Artificial Intelligence

[fisicx]

But does Cuthbert want AI? Maybe he just want a better detection system. He will do his research and may discover AI will do the job. He will then look for AI detection algorithms or whatever.

The problem with your site it it tries to cover every base instead of focusing on solutions. If you want Cuthbert to find you you need to have a page on the site that specifically targets cybersecurity systems. One could argue that you need multiple sites each targeting a niche use for the software.

The key to your success isn’t AI, technology or explainer videos. It’s marketing. If you are really good at marketing you will get customers. Your software doesn’t even need to be very good as long as you are a top notch salesman.

 

[Fennaio]

But does Cuthbert want AI? Maybe he just want a better detection system. He will do his research and may discover AI will do the job. He will then look for AI detection algorithms or whatever.

The problem with your site it it tries to cover every base instead of focusing on solutions. If you want Cuthbert to find you you need to have a page on the site that specifically targets cybersecurity systems. One could argue that you need multiple sites each targeting a niche use for the software.

The key to your success isn’t AI, technology or explainer videos. It’s marketing. If you are really good at marketing you will get customers. Your software doesn’t even need to be very good as long as you are a top notch salesman.

I agree. The site isn't complete yet. It's basically a wireframe. You can obviously see it's split into solutions and industries. All these will be expanded on and with some tech knowhow they'll effectively become targeted solutions.

I'm only in Year 2. Year 1 was developing the software (9 months including intensive research) and the websites, at the same time as making some survival income. The start of Year 2 was making enough income to fund the rest of Year 2 and half of Year 3 which is additional software functionality asked for by users (20%) and like you say...marketing (80%), which i know is key, hence why I'm on here getting valuable feedback, and I've simply not had the time to do it yet. Developing software like this takes a silly amount of time and without it there is no product. I'm now in the position to properly market it.

Rightly so, you are focusing on the absolute end solution and totally agree with that and I'm thinking hard about it.

The other side of it is it solves a key problem in the tech industry that developing AI is extremely costly - the overall solution saves 180k in development costs. It can be up and running, learning literally anything and can integrate with systems in 2 minutes.

So the next challenge, like you say, is marketing it effectively.

 

[MrPetrolHead]

Mostly used as a marketing buzzword for machine learning.

Why do I need an AI enabled fridge?

May I give you a few ? …At minimum AI in a fridge would help it optimise compressor runs. As the unit got older it would adjust those to stay optimised. Add a sensor under one fort and it would take into account the load. Add a teatime clock and the ability to communicate it would learn about your energy tarrifs. the ambient temp cycles and reshuffle its operation to run leaner, add a local space temp and it can tell when it’s out too much heat into its surroundings and maybas well shit done and let that dissolute a while.

 

[fisicx]

May I give you a few ? …At minimum AI in a fridge would help it optimise compressor runs. As the unit got older it would adjust those to stay optimised. Add a sensor under one fort and it would take into account the load. Add a teatime clock and the ability to communicate it would learn about your energy tarrifs. the ambient temp cycles and reshuffle its operation to run leaner, add a local space temp and it can tell when it’s out too much heat into its surroundings and maybas well shit done and let that dissolute a while.

That's not AI. That just a bit of clever programming.

 

[DontAsk]

May I give you a few ? …At minimum AI in a fridge would help it optimise compressor runs. As the unit got older it would adjust those to stay optimised. Add a sensor under one fort and it would take into account the load. Add a teatime clock and the ability to communicate it would learn about your energy tarrifs. the ambient temp cycles and reshuffle its operation to run leaner, add a local space temp and it can tell when it’s out too much heat into its surroundings and maybas well shit done and let that dissolute a while.

What does any of that have to do with AI?

 

[MrPetrolHead]

That's not AI. That just a bit of

What does any of that have to do with AI?

 

[Fennaio]

That's not AI. That just a bit of clever programming.

That's the point. AI works it out in a second and continually adapts . "clever programming" takes a lot of time, is specific, costs money for a very specific task and doesn't adapt. AI would happily adapt and learn continually.

 

[Fennaio]

What does any of that have to do with AI?

It's learning.

 

[Fennaio]

May I give you a few ? …At minimum AI in a fridge would help it optimise compressor runs. As the unit got older it would adjust those to stay optimised. Add a sensor under one fort and it would take into account the load. Add a teatime clock and the ability to communicate it would learn about your energy tarrifs. the ambient temp cycles and reshuffle its operation to run leaner, add a local space temp and it can tell when it’s out too much heat into its surroundings and maybas well shit done and let that dissolute a while.

Well said.

 

[fisicx]

It's learning.

That doesn’t make it AI.

This sums up your whole problem. What you call AI is not what others think is AI. A fridge might query a database and determine the best options for whatever but that’s not AI. Doesn’t even need to be ML, a halfway decent programmer could put this together.

People are finding reasons to shove their version of AI onto things that just don’t need it.

 

[MrPetrolHead]

So your saying if we embed the AI chipset and set the machine off to learn and optimise it’s operation,

That doesn’t make it AI.

This sums up your whole problem. What you call AI is not what others think is AI. A fridge might query a database and determine the best options for whatever but that’s not AI. Doesn’t even need to be ML, a halfway decent programmer could put this together.

People are finding reasons to shove their version of AI onto things that just don’t need

 

[MrPetrolHead]

You’ve just assumed based on what you know. I’ve tried to help you, but you just reverted. I’m a controls engineer. I’ll leave you be….

 

[MrPetrolHead]

So your saying if we embed the AI chipset and set the machine off to learn and optimise it’s operation,

Its more about it learning to find its way through its life. I has to learn to adapt to its own wear and tear and changes to its load, encironment and climate. There will be changes to where it lives, it’s usage pattern, the energy tarrifs, replacement parts. with new chanrxteristics You set it on its way, that’s it.

 

[Fennaio]

That doesn’t make it AI.

This sums up your whole problem. What you call AI is not what others think is AI. A fridge might query a database and determine the best options for whatever but that’s not AI. Doesn’t even need to be ML, a halfway decent programmer could put this together.

People are finding reasons to shove their version of AI onto things that just don’t need it.

AI does it without querying a database. AI doesn't rely on rigid code and instruction. It's all stored and continually learnt and adapted using an Artificial Neural Network.

I've worked in software for 20 years (sounds like you have experience in software too judging by your lingo) - there's a massive difference between making stepwise code instructions and classes/methods/db calls and an artificial neural network continually learning with no programming and without the need of a db.

No one is suggesting to force AI on a fridge (although it would work) but the principle is there, which is the point of that post.

With your db analogy....how do you think you'd achieve natural language processing with normal programming and db calls?

As a hypothetical scenario: just say you took the whole forum database of this website that contained all posts, conversations etc. You know the username and you know their post/message content. Let's say that db is now set in stone.

Your mission, should you choose to accept it; look at the content of the next person who posts something, and predict, from their content, who the username is. You have to program something that will do this and give you an accurate (I'll be very generous and give you 80% accuracy and 1 db call) in less than 20ms. As a bonus mission, predict the order of words that they might say in their next post. As a second bonus mission, now consider you still have the same db, but you now have access to all e.g. 100 data-points related to all users - you now have to predict (with 75% accuracy) the order of words that someone is going to type from a certain postcode at a certain time of day. If you can make any of these happen in less than 20ms I'll believe you that they can be achieved without AI.

This is all tongue in cheek btw

 

[fisicx]

I totally agree with you that AI can do these things but the next question is why? It is going to make the world a better place?

Whilst there may be many benefits in some applications (like energy management or traffic control) at a very local level and for most small businesses the benefits are minimal.

Once you can show how your AI can help an eBay or Shopify business or a sole trader builder and so on you may get some traction.

At the recent UKBF meet-up @Paul FilmMaker gave a demo on how to use video and it was a real lightbulb moment. That’s what we need from you. How to use AI in a small business.

 

[Fennaio]

I totally agree with you that AI can do these things but the next question is why? It is going to make the world a better place?

Whilst there may be many benefits in some applications (like energy management or traffic control) at a very local level and for most small businesses the benefits are minimal.

Once you can show how your AI can help an eBay or Shopify business or a sole trader builder and so on you may get some traction.

At the recent UKBF meet-up @Paul FilmMaker gave a demo on how to use video and it was a real lightbulb moment. That’s what we need from you. How to use AI in a small business.

How do you define "small business" ?

 

[fisicx]

Hairdressers, tradesmen, online stores and service providers, independent shops and takeaways, gardeners, builders, hotels and B&B, accountants, lawyers…. The list is endless.

In other words, almost every member of this forum.

How are you using AI to improve your business? Is it generating more leads, improving processes, cutting costs etc.

 

[Fennaio]

Hairdressers, tradesmen, online stores and service providers, independent shops and takeaways, gardeners, builders, hotels and B&B, accountants, lawyers…. The list is endless.

In other words, almost every member of this forum.

How are you using AI to improve your business? Is it generating more leads, improving processes, cutting costs etc.

Strange question to ask that if taken logically. If I transpose that to the list of small businesses you listed:

How are you using hairdressing to improve your business? Is it generating more leads, improving processes, cutting costs etc.

How are you using your takeaway to improve your business? Is it generating more leads, improving processes, cutting costs etc.

How are you using gardening to improve your business? Is it generating more leads, improving processes, cutting costs etc.

I would guess that the above trades are using their business to make money by selling their trade. My trade happens to be AI.

AI software is a service like every other small business on here.

 

[fisicx]

OK. So how can your (or any) AI help those small businesses?

You did say:

I could be wrong but I still think there's a misunderstanding by many people of what AI is and how it can be used in business.

I just trying to work out how it can be used in business.

For example. I build cusotm wordpress plugins. I get leads from all over the world via a number of sources. How could I use AI to get more bang for my buck?

This isn't me being all contrary, I'm really interested in how AI can improve a small business.

 

[DontAsk]

It's learning.

It has to be taught.

Put your AI system in a fridge. How do you interface it to the hardware? to the internet?

How does it learn how to control the compressor?

How does it learn how to interact with the correct energy provider to get tarriff informations? How long does it take to re-learn when the website changes? What happens in the meantime.

It's a really CRAP example that does NOT need AI.

 

[Chris Ashdown]

One area that is hardly ever mentioned in Weather forecasting is that classified a machine learning or AI, either way it brings useful information to all sizes of companies

 

[Fennaio]

OK. So how can your (or any) AI help those small businesses?

You did say:

I just trying to work out how it can be used in business.

For example. I build cusotm wordpress plugins. I get leads from all over the world via a number of sources. How could I use AI to get more bang for my buck?

This isn't me being all contrary, I'm really interested in how AI can improve a small business.

I know, I'm very tongue in cheek, taking on board your comments and respect them, genuinely.

Anything where predictions, insights and recommendations can add value is the honest answer...and the value of the prediction carried out in real time by AI has to outweigh that taken by a human to do it in their time (including the downstream consequences of getting a prediction wrong, factored in with the time investigating, pursuing it, loss of sale etc.).

In your case with lead gen, you will have built up a dataset of leads that turned out to be rubbish, some that came back eventually, some that were great, repeat customers etc. AI would learn the patterns (link all the data you have about the lead) quickly and give you an output or outputs instantly of what leads to pursue. Sounds a simple thing to do on the face of it but it isn't, especially if you've got 10000s of samples with multiple links within the data. AI will learn from any number of inputs and outputs all the time and it would give you an accurate output based on as many factors as you can think of - and what you tell it to tell you.

 

[Fennaio]

It has to be taught.

Put your AI system in a fridge. How do you interface it to the hardware? to the internet?

How does it learn how to control the compressor?

How does it learn how to interact with the correct energy provider to get tarriff informations? How long does it take to re-learn when the website changes? What happens in the meantime.

It's a really CRAP example that does NOT need AI.

Nothing to do with business but I use AI to feed my chickens. I've got my own software running on a raspberry pi connected to 2 actuators and 2 motion sensors. It has learnt the feeding times of the chickens and knows when to open and shut the coop door in the AM and PM and drop food. Means we don't have to get up at 4am to feed the noisy bastards. Back on topic now....

One area that is hardly ever mentioned in Weather forecasting is that classified a machine learning or AI, either way it brings useful information to all sizes of companies

AI is definitely used in weather forecasting.

 

[fisicx]

In your case with lead gen, you will have built up a dataset of leads that turned out to be rubbish, some that came back eventually, some that were great, repeat customers etc. AI would learn the patterns (link all the data you have about the lead) quickly and give you an output or outputs instantly of what leads to pursue. Sounds a simple thing to do on the face of it but it isn't, especially if you've got 10000s of samples with multiple links within the data. AI will learn from any number of inputs and outputs all the time and it would give you an accurate output based on as many factors as you can think of - and what you tell it to tell you.

But that's the issue. There isn't the corpus of data to work with. Same are most small businesses. The bloke who did our patio gets most of his work via nextdoor and referrals. He has a very basic godaddy website and does all his business using his phone. There is nothing for your AI to learn from.

It will be the same for most small businesses. Not enough useful data for your software to work with.

 

[fisicx]

Nothing to do with business but I use AI to feed my chickens. I've got my own software running on a raspberry pi connected to 2 actuators and 2 motion sensors. It has learnt the feeding times of the chickens and knows when to open and shut the coop door in the AM and PM and drop food. Means we don't have to get up at 4am to feed the noisy bastards. Back on topic now....

Don't need AI to do that. You can buy kit to open/close the door and raise the food. Chickens get up with the sun and go to bed before it gets dark - all you need is a light sensor.

This is the sort of thing I've been saying. AI has its uses but that doesn't mean everything has a use for AI.

 

[Paul FilmMaker]

I totally agree with you that AI can do these things but the next question is why? It is going to make the world a better place?

Whilst there may be many benefits in some applications (like energy management or traffic control) at a very local level and for most small businesses the benefits are minimal.

Once you can show how your AI can help an eBay or Shopify business or a sole trader builder and so on you may get some traction.

At the recent UKBF meet-up @Paul FilmMaker gave a demo on how to use video and it was a real lightbulb moment. That’s what we need from you. How to use AI in a small business.

Thanks! As soon as everyone sees how video generates more sales and they even see it in action, then everyone 'gets it.' They want it. Like you say, it's the 'lightbulb' moment.

So it's a fair question.

How does it work to increase sales?

 

[campbeji]

So AI?

In my head, AI will have to be Intelligent, which means that it makes decisions based on things it has learned, but not just that, it also needs to be able to decide what to learn.

The best example of what AI should be, and probably will be in the future, comes from Douglas Adams and 'The Hitchhiker's guide to the Galaxy

“And to this end, they built themselves a stupendous super-computer which was so amazingly intelligent that even before its data banks had been connected up it had started from I think therefore I am and got as far as deducing the existence of rice pudding and income tax before anyone managed to turn it off.”​

The systems we have at the moment are not 'Deep Thought', probably closer to the Acorn Atom (I just love that name).

When my son was born I was convinced he was defective, he couldn't do any mathematics, or communicate beyond several varying dB levels of high-pitched screech, and when I sat him in front of the computer to do my accounts, well let us say that was less than successful. 25 years later and he has improved a bit, got a 1st in his degree that he did in two years, he can talk quite well now, mind you, he still won't do my accounts :-(

In 25 years we will all be saying either Isn't AI wonderful and so helpful or How may we serve you our wonderful AI overlords?

Jim