If you had to launch a startup today and you could easily find all the resources you need, what niche would you go in? What is the next blue ocean?
AI right now is like the beginning of the internet, blockchain, the industrial revolution, the printing press, and electricity. So, in my opinion, the integration of AI with every existing technology creates a new blue ocean of possibilities, paving the way for unprecedented advancements across industries.
That’s why I’m waiting for the ultimate programmer sub to know what exactly I want to create.
3 Likes
Yeah AI is literally in every tool right now.
Image creation and text creation with AI is a thing right now.
Video is still in its early shoes.
1 Like
it’s probably super easy right now to get funding for anything AI related if you have the right background.
so simple answer is anything AI related
2 Likes
That would depend on how much risk you were willing to accept and mitigate.
In my opinion there is no one blue ocean or next big thing. You have to look at solution-based thinking. I’ll give you an idea for one thing I’d been mulling over but will probably never take any action on.
Elon Musk recently spoke in his Tucker Carlson interview about his previous work in AI and his disillusionment about where the products he had helped work on were going, particularly how the LLMs had a clear liberal bias. He suggested he had been wanting to work on his own AI idea called Truth GPT.
Now, in going down that path while your competitors are trying to regulate the sh-- out of anything able to be classified as AI presents some rather formidable legal hurdles. Products already in existence will have their own intellectual property rights assigned and its fairly safe to say the definition of AI will have to do with existing concepts in the field such as LLMs, traditional neural networks and so on.
What if you could create something which not only takes data input from the internet or humans, as is traditional nowadays, but adjusted traditional algorithms for gradient descent or weighting the networks to harmonize with ambient input from the local environment? Or similarly, have the activation function be a function of that ambient energy in the environment through sensitive magnetic sensors for example to give it a degree of creativity in its learning process. It could be built onto a chip and the implementation for the chip subject to intellectual property and copyright safeguards.
I think ideas like that to compete with traditional models of AI that just simply do what any nefarious human might tell them to do would be a great idea, especially if a proof of concept of an ethical machine which is apolitical and yet creative could be demonstrated.
EDIT to add: I suppose I should explain my reasoning here. It should be possible with a sufficient number of sensors to nudge inputs in a neural network in just the right way that you can tune to give a machine the ability to accurately sense the external world and compare the data it is being fed to reality. And also adding an analog or natural component to a chip through this method would give you the opportunity to no longer call it an artificial intelligence, but one with a natural component to it, thus opening a loophole.
The idea with being an entrepreneur is that you look at problems that exist or assumptions we make that might be flawed and see how you could monetise the solution or disrupt those assumptions in a way that aligns with human values. And there are no end of problems in our current world we could work on, but any of the big ones (and consequently, possibly high reward ones) come with a lot of risk due to the financial power structure which is in existence. So I’d start with things those big three letter orgs might not be expecting to enter the business market.
1 Like
If you have sufficient funds in your bank account, it might be worth considering exploring crowd funding projects which are investing in AI technology, as they can provide exceptional opportunities for generating passive income.
Crowd funding is where communities of individuals pool their financial resources to kickstart potentially lucrative venture.
2 Likes
Passion is what makes us feel excited and eager to get going. It brings pleasure to what we accomplish and it drives us to be the best we can be.
If you run EOG it will lead you the way, the resources will become available. Don’t ask for the niche, or the blue ocean, your heart will tell you where to go.
3 Likes
2 cents
You also don’t need a blue ocean, plenty of massive success comes from saturated fields full of competition. Better that if your heart’s in it then just looking for the next goldrush.
1 Like
1 quantum computing
2 consult AI to answer the questions:
2 Likes
why not just run quantum limitless for a year ??
Trying to find the next blue ocean will most likely lead you to frigid water, slam into an iceberg on a cold night and sink before anyone can help you because you didn’t bring enough lifeboats.
-
Run EoG until you achieve whatever goal you are going after.
-
Find something that you like to do so much that you might not mind paying to do it.
-
Get a product out at fast as you can to test if you are wasting your time. If failure repeat previous steps
-
If 3 is successful, work at it until you are a pro at it.
-
Promote yourself early. (Do this while doing step 4)
-
Build off of your business get really good at one thing that is your core business, then when the time is right build off of your core product.
-
Scale.
-
Whatever the hell you want after you captured your goal of the market share.
Very rudimentary, but you get the picture.
3 Likes
One thing I figured I’d add here is that since starting IGUPX, I started considering generative adversarial networks, variational autoencoders and the like, and this whole idea of generative AI art. If you haven’t been living under a rock (no offense intended to rock hoppers), you’ve probably seen just how many AI art products there are out there that offer the same boring stakeholder capitalism based subscription model to use their services and strong arms the user into playing by their rules if you want to use the services. And there are very few (read, probably none) desktop based programs that let a user run the VQGAN-CLIP model on their own PC without having extremely powerful hardware. Ditto for photogrammetry applications.
I think a community funded implementation in the Rust language or something else similarly low level would be a great idea. All it requires is for someone to understand the complex math behind it and implement their own version which doesn’t rely on the existing open machine learning libraries and get subject to performance costs due to the large number of internal manipulations of data which could be performed more efficiently. Then simplify an interface that allows a person to control the training of the model so that they have the final decision on what content to run it on, to avoid any potential copyright or legal concerns.
If resources were not an issue than I would say open source and free software with legally separate paid components is the way to go. Disrupting the industry and putting the creative power in the hands of individuals who are low income does not come without a sacrifice, however whoever does this will gain incredible publicity and personal respect from the users of such software. Just a thought.
For more content, here is a good summary of the tech behind generative AIs:
The main problem that VQGAN-CLIP solved was allowing transformers to scale by discretizing the input and using an image classification network to drive the discriminator for the GAN. At a high level its really pretty simple stuff, all artificial intelligence right now works through dimensional reduction / input compression and learning network weights through trial and error; even LSTM is simply moving part of the problem of learning context into a special type of neuron whose implementation takes a certain amount of prior data context into account. The same mathematics that underlies the perceptron scaled up in complexity is present in any network.
1 Like
Technology can be disruptive, it can have an advantage or disadvantage on small-medium enterprise. Innovation is a priority investment for every entrepreneur. But don’t just choose technology to bring peoples attention into your presence, rather use it to enhance what you already have.
Now a days we have a choice between modular or proprietary solutions.
2 Likes
Of course, although as an entrepreneur you have to think about your ROI and what the value proposition is. Good publicity is invaluable to drive things like sponsorship, trust in your products and your business, and so on.
What I’ve noticed in my own examination of these technologies for potential inclusion in my own startup is that there are restrictions due to a subscription model and risks that any start up owner has to mitigate when dealing with powerful technology like AI, because of all the emergent legal issues and the risk of relying on these other small start ups for creating your own product (eg what is the business risk if the generative art company went bankrupt? What about if an image you used in a video happened to be based on some other artists work and you got sued? etc). Risks that are easier to mitigate when you are in control of the network or at least understand the content used to train the models.
My own personal preference in my own business model is to provide other creative people with the tools they need to be at their own very best and form a mutually reinforcing rather than competitive approach. That might seem naive at first glance but with good networking I believe it ultimately will win out over focusing on the competition.
AGI enters the chat