2023-02-20 07:19:49 UTC


NFTs reached euphoric heights in 2021. The hangover this year has been equally painful. Amongst the wreckage, many of the usual skeptics piled on and predictably called for the end of NFTs. But I'm licking my chops for NFT infrastructure opportunities this year. I predicted last year that the NFT market would play out like 2013-2015 in bitcoin:
"Yes, they're like the ICOs of this cycle: sky high hype, crazy volatility, lots of early lottery winners and complete garbage. But as a new asset type and class, they will transform the world…they represent verifiably scarce, portable, and programmable pieces of digital property… which makes the potential for NFTs essentially unlimited as blockchains become global transaction ledgers for both natively virtual property and physical property.
Physical art is a $1.7trillion asset classwith ~$60 billion of annual sales volume. At the end of Q3, NFT market cap was just $14 billion according to DappRadar. Digital art NFTs have represented less than $2 billion in total sales to date. So NFTs are still less than 1% of the physical art market, and digital art is just one tenth of the total NFT market.
Do those numbers remind you of any other four-year-old asset?
Despite the froth of today's NFT market, this early flight from physical art to digital art could end up looking like 2013's bitcoin "bubble," which crashed 80%+ in 2014, but also marked the beginning of bitcoin's decade-long ass whooping of physical gold. Bitcoin's market cap crossed 0.1% of gold's in November 2013. Wouldn't you know it? The "non-collectible" digital art market is now 0.1% of the physical art market. The digital art/ NFT market crash will eventually be even more nauseating than the 2015 bitcoin bear market (because these are highly illiquid assets by definition), but the 10 year trajectory of the overall market will be the same: 100x+."
I'm sticking to that thesis, and everything we've seen this year hardens my conviction in the NFT space and its potential moving forward.

"Blue Chip" Pixelated Art

Given their non-fungible (and thus, highly illiquid) nature, the size of the NFT market is hard to assess. The best estimate that I've seen comes from the team at NFT Valuations, which is supervised by the crypto faculty at the University of Nicosia (one of the earliest academic institutions with a real crypto footprint).
They peg the current NFT market cap at about $7.5-10.0 billion, with two thirds of that coming from profile picture ("PFP") collections like Crypto Punks and Bored Apes.


Is this sustainable? Allow me to analogize.
I'm sort of obsessed with Yellowstone and its Kevin Costner-led cast of cowboys and misfits. It's a good show, and you should watch it. (Succession and Yellowstone are the same show but one is set in New York City and the other is in Montana.)
One of the things the guys (and one gal!) on the Yellowstone ranch do is stamp new recruits with a branding iron once they've proven to be sufficiently loyal. You wear the big Yellowstone "Y" scar tissue with pride and people know you're part of the family. Since that family is formidable, people know not to mess with folks that wear the brand.
Anyway, if you understand why ranch cattle brands and gang tattoos are interesting community markers, it might help you grok PFP NFTs. Punks are arguably original collectors' items because they were the first big NFT collection. Everything else seems like more of a head scratcher unless you think about them as clique identifiers.
If you want to feel like you belong somewhere, it's a worthy investment for anyone who was early in crypto, has spare cash, and is looking for a new family, now that their old family hates them for shilling crypto last Christmas.
As for me, I'd prefer to sear my flesh with scalding iron vs. dropping $5k on a Pudgy Penguin.
Jokes aside, the Bored Ape Yacht Club ("BAYC") creator Yuga Labs had an interesting year (they're now the second largest NFT project by "market cap"), and the project checks my box for "looks like innovation wrapped in a joke/toy."
I won't comment on the merits of the ApeCoin token itself, because frankly I don't understand it, and that's why I pay these guys.
That said, Apecoin is a novel experiment. Basically, Yuga Labs said:
"BAYC (and its sister projects Mutant Apes Yacht Club and Bored Apes Kennel Club) want our pfp to be useful, and to represent unlimited access to a vibrant community of NFT OGs that meet up, wear the same hoodies, and colonize the metaverse together. Let's do a governance token, and then let people buy plots of the BAYC metaversewith that more fungible currency."
And the community said, "Nice."
While I don't understand the allure of this particular virtual community of nearly 100,000 human souls,  I also don't understand TikTok dances, the Kardashians, or how Dogecoin and Shiba Inu are still more valuable than Uniswap. In the meantime, their community is one thing that actually seems to be growing in this hellscape of a year.

Generative Art

I understand digital art slightly better than I do PFPs. Art Blocks, the "premiere" generative art NFT collection, is the fourth largest NFT project by volume, and its various collections represent four of the top twenty collections by floor price. Art Blocks sells things like pictures of squiggles for thousands of U.S. dollars that less enlightened heathens might otherwise use for things like food and shelter.

The NFTs look like this:


Art is art, and its value is in the eye of the beholder, and the market. I also get that most physical and digital art pieces are Veblen goods, so I cringe conceptually as to why these things have exploded.
But that's not why I'm writing about literal squiggles. Instead, I try my best to find the innovation that's buried under the surface of the toys.
For generative art, it could be AI-generated NFTs that combine to create game assets, characters, icons/designs, music, content, etc. and secure all of that prompt-driven IP. Think of a designer or copywriter proficient in GPT-3 prompts that wants to secure all iterations of a creative work or mark.   A photographer will take many photos and select one. Likewise, generative artists will tweak images with computer code and sell the best of the bunch. Or you might be able to combine NFTs and have generative art reputation NFTs that look and feel "elite." Much like a military officer's regalia, you can use these as pictures that speak a 1,000 words about your background, character traits, community, skills, and accomplishments.
Or you could run genomic simulations to create the perfect offspring. 

Dystopia or evolution?

Depends on your worldview, I suppose. But given what we've seen from OpenAI this year, I'd get used to the realm of science fiction becoming real sooner than expected.
(Required Reading: Revisiting Art Blocks)

The Institutions Are Here, Just Not the Ones We Expected

NFTs proved to be big business for big brands last year. Luxury brands (Gucci, Dolce & Gabbana), professional sports leagues (NBA Top Shots), and food and beverage companies (Starbucks) alike all hopped aboard the NFT hype train, and the low marginal cost (windfall profit) products it gave them last year when the market was hot.


Very few of these collections maintained their value over the past year, but we've continued to see strong moves from legacy brands and non-crypto tech companies (bully for Salesforce who launched an NFT Cloud!) looking to integrate NFTs into their platforms. Reddit launched a collectible avatar program, which has onboarded 3 million unique users. Meta is creating an NFT marketplace that will be integrated into the Facebook and Instagram platforms. I'm betting that we'll continue to see more experimentation with NFTs under Elon's Twitter (more on this in the decentralized social section below).
Fashion may be going digital the fastest. Though digital fashion is not a new phenomenon - in-game virtual goods and skins is now a $50 billion market - tokenized fashion opens up a world of new opportunities for brands.
Whether it's digital only (NFTs that can be worn in games and metaverses and transferable and composable across platforms, ideally) or digital/physical hybrids (NFTs that can be redeemed for physical items, or rendered on a person in a picture or video), there is plenty of demand. Gucci sold a digital version of one of its physical bags on Roblox, and it sold for $800 more than the "real" thing. Digital drip will become more commonplace as gamers demand more customization around their digital identities.

I'm more excited for brands and artists than I am for speculators. I think the mini-era of NFTs as "investments" is largely over. The era of NFTs as digital "consumables" is just getting started.


Whenever I hear someone pitch crypto gaming as a key sector to watch, this is how I feel.

Don't worry. I understand the thesis, and our team has done a good job explaining STEPN, Axie, and other early top games.


Web3 gaming is all about giving players ownership of their in-game digital items and using tokens to create in-game economies (and potentially, game governance). Players with a stake in the game they're playing, whether through in-game items or quasi-equity (tokens), will feel more invested in the game's success and spend more time and money playing. If you hit critical mass amongst players, they'll want to buy and sell items (NFTs), and an in-game economy can emerge that shifts the traditional gaming business model.
Instead of games extracting value from players via Downloadable Content (DLC) and micro-transactions, game creators can make money as marketplace operators (transaction fees). Users are incentivized to buy or earn top game items (the Play-to-Earn ("P2E") model popularized by Axie Infinity).
There's three problems, though.
  1. To build a sustainable P2E gaming franchise, the game needs to be, ya know, fun. Was Axie fun? Or was it a flash in the pan, bubble, or success story on par with the 2,000% APY DeFi pools that spiked for a moment, then tanked? My hunch is that more people were treating Axie as an odd job last year and not a pastime.
  2. To make a fun game, you generally need a lot of money. Development timelines are long, and final products are expensive. Game development is significantly longer than software due to the art components and additional complexities. The higher the quality of the game, the longer the timeline. When NFT game items and complex token mechanisms are also thrown into the mix, game development might be even longer and harder to change compared to other areas of crypto software.
  3. There's no denying that Axie knocked it out of the park with $1.35 billion in lifetime protocol revenue, more than any other dApp and $450 million more than the second highest grossing protocol, OpenSea. But we're looking forward, not backward, and here's the trend:

Token Terminal)
You literally cannot see fees since the spring using this y-axis because they are down to less than $1 million / year on an annualized basis.
If you think token trading, gambling, and financial negotiations are interesting, there are ways to express that passion productively in society. Namely in finance (legacy or crypto). I'm not sure adding tokens, financial alchemy, and creating steep player buy-in costs is a good way to build a fun game, or at least not one with a very large community.
I fully understand and appreciate that the gaming market is enormous at nearly $200 billion per year, and I fully expect some teams to knock it out of the park with a crypto game at some point. It might even be one of the sparks for a "cross-the-chasm" moment for crypto or AR/VR.
But it's the most overhyped sector of crypto right now.
Last year I said, "The play-to-earn gaming trend is here to stay. The amount of money these platforms have raised is nuts, and they're set up for a full cycle of iteration and development regardless of whether the sector's frenzy subsides next year."
Though I did expect a pretty rapid cooling of Axie's growth, I thought some of the big checks going into crypto games last year were based on some substantive development efforts. If they were, I haven't seen them. It's just as likely that they were part of the foie gras-ing of the broader crypto entrepreneurial class.
I'm bearish, and will only update my thesis when I see some signs of life or startup valuations that don't pull forward ten years of credit for the entrepreneurs building in this sector. It sort of reminds me of early social networks. I'll invest as much as possible in the first actually fun franchise at a higher valuation and pay for the derisked assets then. I don't feel the need to be early or tinker here. I'm happy to write about them next year if I actually play anything fun.

(Here's a good thread and Galaxy report. Great graphic on the gaming sector pp 31-40, too.)

The Cryptoverse

Using an Oculus for the first time and building that blue virtual mesh cage around myself was one of the coolest tech experiences I've ever had.
I also haven't used the Oculus in over a year.
Maybe it's lockdown fatigue, and I just prefer to see people in person when I can. When I can't, Zoom is still better than floating metaverse avatars. It still feels like we're years away from hardware that makes the cryptoverse not suck. The exception might be for gamers (I'm not one), teledildonics (no judgment, but again, not my jam), and gambling (if you're in crypto, who needs blackjack?).
We'll continue to cover the crypto metaverse, but I'm only likely to buy land in the metaverse if a distressed Decentraland plot comes on to the market. Otherwise, one of the only apps I've been intrigued by is Decentral Games, a DAO with crypto metaverse poker rooms with a real business model that does $2 million / year in fees.
Though with less disposable income, cryptoverse players aren't biting as often...that $2 million fee figure is down from a $36 million pace in Q1. Ouch.

I remain bullish on our AR/VR future, but I'm not about to bet the farm on it happening any time soon. The people who have done that have gotten their faces blown off this year.

Decentralized Social (DeSo)

An irony of Meta's collapse this year seems to be that the other wing of Mark Zuckerberg's empire (ya know, the one that makes money) is one of the areas most primed for disruption (and opportunity) thanks to crypto. I am excited for Decentralized Social (DeSoc) experiments to take root in 2023.
DeSoc rearchitects Web2 social media on crypto rails. It has the potential to serve as the underlying substrate that allows people and capital to organize and exchange value on the internet. DeSoc gives people on-chain reputation, organization (DAOs), and ownership of their content. It also introduces new ways for people to better capture the value of their online IP. DeSoc will change how value flows on the internet.
Currently, the DeSoc sector is split into three layers: front ends, the social graph layer, and content storage. Front ends like Lenster resemble traditional social media apps, but they read and write data   to the social graph layer (Lens, Farcaster) and the content storage layer (IPFS, Arweave) instead of their own app. The core piece of DeSoc, though, is the social graph layer.

Unlike traditional social, a DeSoc social graph is entirely open and composable.
Any developer can create an app on top of a DeSoc social graph and tinker with applications and monetization models. Users can use these apps and actually visualize their social graphs due to the openness of the data.
Imagine being able to chart which friends are triggering you online. Instead of blocking them, you could turn amplification down 90% until they lighten up. Having a bad day? Dial up the feel-good meme impressions. Feed me substantive business-related posts from 9-5, and witty banter and shitposts in the evening. DeSoc won't feed the algorithm and exploit your attention. It will help you conquer the beast and take care of your chicken and your mentals.

Most value in DeSoc will likely accrue to the social graph and content layers. Once a user's connections and content are stored in the social graph layer, they can use them in any front end. The runaway network effects of Web2 social media won't be locked into a single (censorable) front end like we see today.
A core infrastructure challenge facing DeSoc is the scalability of base layer networks like Polygon and Etheruem. Polygon's peak daily active addresses is 1.5 million, orders of magnitude more constrained than even mid-tier social networks like Twitch that have over 140 million MAUs. To achieve broad adoption, DeSoc protocols will need to operate on the cutting edge of scaling solutions like rollups, parallelization, and data availability.
Investment in this sector is good for the broader ecosystem's scalability challenges. DeSoc might also ultimately serve as the communications backbone for many DAOs and guilds.
If you think I am too negative on gaming, you might think I'm too blindly optimistic on DeSoc.
That's because a) I think it's easier to build DeSoc communications networks iteratively (this has been true historically when you look at social media vs. game studios' capital expenditures), and b) more of the value in gaming lies in the front-ends (best built by centralized visionaries) vs. the back-end (easier to decentralize via open-source software). The Web2 social industry happens to be nearly identical in size to the gaming space, $220 billion in annual ad revenue alone.
If you include digital entertainment and content (direct consumer purchases and subscriptions) in the DeSoc addressable market, you get closer to $1 trillion in annual revenue.
Transparency, user control over the algorithms, portability, revenue shares, 8 billion users.
Like I said. BOOLISH.
(Required Reading: The Open Social Mapand more on Lens, Farcaster, and Ceramic)

ENS & Identity

Vitalik likes decentralized identity and so do I! But I don't think proof-of-humanity necessarily requires you to submit to a global Very Good Citizen Registry™ or look into an orb and sell your retina data or genomic sequence.
Let's start with the good and the bad of crypto naming services.
The Ethereum Name Service ("ENS"), Lens, and Unstoppable Domains allow you to add human-readable NFT wrappers around your crypto wallets, much like domain registrars offer human-readable website domains that point to different IP addresses. You can send ERC tokens (fungible or non-fungible) to ryanselkis.eth instead of a long crypto address. Though, you'd be incorrectly assuming I own that ENS domain. (I thought I had renewed it for three years, only to realize a squatter snapped it up over the summer. F word.) You can then use these persistent identifiers like a twitter handle or email.
It's much easier to remember, and it's big business. It turns out people want to use crypto more often when blockspace fees aren't astronomically high.
DeSoc and gaming have proven to be good tailwinds for namespace adoption, and ENS has created a helluva schelling point for crypto users by securing the ".eth" namespace. Other competitors and breakout apps may choose to develop their own native naming protocols, but odds are good we get .eth, .crypto, .lens, and a handful of others as the most popular identifiers, much like we have .com, .org, and .io as the most popular domains today.

I think it's problematic that ENS and other wallet identifiers are globally searchable by default, along with every single trade, payment, vote, and other activity you've done on chain. That feels quite a bit more invasive than Venmo, and I'm not sure the solution to censorship and financial surveillance is radical and unprecedented transparency rather than privacy and selective data permissioning. Could crypto allow us to do identity the old-fashioned way, and show the world a slightly different person depending on the context of a situation?
You are a different person at work vs. at home, and with adults vs. with your kids. There is no reason everyone on the internet needs to know everything about you.
There's a number of teams working on this already.
Maybe the most interesting one you haven't heard of is Urbit, an enigmatic project that aims to create  a sort of open-source and privacy-focused version of WeChat; a digital homestead. Urbit is designed to give users digital autonomy, privately store their own data and digital assets, and communicate via direct, encrypted messages with peers without any third-party interference.
GDPR pop-ups are the internet's floating garbage patch. Crypto-powered identity is the future.
(Required Reading: ENS: Evaluating Growth Opportunities)

Creator Royalties vs. Monopolist Taxes

Apple charges a 30% tax on NFTs in its app store in order to enrich the company and its shareholders in exchange for building a world class mobile operating system. Likewise, OpenSea made the decision to enforce on-chain NFT royalties, reaping $1 billion in exploitative gains for…its creators?
I've been a casual observer over the furor that's risen up over OpenSea's royalty change last month. But, to be honest, it didn't really register for me given the broader market backdrop and my interpretation of (much of) the anger as coordinated outrage amongst competitors that haven't quite been able to nip at OpenSea's heels the way they wanted to. (More on that next section.)
I will grant you that royalty enforcement sucks at the user level. If you're wrecked on an NFT "investment," you want to recoup as much as you possibly can. If you are looking to scoop up new NFTs on the cheap, you want your marketplace aggregators to have a full selection of assets, and you'll will be pissed off if royalty enforcement reduces your access to certain collections or breaks your filters.
But I happen to think the right default in the long-term is to side with the creators. If creators want to opt out of royalties given the current market conditions as a way to help their fans get through the bear market, that makes sense. But I also think it's a sign of how baby sh*t soft some NFT punters are. You buy a picture of a second-tier cartoon PFP that had embedded royalties baked into all transfers at the mint, a work of digital art that just "really spoke to you," and now that the market has turned, what, f*ck the artists?
Nothing screams "Web3 is the creator economy" quite like Brad from SoHo demanding that we disable the valuable, differentiating function we were hyping as a game changer for artists less than twelve months ago, simply because his paper net worth evaporated from bad bets.
If this take upsets you, Brad, I'm sorry (that you feel that way).
Let's keep our eyes on the prize. Elon Musk seemed like he was ready to go nuclear on Apple for the app store VIG, but relented. Maybe the Coinbase team will hold the line? The tide may be turning for America's favorite monopoly, if only slightly.
Apps shouldn't need to give up 30% to Apple, but creators also shouldn't have to abandon

agreed-upon income streams because you had a bad year. More on the need to route around the app stores in the next chapter.
(Bonus: DCinvestor had a good rebuttal and mini-thread to my critique re mandatory royalties.)
(Required Reading: The Decline of NFT Creator Royalties)

OpenSea's Down Bad Yarrrr.

Last year, OpenSea was coming off one of the fastest revenue ramps of any business in history, processing nearly $5 billion in NFF sales last December alone. They went from a seed-stage startup to a $13 billion business in under two years, and I said at the time that I thought they could eventually be a $100 billion company (or network) if they continued to execute. That was easy to say at the time! But I also still believe it.
Yes, there are competitors. Of course they are after a bull market like we saw last year.
But Coinbase's entrance flopped. FTX is now gone, as are the NFTs it hosted on its platform. There is no second best, and the competition seems to be focused on marketplace aggregation or unsustainable, near-zero fee models.
Uniswap's "Genie," Paradigm-backed Blur, and even marketplace competitor Rarible launched aggregators in Q4. OpenSea bought Gem, its leading aggregator competitor in Q2.

I'm skeptical if vampire attacks from competing marketplaces will displace a market leader with a strong brand and sustainable revenue model in the middle of a bear market. Paradigm-backed Blur is the latest to throw its hat in the NFT marketplace ring. Early data seems encouraging, but the same was true with previous competitors, LooksRare and X2Y2, who captured early volumes through incentives and discounts but now languish at sub-10% market share.

And that's not even the full story. Most of the alternate marketplace activity has been wash trading designed to spoof the appearance of liquidity and farm token rewards on near-zero-fee marketplaces. OpenSea's dominance becomes more clear when analyzing transaction and user trends in parallel to volumes.

Going back to the point I made earlier on bitcoin's history vs. digital art, it might be helpful to think of OpenSea in the context of Coinbase in the early days.
You young bucks don't remember every single upstart competitor to Coinbase and how adamantly they argued about how Coinbase's fees were unsustainably high.
It's not that the logic is flawed! Other things equal, a competitor that charges 90% less will ultimately begin to eat your lunch. But that's true in sectors that don't routinely have 90% dislocations in market structure. Coinbase's competitors never had a bear market buffer. Coinbase made money on the way up in each market cycle, and its competitors picked up (temporary) market share. But not enough to ever really make a dent. And when the tide went out, Coinbase had its retained earnings, and its competitors had nothing.
They all got crushed or pivoted.
I believe the same thing will happen in NFT marketplaces. If OpenSea fumbles operationally, or stops innovating, or experiences extensive brand damage, it could lose its lead. But it's got the same SEO marketing benefits as Coinbase did, the brand advantage (X2Y2 is a terrible name), and, as discussed above, that whole creator royalties thing. If this chart has plateaued, it will shake out also-rans but keep OpenSea operating in the black until things rebound.

(Source: Nansen)
In full disclosure, I should remind you that I'm an early (lucky) OpenSea investor. But I invested, and will continue to hold, for three reasons: 1) the company's positioning reminds me of Coinbase for a slightly different asset class, 2) the design space for NFTs is significantly broader than it is in fungible currencies, 3) it seems probable that the feds come for NFTs only after everything else in DeFi, privacy, and DAOs.
Truth be told, I haven't spent much time on the NFT markets over the past couple years. At least not "investing" in (let's be real, "flipping") JPEGs themselves. It's because I'm bad at trading and have a finite attention span, but I do think NFTs will become a ubiquitous standard for wrapping financial assets in the same way they currently wrap monkey JPEGs.
(Required Reading: Exploring the Unbundling OpenSea Thesis, Comparative Analysis of NFT Marketplaces, Blur, Financialization, and Renewed Interest in NFTs)

OpenAI & ChatGPT

I'm still, like everyone else, trying to process just what the hell is going to happen next as OpenAI transforms art and information. My quick reactions are here, if for no other reason than I want to look back on these next year and see how well they've aged or how silly my initial overreaction will look.
  1. Prompt writers will inherit the earth. Who knew Twitter's character limits would keep me gainfully employed as a writer once my ability to write the Theses next year becomes a redundant skill! Seriously though, I wonder if this is going to upgrade or downgrade our cognitive abilities, or simply rewire our brains much like GPS and social media did? We're worse at directions and spatial awareness than we were pre-GPS, but better off. We're better at connectivity post-social media, but less mentally healthy. Which way will ChatGPT trend? And what does the AI think of itself?
  2. The lawyers always win. This is going to be an IP and copyright lawyer's bonanza. Well structured, unique, historical data will find fresh monetization schemes. I could also envision this being great for individual users and hell for corporate users who might be opening themselves up to copyright infringement or other IP disputes in leveraging this tech. ChatGPT passed the bar, so it's clearly in the pocket of Big Law!
  3. I agree completely with Ben Thompson's observations about ChatGPT in Stratechery. This will  be one of the biggest boons to productivity in the next decade. Today, it reminds me of early Wikipedia and early Google searches (both are still flawed, by the way). This might help students complete a term paper, but it won't do the work for them. If anything, this might free up good students to spend more time on diligence, assumptions, logical structure, and source review (vs wordsmithing), much like Excel allows you to run complex models, but those models are still mentally taxing to structure.
  4. It's a short hop from ChatGPT to ubiquitous private communications. It's literally horrifying to think that all tweets, texts, and emails could be mass uploaded to the AI, and investigators, authorities  (or enemies) could more easily query your own words for evidence of your state of mind and psychographic profile. The NSA probably already does this. But now it's hypothetically in everyone's hands? Bullish Signal. And any crypto protocol that makes it easier for you to own your own content and speech. I'm revisiting Ocean Protocol, and it looks like others are as well.
  5. I think the imminent demise of Google has been greatly exaggerated. I really liked this line from Parker: "Remember before OpenAI when Google had a chatbot so powerful some guy believed it was sentient and tried to free it? That was the stuff they didn't think was worth shipping." Scary! We must decentralize these data silos if we are to defeat Skynet. And hope that China stays isolationist and runner up in AI research.
  6. AI is simply a reflection of the data sets underpinning the large language models. The current data is pulled almost exclusively from the internet or intranets of tech companies. Content produced by AI is not representative of humanity or even a global hive mind, but it is heavily weighted by who is on the internet and, among that group, who dominates in producing content. AI isn't good or bad, but the data sets might be.
I could go on about this for hours. But so can every other Thought Leader in Tech™ so I'll spare you the word vomit. Go ask ChatGPT what it thinks about crypto.