jakiro
Author: Delphi Digital Lianchuang Tommy, Source: Author Twitter@Shashnessy119; Compilation: Deng Tong, Bitchain Vision Realm
When I tried Midjourney for the first time and later ChatGPT, their powerful ability initially made me scared.My ability to understand their abilities has aroused a slight crisis.This made me realize that although the current LLMS is only good at perfecting our sentences, they inevitably closer to the AGI that is fully developed, which can profoundly affect our thoughts.
The media’s portrayal of AGI usually values its real impact and regards it as an all -around application or IROBOT.AGI will penetrate into every aspect of your life, even if you are not realized.Imagine the AI2041, which is an outlook for the future. There, the AI insurance application of a family will evolve into a love choice based on social class and risk analysis of true love.This shows how deep and influential AI will become.
Each science fiction film depicts a future of artificial intelligence full of pessimistic emotions, because such a plot is more selling, and this is a realistic final.No matter how morality and ethics we think of the OPENAI board of directors, due to the nature of human existence, there are some prejudices that cannot have negative effects on all applications and use cases based on these basic models.You may like to communicate with ChatGPT, but if you are in court, your artificial intelligence jury uses facial recognition to see your skin color or the unique spelling of your name, and double the time to send you into prison. What will you think?These influences are disturbing.
Centralized artificial intelligence is inevitable.Once Google requests your license to access your GDRIVE, GDOCS, and Gmail, your personalized artificial intelligence will have your own life.I expect Apple to launch a personalized artificial intelligence for each device, because they are behind in the global artificial intelligence competition and need a product that uses their security brands.If the OpenAI fine -tuning model, this has a chain response to society and affects the millions of customized ChatGPT built on it. Do you feel comfortable?
We need an alternative.Cryptocurrencies are the perfect combination of artificial intelligence, because the transparent global human coordination that supports this movement can use artificial intelligence globally to achieve good results.Create and fine -tune the open source model of crowdfunding (cash or GPU), and anyone can review the prejudice or problems in real time. This is the safest way forward in the accelerated development of artificial intelligence world.
I believe that we are moving towards a world with billions of artificial intelligence models. Whether it is downloading and personalized everyone’s open source model, or the project and the company to build their own model set for specific cases (think of the Uniswap LP provided, the exchange risk of the exchangeAnalysts, artificial intelligence) analysts).
Encryption and artificial intelligence are a perfect combination. Its cornerstone is the audit, community ownership and community -oriented of our generation’s most powerful technology.Whether using everyone’s GPU to train the model and have ownership for them in the model, or DEFI and smart contracts use artificial intelligence to expand their capabilities in their use cases, or customize your personalized artificial intelligence (instead of your personalized artificial intelligence (instead ofThe artificial intelligence that must be summarized throughout the world like Bard), this combination is reasonable.
Decentralized artificial intelligence will transparently share the internal operations and ownership of our generation’s most powerful technology.The centralized artificial intelligence cannot provide this core value.
In the end, AGI will use cryptocurrencies because it trusts code and mathematics, rather than the sudden imagination of the branch branches and human nature.The artificial intelligence creation we evolved in the future will use cryptocurrencies, and we should do this.
Encrypted X AI theme and idea
Artificial intelligence values of crypto punch:
Ethereum core Cypherpunk values are applied to artificial intelligence from @Vitalikbuterin overview: not to de -platform, open global participation, resistance review, neutrality, cooperation, etc.It is ridiculous to rebuild artificial intelligence under the centralization of Web2.
AGI and noCurrency required currency
AGI’s utopia or anti -utopia is conducive to such an idea to interact with unauthorized money to realize its ideas.In the future, there will not be a Chase check account.It will further build decentralized artificial intelligence and cryptocurrencies with encrypted technology and is not controlled by the Fed or OpenAI board.
DEPIN and AI are a clear example
In the past ten years, all our research has been concentrated on how to improve the performance and efficiency of super large -scale data centers.In the next ten years, I expect technology to use potential GPUs and user hardware to train and reasonable AI models.The demand for H100 that NVIDIA can ship is obviously limited, while technology companies master the available products.Let our Mac Pro GPU and other hardware use the large -scale use of training and reasoning is an obvious case.Leaders include@t_official,@akashNet_ and @Gensynai.In the future, there may even be a path that NVIDIA turns instead instead of selling its H100, but simply establish its largest cluster.
Incentive new mode creation
To establish application -level creativity, we need to motivate the development of the new model itself.This includes incentive measures for training funds, specific training data for crowdsourcing, and hosting and reasoning models.Large language models are only a type of artificial intelligence model. Even so, there are dozens of leading models (Bard, Chatgpt, Claude, etc.).Users around the world can provide their GPUs, funds, or data to large -scale training and fine -tuning models, and have part of the final model.
More intelligent applications and intelligent contracts and artificial intelligence
Unhindered decentralized artificial intelligence will provide better applications.Quoting the intelligent contract of artificial intelligence models can expand the design space of the application and greatly enhance its logic and functions.Imagine that the UNISWAP liquidity supply is affected by the large -scale under -chain model of ZK to ensure that the model is not tampered with.Examples include @infrence_labs @gizatechxyz and @moduluslabs. Just look at @testmachine_ai. It provides a predator mode to review your encryption code in real time and learn from it for 6 months for expensive manual review.Or check the large -scale machine learning model for NFT through @upshothq.
Artificial intelligence makes cryptocurrencies easy to use
In the future, most cryptocurrency users will never see the various fine branches and endless letters and vocabulary we talk about in the field of cryptocurrencies.They only need to enter the intention to enter a low -level machine learning (LLM), and a solution network will process all the complex steps of their transactions.This LLM will learn, personalized and simplify your life.Few people need manual bridge assets, which is made by the solution to the solution.
Best model decision maker win
I believe we are moving towards a world with millions of artificial intelligence models. Whether everyone has their own personal models, or every project and company have their own models.We already have more than 490,000 open source models on Hugging Face, and have 3 million custom CHATGPTs on the OpenAI app store.When the mature artificial intelligence services of encrypted services became the mainstream, I think it can effectively choose the protocol to which models to use for each situation will be very valuable.
Just today,@nousreSearch has released a new Bittensor subnet program. The subnet can evaluate the open source model and use the logic of this rating next step to target the request to the correct model.
Morality, ethics, and legal issues limit centralized artificial intelligence
Central participants are constantly facing moral and ethical issues, and therefore damage the model.Imagine that a decentralized low -level machine learning (LLM) pays people to pay data by signing an agreement instead of a lawsuit against the New York Times.Compared with the centralized system, this limits the centralized development, and the open system can be released directly (for example @bittersor_).In the middle of the center of the center, when struggling with intellectual property -related lawsuits and the moral and ethical concerns of publishing more intelligent artificial intelligence, the encrypted network can start and deploy these networks without the need to win.
Decentralized artificial intelligence provides transparency
People will expect transparent training (“we build this model in the way you say”) and reasoning (“My request is not messy”).Centralized artificial intelligence cannot provide this core value.Even if ordinary people are difficult to review the model, similar to cryptocurrencies, our idea is that you can sign a contract with someone or artificial intelligence who can review the model.
Real -time insights in the future
I think people want to understand the future of artificial intelligence in real time, not just to update when Openai wants to share.This is only possible to achieve through transparent and decentralized systems.Do you really want to figure out when AGI is found?
Encrypted X AI visualization
We need more platforms to provide the visualization of what happened behind the artificial intelligence model that encrypted projects is using.When you use Bittersor’s subnet 1 for text generation, how can you determine that it is not just running your prompts through Bard or ChatGPT?I don’t say this is a bad thing, but I don’t know the answer.
Tokens in the artificial intelligence world
The use of tokens to promote the ownership and coordination of artificial intelligence projects will be interesting.At present, tokens have attracted supplier users (and speculators), but developers of centralized companies such as OpenAI have locked the demand party and can contact a large number of users.See if the encrypted project can effectively guide the needs of the demander to surpass the tokens of the supplier, this will be a very interesting thing.
Attract real artificial intelligence talents
The Crypto X AI project must attract real AI talents from Web2.In view of the general performance of cryptocurrencies in the minds of web2 builders, this is an obstacle.It will have significant advantages in projects that can attract real artificial intelligence talents from Web2.I just think that learning cryptocurrencies is easier to learn how to build basic artificial intelligence models.
Delphi Ventures Crypto X AI Investment Portrait Company
Delphi is very active on the binding point of Crypto X AI.We are honored to support leading projects in this field.
-@ionet_Offical: GPU cluster with large -scale heterogeneous hardware.
-@inference_labs: Allow DEFI and smart contracts to use the linked model through ZK.
– @0G_Labs @mheinrich: Data availability of artificial intelligence on the chain.
-@Upshothq: Artificial intelligence network for the next generation of decentralized applications.
-@testmachine_ai: The proprietary algorithm driven by artificial intelligence -driven for auditing smart contracts
-@taofuxyz: mobile pledged token of Bittensor, etc.
-@altStatemachine: You can have a unique Metaverse AI that can have, train and trades.
-@GepPettoai: artificial intelligence game and video creation.
-@Stabilityai: open source tools for artificial intelligence, through @SeedClubVC (H/T Nima and Anthony).
-@Mythosventures: The leading early artificial intelligence venture capital fund.
-@Mypeachai: NSFW companion (angel).
