Crypto and AI: Is something happening there?

As a rule, every tech trend can also be found in the crypto space, sometimes as a parody, sometimes as a tragedy, and sometimes as a useful application. Artificial intelligence is considered the major trend of 2023. How is the crypto economy reacting? Are there any promising investments?

The tech trend of 2023 is almost certainly artificial intelligence (AI), at least since OpenAI with ChatGPT has delighted almost everyone who is interested in technology. This also has an impact on the crypto market.

One could construct thousands of plots of how cryptocurrencies and AIs go together. For example, an AI could own Bitcoins, accept them for services or use them to buy domains. Crypto could turn AIs into autonomous entrepreneurs. But we want to theorize here, we want to experience what really happens.

Which cryptocurrencies, tokens and blockchains do something with AIs? How serious do they appear, what plans are they pursuing?

The top 13

We find a first indication on ranking sites such as CryptoSlate or Coinmarketcap, which have set up the “AI and Big Data” category for coins. From these lists I sorted out the coins like the Graph (GRT) only work with data or like Phala Network offer cloud computing.

The following list shows the remaining top 13:

That would be a start for our inventory. In the next step, we will briefly present the 13 coins before we dare to draw a preliminary conclusion.

In the brief introduction, we summarize what the coin says it does. It will rain bullshit, I promise, but also inspire. As a first assessment of whether it is a “dead coin”, we also look at when there were the last signs of life in social media and on the blog.

The market capitalization in euros is in brackets after the name.

Brief portrait of 13 AI tokens

SingularityNET (218 million)

SingulartyNET is the “first decentralized AI marketplace” on a blockchain. Instead of a central provider as with OpenAI, there should be many servers and providers of AI services. Coindesk magazine described SingularityNET as “one of relatively unserile projects to emerge from the 2017 ICO hype.” different instances – speech recognition, navigators, image generators – work together decentrally. He hopes to use it to decentralize AI development “in the same way that cryptocurrencies decentralize finance.”

14 hours ago SingularityNET tweeted for the last time, the last news was published on January 25th.

Fetch.ai (236M)

It’s quite difficult to understand what Fetch.ai really does. It is an “interchain protocol” in the Cosmos universe that allows “peer-to-peer applications with automation and artificial intelligence” to operate. The Fetch platform “lets anyone build and activate AI services.” Fetch is being developed in the university town of Cambridge, and there’s news of collaborations with the IOTA Foundation and Bosch, as well as first Fetch apps like commodity derivatives marketplaces and micro-mobility.

Last tweet: 19 hours ago, last news: January 27th.

Alethea.ai (118M)

Artificial Liquid Intelligence, Alethea AI for short, is a “research and development studio” that deals with generative AI and blockchains. From what I understand, they are porting CharacterGPT, a “text-to-character” AI, onto the blockchain. So you can use AI to create avatars for the Metaverse, presumably as NFT. Alethea works with the blockchain Polyhon for this. A cooperation with Hansen Robotics from Hong Kong is also intended to tokenize the popular humanoid robot Sophia through an NFT series.

Last tweet 12 hours ago, last news on January 19th

Numerai (94 million)

Numerai has been around for quite a long time, I wrote about it back in 2017 . It’s a kind of AI hedge fund that uses machine learning and DAO-style tournaments to try to make good trades through artificial intelligence. So far it doesn’t seem to be a resounding success, otherwise Numerai would probably be worth more.

The last tweet is January 30th, but the last news was March 22nd, 2022 a while ago.

Cortex (42 million)

Cortex is “the first decentralized world computer capable of running AIs and AI-based apps.” The “GVM” is an EVM-compatible language for smart contracts capable of “AI inference” to process, for which it uses the GPU instead of the GPU. So you can “execute non-trivial AI models” on the Cortex.Chain. Ambitious! The blockchain is already up and running, and the first AI dApps appear. But not much has really happened yet.

Last tweet January 29, last news 22 November 22

Vector Space Biosciences (25 million)

Vector Space Biosciences is more about data processing than AIs. But it sounds interesting: Together with a scientific laboratory, Vector uses “advanced language models and a network of specific data engineering pipelines” to make scientific discoveries “that protect and repair the human body during a space flight.” Somehow you have to go to Mars come. Special. But the role of the token is somewhat puzzling.

The last tweet is from January 19th, but the last news from August 2022.

0xmons (25M)

0xmons is explained more simply: They are animated NFT collectible figures that were created by a generative AI. They are pixel style and animated from Pokemon, SCP and Lovecraft. It seems to have a market value – the floor price on OpenSea is 12 ether and it has turned over 3,000 ether. On the other hand, the value of the 0xmons utility tokens seems to be low – perhaps because nobody knows what they are really needed for.

The last tweet is from January 31st, the last news from March 29th, 2021.

Vaiot (20M)

VAIOT offers “a portfolio” of AI assistants on a blockchain that provide automated services and transactions. Including a paralegal, a broker and a sales assistant. However, concrete details remain disturbingly nebulous. Even after looking at the white paper, I don’t know how concretely Vaiot connects blockchain and AI.

Last tweet 19 hrs ago, last news on Jan 24th.

SingularityDAO (17M)

The aforementioned SingularityNET, the highest endowed AI coin, also has its own DAO. She forms AI-generated risk management strategies for DeFI, for which she presumably uses SingularityNET. I suspect this DAO is still in its infancy, but I also suspect that the DAO is not directly related to the actual Singularity platform, but uses their brand to collect money for trading.

Last tweet is 18 hours old, last news from December 27th

GNY.io (12M)

GNY develops machine-learning based forecasting tool for crypto trading.

Last tweet January 31st, last news October 19th

DeepBrainChain (12 million)

DeepBrain Chain bills itself as a “decentralized neural network.” However, the platform seems to function more as a decentralized cloud computer, with participating nodes providing computing power to AI companies in exchange for tokens. The blockchain is live and, according to Blockexplorer, is already processing the first operations. However, I cannot judge to what extent cloud computing or something else that makes sense is behind it.

The Reddit is in Chinese, the last blog post is on February 28, 2022, the last tweet is today.

Oraichain (9M)

Orichain is a “fundamental layer” for “a new generation of smart contracts and dApps”. Like Fetch.ai, Oraichain is based on the Cosmos architecture. It promises to verify the execution of AI smart contracts through a consensus proof of correctness, which sounds pretty bold. However, there is already an AI marketplace where at least two providers have been selling AI services.

The last tweet is January 31st, I didn’t find any news.

Matrix AI (5.4M)

The matrix network evolves in three stages. The first is an “AI-optimized blockchain platform”, the second a “blockchain-based AI economy”. Form a “self-evolving AI platform”. The third tier will be a “general purpose platform for avatar intelligence”: each user can “upload and save their brainwaves to create their individual avatar. freaky.

The last tweet was from January 31st, the last Reddit post was 3 days ago.

Some theses on AI tokens in cryptospace

How is all of this to be assessed? I formulate a few theses on the overall impression. These theses are based solely on this first impression, which I have obtained. They are neither fully thought through nor thoroughly researched.

1. The market remains cool. There is not (yet) a flood of AI tokens and their clones, and it is apparently not (yet) enough to write “AI” on your token to make a profit. Instead, we have a relatively moderately rated group of AI projects, almost all of which have their own approach. Only three achieve a market cap of more than $100 million (SingularityNET, Fetch.ai, Alathea). Oraichain and Matrix AI Network even have a market cap of less than 10 million, VAIOT, SingularityDAO and GNY 10-20 million. No token makes it into the top 100 in the overall ranking.

2. However, the market does not distinguish between AI, big data and cloud computing. He pulls up indiscriminately in the wake of a small AI boom, which falls into this block explorer category. Even with typical articles like “5 AI Coins You Could Invest In” there is no disconnect. Possibly because of the assumption that a decentralized AI needs data and computing power from somewhere, but probably more because the market is still hardly sorted.

3. Most of the projects are relatively young. The oldest token, Numerai, is from 2017, Agix and DeepBrainChain from 2018. The youngest are Alethea (2022) and Xmon and SingularityDAO (2021). Aside from Alethea, all AI coins predate Dall-E and ChatGPT — and are therefore not products of this hype.

4. There is little evidence of “dead projects” . Each project has an active account on Twitter, with most having last posted to the blog in January or December. Of course, this says little about the progress of development or the status of the project, but it serves as a kind of first sign of life and minimal protection against a rug pull.

5. Many throw around buzzwords and phrases. An algorithm is beefed up into machine learning and that jazzed up into AI and so on. Branding is everything in the crypto market. This could apply to Numerai, a rather moderately successful decentralized trading fund, as well as to GNY, which develops trading via machine learning, or Vector Space Biosciences, which finances scientific cloud computing with tokens.

6. Other projects are centralized providers who sell their service on a blockchain or for other blockchain applications. That’s what I think of Alethea, who offer CharacterGPT as a service, Vaiot, who rent out AI assistants for blockchains, and maybe also GNY, who are developing a trading bot using machine learning and can be paid for by tokens.

7. Some tokens promise something that seems unrealistic: When Orichain or Cortex say they want to bring AIs into smart contracts, and thus onto a blockchain, it seems implausible, since AIs consume far too much data and processing power to run so redundantly as a blockchain does by definition. Fetch.ai is also going in that direction, unless they’re actually porting a central server, and Matrix AI sounds way too far-fetched to be realistic anyway. Either the blockchain is centralized – or the AI ​​works with rudimentary functions. But who knows: maybe interesting, semi-decentralized blockchains will emerge that are connected to a blockchain as sidechains or rollup?

8. The most promising projects try to decentralize AI services. Instead of powerful companies like Microsoft or Google owning the AIs and renting them out via API, in part and with a handbrake, AIs should be operated decentrally. A DAO-like construction such as SingularityNET and the associated DAO could provide the business model for this, without which the infrastructure for a serious AI can hardly be financed. Fetch.ai, Orichain or Cortex could also go in this direction with the appropriate configuration.

9. Blockchain projects that sell computing power and data could participate in the AI ​​boom. Since AIs are very hungry for computing power and data storage, blockchains or DAOs that sell computing power to AI providers can be worthwhile. Why not if the computer or server is already running anyway? That could make coins like DeepBrainChain attractive.

All of this is not to be construed as a recommendation for anything. At best as inspiration to deal with individual projects in more detail. I myself have no idea which of the coins presented here is a serious project and which is fraud. And even if there is an actual project, the first thing to check is the tokenomics – even a real project can issue tokens in a way that is cheating.

Irrespective of this, the emerging market for AI tokens looks interesting and in parts also promising. I plan to present some of the coins in more detail in the near future.