hilariously large nonetheless.
“Multiple companies are now each spending >$1 billion/yr to train. “We see a massive spike in demand for latent GPUs”
Result: Not quite at $1b to train but the capex spend makes this number feel insignificant in the grand scheme of the AI buildout. I suspect there will be 2-3 massive winners in the adjacent crypto space.
“We’re not quite there yet but there are smaller groups of people across the energy & crypto spaces who are seeing real traction building independent infrastructure models, specifically distributed energy resources”
Result: This is absolutely happening already and incidentally became a narrative that investors glommed onto at some point in 1H ‘24.
Some quick back-of-the-napkin math says we went somewhere around 30 for 46. Honestly that feels about right. The space has directionally moved the way we expected with many of the most critical pieces falling in-line with the future we initially believed in. After re-reading the original piece I’m feeling a bit nostalgic for that period of time | there was quite a bit of despair on the timeline, a lot of people were still reeling from 2022 and feeling unsure about crypto’s future. It felt meaningful to share something much more optimistic, that we spent a lot of time on and genuinely believed in. Seeing crypto move forward4 since then has been exciting at times, frustrating at others but entertaining all the way through. There is no second best industry.
“A wrong concept misleads the understanding; a wrong deed degrades the whole man”
i am giving myself an extension to tuesday aug 5th
also why i am a journal maxi
this is the proper way to cope about being wrong
albeit with plenty of frustrating realities still in place