DEEP Funding brainstorming

Hi all – as indicated in the SNet Phase Two overview we will be launching later this year a "Decentralized Ecosystem Projects funding pool ( DEEP Funding pool) via which the community will vote to allocate tokens among community-proposed projects.

This will likely start 1-2 months after the initial issuance of AGI-ADA tokens (not the real ticker symbol I guess), we are already looking into the logistics. Initial AGI-ADA tokens we are aiming for April or maybe May though there are some technical dependencies we’re working out.

But I thought I’d start a topic here to nudge the community to share ideas regarding what may be interesting DEEP Funding projects. Folks can respond to this thread directly or start new Topics within this same Projects category. I’m interested not only in people discussing their own projects but also in folks suggesting projects they’d like others to pursue. Since the allocation of DEEP Funding tokens is going to be community-voting-based, getting community feedback here on preliminary ideas would seem to be directly valuable…

While I haven’t used this Forum much so far, for this sort of purpose it seems a far better choice than say Telegram because ideas proposed on Telegram tend to get lost easily in the endless scroll…

Thx everyone for your participation!

19 Likes

Hello,

I think that as a source of inspiration it may be beneficial to have a look at what the big cloud platforms are currently providing, either to imitate or stand out against them: What ML and AI services are offered on major cloud platforms? How are they categorized?

Here is a list of what seem to be the best overviews (after a short walk through search engine results):

  1. Amazon Web Services (aws): Machine Learning, AI services
5 Likes
  1. Google Cloud: AI building blocks
  2. Microsoft Azure: Cognitive Services
3 Likes
  1. IBM Cloud: AI solutions

And here is an article providing a similar overview: Best MLaaS platforms (2020)

(split into multiple replies due to hyperlink limits for new users)

3 Likes

How about a SingularityNET version of the OpenAI DOTA bot - decentralised so that it can train against real players as well as itself.

Then the bots can share data with each other over the blockchain, and update themselves.

One day it can challenge the OpenAI bot to a battle!

This implementation of AI already has a lot of eyes on it, and it would be a fun thing for the community who cant write code, to take part in, just by playing against the bot!

  • Proof of concept of SingNET interacting with games.
  • Proof of concept for an AI that lives within blockchain, learning from multiple sources
    simultaneously.
  • Health, Longevity, Finance, are all great areas to work on, but maybe something a little more “fun” could add more dimensions to the project.
  • Earn some kind of reward for matches played.
  • Draw the attention of competitive gamers that have followed AI development due to the OpenAI version and bring them into the Decentralised AI space.
  • Attract developers to do something fun!
11 Likes

Ben,

First things first. I love that you look like a stoned old hippy and you are head of one of the most exciting projects in the crypto space!

Thank you

5 Likes

i was think of something easy simple and useful like a bot that can do TAX scan your info like id and you tax forms T4 this is very random just want to put it out there.

2 Likes

How about funding this?:

1 Like

DAOs on Cardano sound great but I’m not clear what’s the AI/SNet angle? I suppose the reputation system may be one good potential. But it would seem maybe “DAO on Cardano” is better for the Cardano Voltaire/Catalyst programme? … whereas “reputation system for DAO governance” is core SNet / DEEP Funding stuff…?

Am def. an aging hippy but most of the time I am drugged only by black coffee these days :wink: … I don’t need psychedelics much anymore, my brain micro and macrodoses itself automatically ;D

2 Likes

:+1:
Keep on keeping on!

I completely agree and would suggest undertaking a detailed analysis of competitor services, if not already done. At first glance, the centralised services offered by Amazon, Google, Microsoft & IBM all appear pretty strong and SNet’s offering will need to compete with these established players.

There will be gaps in the market which SNET can target and there may be specific applications or algorithms more suited to a decentralised approach.

1 Like

A decentralized personal CDP (Customer Data Platform)

Summary:
This is quite a big (too big?) and ambitious endeavor combining features from different Blockchains, including Cardano’s Prism, Ocean protocol and SNET. In short, the idea is to collect, store and analyse my behavior and :

  • Let me be in charge on which sites, social media etc can use this data for what purposes, and against what fee
  • Create a personal assistant the recommends interesting content or products or other tips to me.

I jotted down my ideas in this One-pager : https://docs.google.com/document/d/1kxjXkzTMwOVtiqIN02Z9tJ8SajgzEu4xG5W7QQLTdFs/edit?usp=sharing
(Thought it was a bit too big for this space)

1 Like

I would like to see someone make use of the recent breakthroughs in Poker playing and create a poker playing alg that mimics Pluribus or ReBeL which takes screenshots as inputs and outputs action suggestions for the most popular poker websites. Some AGI cost associated with every suggestion ofc.

I don’t actually play poker myself, and though I might be interested in considering it if I could have suggestions from a superhuman level player, the reasons I would like to see this happen are aligned with singularity net utilization. As a part of this, it could be big news in the poker field, exposing many people to singularity net, and attracting attention to the projects.

Agricultural General Intelligence

From bunker-like seedbanks in remote lands to Bill Gates being the number one farmland owner in United States, it is pretty clear how food production will be a key issue in the future. Climate change and an increasing world population are only two of the most important factors in the game.

Can we decentralize food production?

Some AI assistant (or even better, an AI platform) that allows every person in the world to have access to data and algorithms that help grow their own food in a sustainable and efficient way. Services that can help in preventing land degradation and desertification, manage resources like water in an efficient way, diagnose plagues and diseases in the vegetables, knowing the symbiosis between plants, trees, fungi, animals, humans… Helping people with info that allow to design and build their own kind of “permaculture garden” or “food forest” even in one square meter. And the algorithms will be feed with the info that people will provide in an app-like UI. Data about temperature, water, soil quality, the size of the fruits they are producing or even the taste and much more variables.

This way, people from Africa can teach the rest of the world different techniques to prevent desertification and, in the other hand, people from the north countries can teach how to grow food on cold weather or some of the techniques learned with the experience of years of mass production, just to give a couple of examples.

Everyone able to have access to every knowledge about food production just in his smartphone. Let’s take AI out of smart cities and let’s make smart our food, our farms, our lands and our countryside!

I have been running this idea in my head for years and I have enough basic knowledge in AI to know that’s really possible to do in an empowering and democratic way. Just I don’t have the resources to do it by myself, but I would be very happy to live in a world where this kind of things exists. The ultimate revolution, the ultimate freedom for people all around the world.

Then, one step beyond would be decentralize the food distribution, where the small food producers can sell or exchange their products with the people around according with demand, no need of supermarkets and of course no need to bring food from the other side of the planet.

Decentralized food production (every people growing food even in the rooftop) + Decentralized food distribution (people distributing to each other the food they produce)

The AI + Blockchain revolution will grow on everything!

9 Likes

Due to my history with livestreaming I tend to think of a few real-time applications:

  1. Streaming voice to text. Pretty basic caption generation. Although can be a hard problem when the topics being audibly discussed get more niche and domain specific.

  2. Customizable synthetic voice. Would be cool to use audio clips to “shape” a synthetic voice that could then be used, perhaps via text to synthetic voice, and real voice -> synthetic voice.

  3. Facial recognition from webcam. Again this can be used with streaming services, there is actually now a trend to use anime characters (virtual youtubers, etc) with facial recognition tech to convey emotion, and to always maintain a high degree of aesthetic beauty without worrying about the underlying human :slight_smile:

  4. This one isn’t so much about live streaming, but I have seen plenty of apps where you drop in your face and it creates a painting of it, or a cartoon version of it, these tend to spread like wildfire because it’s easy to use and fun to see the results. One that recently got popular was toonme.com, but there are others. These types of applications are fun and widely appealing, although not critically important.

  5. I think an AI artist in general could be interesting, ESPECIALLY if you can feed it images and descriptions, see the result, and then critique it and give it suggestions, continually iterating. It’s a pretty massive project but I think something that AI can really do to empower humans is to allow the human to tap into their inner critic, and use that as a command center while the AI does the heavy lifting with wizard artistry.

Just some fun ideas.

4 Likes

First Goal.

  1. Increase the value of the AGI coin by no longer discouraging the trading of the AGI token.
    The more the coin is traded and deemed an investment vehicle the more leverage each vote will have when voting / buying Deep Funding projects. The worker / contracted individuals will accept AGI with the purpose of either holding for future value or selling for other coins or fiat. Lets not pretend people work their best when working for free. AGI is deemed an investment vehicle or no one would hold it.
  2. Financial AI Lending Project or similar: Immediately Develop an automated AI driven Lending platform owned by AGI shareholders and trading and lending AGI coins similar to AAVE. The video games are great but they will not bring global mass adoption and value to the coin. If Singularity net runs out of money before Artificial General Intelligence is realized then it was just a pipe dream. Go where the money is first, create immediate global need and value, then have fun creating chess games.
  3. And possibly most important: More vested interest brand ambassadors from top AI schools or organizations other than Hanson and Dr. Ben. We should have no less than 10 PHD level Active AGI Brand researches actively promoting their AGI. AI experts from Medicine, Robotics, Finance, Engineering, Automotive, Aero, Govt, Marine, Travel, Entertainment etc… 1 or 2 experts whom are compensated with locked AGI for a set period of min 3 to 5 years once they have achieved specific milestones contributed to AGI. This should be a substantial investment in the 6 figure equivalent (USD current value for each). This will take the pressure off of Ben being the only spokesperson and spread our AGI and value worldwide instantly and eliminate the fear of Ben being the only one at the helm. Again the more AGI value, the higher the AGI price, the more incentives we can offer towards projects.
1 Like

Hi,

I just wanted to say that I think this project is awesome - the team should be very proud of itself for getting SNET this far. It’s great to see a blockchain project with real world application and this DEEP funding is an excellent initiative to accelerate adoption.

I must admit my AI knowledge is a little rusty and I’m still getting my head around SNET, but will make a couple of suggestions:

Firstly, on the overall strategy I would suggest funding a mix of projects ranging from simple “nuts and bolts” algorithms to larger projects, with emphasis on the former. I’d also keep away from projects with unrealistic goals, e.g. “predict the stock market” and General Intelligence.

Secondly, a project idea was to use AI to analyse one’s publicly available data (e.g. social media interactions and google location tracking) to generate suggestions to improve one’s physical and emotional health. For example, the code could advise when a person’s lifestyle has become too unhealthy or if there is increased risk of burn-out due to overworking. It’s arguably too “big brother” but I’m sure there’s similar interest in monitoring employee wellbeing.

Really interesting to read about others’ project ideas.

Ben, you probably need to read the whole post? - it looks like you only read the bit that is visible in the link above?

In companion with the CDP above, You could also create and E-commerce recommendation engine.
There are already enough on the market, but they can be costly. It is in any case highly AI driven. The engine usually consumes order data, clickdata and profile data. This can be collected by the engine, or (sometimes) the engine can be fed with historical data. The more data you have, the better the recommendations. Except perhaps price, other differentiating feaures could be:

  • Mixing ‘public data’ as mentioned above with personal, private data from the end-customer, without exposing this data to the ecommerce shop
  • Crowdsource the best algorithms. and enable the customer (the ecommerce shop) to switch algorithms
  • Creating leverage by connecting many smaller shops, instead of going for the few big ones. E.g. by connecting with open source shops.
  • Perhaps create an opt in for the customer to receive the best personal recommendations as a push message from all connected stores. e.g to 5 per day, top 10 per week, specific categories, etc.
  • Some user based mechanism that improves the results by giving feedback
1 Like