I quit my job and started a business 6 months ago to detect things (cars...

>Do you have any direct competitors?
Lots of companies doing similar things, i.e. there is one detecting apple orchards

I don't know what specifically to build, what industry to go for etc.
>Who's your target audience? How have you been targeting them?
I spoke with energy industry, so I thought of estimating power consumption from night light intensity but apparently that's only useful for large-scale utility companies. I could only reach to small producers (solar plants, one hydroelectric plant manager etc.) and they said this is really cool but not useful for us, sorry.

I also spoke with a guy in oil and gas, he said it would be useful if I could detect stocks in oil production sites, but my sattelite imagery is too low-res to do that successfully.

So I guess I'm at a point of shutting down or thinking of something different to do.

expand
>detecting things from lowres photos
How you do dat?
>hurrdurr that looks like a cat to me
2000% accuracy rates

Well another thing is I can spend my time either trying to find clients or building something to show to the clients. My attempts at #1 failed so far, and if I do #2 I might end up building something completely irrelevant no one wants

How about a spy app btw. Phone zombies might pay 1.5 to spy on their neighbours roof

It was just an example but seeing how many cars are there in a parking lot, every day, could be used to see how busy a shopping mall is.

Anything about as large as a car, once a day (provided it's not cloudy). Idk what else can be done with this tech.

>Who are you targeting?
Anyone at this point
> What are you selling?
Subscription service, giving an annual graph of [things] in [locations] you want
> How much is your product/service? How does it compare to your competitors?
I never sold so I don't know, if it's a one-off job, $1000, if it's a subscription, $99/month?

competition seems similar, though most are "call us for a quote" model

how about looking for customers outside the US. If you are willing to deal with shady people and willing to be paid in kind, you could explore the "military" sector and smuggling

>>detecting things from lowres photos
>How you do dat?
I have satellite photos, updated daily. Like google maps. They aren't very high resolution, so I can detect a car but can't see a person. Similarly I can't see what kind of car it is.

Within that resolution, I can train an AI to tell apart the thing I'm interested in from the background. Then I define an area, let's say a city or a neighbourhood, and then the AI detects everything it thinks is similar.

>muh captcha
google beat you to it retard.

I live in the UK, have only looked around here
I can't even reach legit clients, let alone underground stuff. Not that I'd be interested in it.

You can pretty much only tell whether they have a roof and roughly what colour it is from that resolution, not very good.

You can, however, tell whether their car is parked in front of their home in any given day. But I doubt anyone would pay for that shit. And wouldn't work if someone else parked in that spot because the resolution isn't fine enough to tell apart what kind of car.

Yeah it's the same technique as captcha, it's just that I'm looking at stuff from space instead of from random roads. And I don't have millions of channers to click on shit to improve my accuracy

Have you thought about trying go get local government and business councils to buy? You could sell your product as a way to detect traffic/congestion during peak hours. This could be used to determine if a site is good for building a new store/restaurant. It could also be used to help city planners expand infrastructure and such.

How about cows? You can count cows for huge farms. They own so many.

>let alone underground stuff
Then get in contact with people who can. It's risky, but high risk high gain.
>But I doubt anyone would pay for that shit.
user, look around you. This is mostly a board where people put hundreds of thousands in 2 week old Tokens they can send back and forth. You could have a working product. Just do something or give up, There are no grilled doves flying in your mouth, except of you are an illegal immigrant and apply for asylum

>You could sell your product as a way to detect traffic/congestion during peak hours.
Don't have hourly imaging unfortunately
>This could be used to determine if a site is good for building a new store/restaurant.
That's an idea, like looking at how busy a parking lot is in an area, and having a graph of it.
>It could also be used to help city planners expand infrastructure and such.
Maybe. I only thought of real estate developers, might try those guys (though selling to government is always slow and no one knows wtf is going on)

it wasn't enough for him to bankrupt his own businesses 6 times, he had to bankrupt the whole fucking country

I can't get venture capital later if my invoices aren't legit anyway.
>This is mostly a board where people put hundreds of thousands in 2 week old Tokens they can send back and forth.
The trick is finding the right people to buy your shitcoin/shitproduct. Which is where I'm failing. Too much of an autistic engineer.

Too small to detect. Would've been a good idea though.

Newb entrepreneurial mistake bro.

You have to build the buzz and marketing before you build the product. Or you end up building in a bubble with no end customer base.

Always sell first and deliver on deals.

You are making a classic mistake. You don't know who your customer is and you don't know what value you are providing for them.

Until you know the answer to these two simple question you are doomed.

WHO is your customer?
WHAT is the value you are offering them?

Figure it out. Look into forestry, maintaining power lines and agriculture

That's the thing, I set out to find customers first before building something no on wants.

But couldn't get anyone interested within my limited contacts, even when I'm telling them I could build to their specs.

So here I am, six months later, with literally nothing to show for.

Today tanker was lost, what about searching for ships?

>Look into forestry,
There is a thing called change management, observing how much a forest changes over years etc. Environmental protection agencies already buy this shit.

But these are all large, institutional customers.

If I could figure out something useful for a small customer, solve some problem for a small-to-medium business, things would be so easy.

Bro that was your sign to not build it.

I get a picture once a day so probably not the best way to go about it

Well I have my sign then, now what. Look for a job?

You got a solution in search of a problem.

If you really want to salvage it best bet is to network far and wide. meet as many people form as many walks of life as possible and see if any one needs it

>Too much of an autistic engineer.
A sea. You need at least an autist who is autistic at social stuff. I would say hit the bars get drunk and don't stop talking about your project and its potential, no matter how often you get thrown out. There is no bad marketing.

Following is an anecdote but a friend, an autistic Russian engineer had an idea building some shit but was not able to get it out. Other friend who can sell skateboards to grannies, but is not able to count to three got into talk with him at random at a bar and 2 years later they sold the prototyp and patent to an investor in the UK for 5MM.

What I like to say is, you need to do something. Don't fear failing it sound as you allready hit rock bottom.

Lost ship is not going far within one day.
Find all ships, compare them to the database, exclude traced ships, BOOM, you found all lost ships, you get tons of reward, hero status, media coverage and future contracts.

Who would even care about that besides the military?

Smugglers. If I would be still in the biz I would have hired OP. Imagine being able to scan border before getting the product over. Risk reduction military level.

Hey OP, founder of an AI company here. We're dealing with Deep Learning. AI as a service business model. Mostly computer vision and NLP stuff.

You need data and ideas? Go to Kaggle. Do some competitions. Implement papers from arxiv and do demos. Doesn't have to be original stuff, just well-made. Businesses get super impressed. If the management aren't complete retards, they know AI is coming and once you prove you can do shit, they'll give you data and money.

Leverage your network and find contacts. You'll get to decision makers eventually. DYOR before getting into a meeting and think how you can you solve a problem for your potential clients.

If you want to show some credibility, write tutorials, write papers if you have the time (you probably don't) to prove that you know shit.

We're now 20+ people and bootstrapped. About to raise capital to get out there and fuck around with the big tech companies. Raising capital for AI is very easy these days once you prove yourself.

Not LARPing. Can't post proof cuz it'll reveal too much and /biz will nuke our website and blackmail me. You just gotta beleive what I say.

>not using RNNs on crypto and selling off your old bots to Veeky Forums so you can work on interesting projects instead of worry about money