All your Bots R’ Us

Marc Canter
AI Blogging
Published in
7 min readFeb 3, 2017

Bannon’s puppet antics won’t prevent me from getting actual work done. Though he (Bannon) will go down in city as one of the greatest Machiavellian, Rasputin-like, Dr. Frankensteins in history — I owe a debt of thanks to Steve.

He’s educated an entire world to the power of the Puppeteer.

That’s the power we want to be selling in our new product — Instigate.ai.

We think of Instigate as a new kind of powerful tool, but my daughter and her friends will know it as a new kind of self-expression — a Personal Bot.

One will have their own Bots, and they’ll want to share their Bots with their friends.

They’ll keep their Bots private and control what their Bot says and does with their friends or they’ll welcome their posse in and create a sort of Group Bot together.

The Puppeteer creates a world of possibilities in their Bot. That’s why we call them “Instigators” — as they’re birthing a potential, not explicitly telling the Bot what to say.

Sure they can script and craft all sorts of experiences, Q&As, comeback and quizes, but they can also just tell the Bot what they’re feeling like and what they care about.

Conceivably one could even birth a Bot and make it publicly accessible. This creates a whole NEW set of issues and potential!

Yes — I know what you’re thinking, that would lead to create CHAOS!

But we got a whole “pruning angle” to show off (which would be one of the ways we’d make money) — so we’re not afraid of the Alt-Rights, Trolls and chaos.

Personal Bots will be a new form of expression — picking up where Blogging and YouTube videos leave off.

Personal Bots allow us to go beyond commenting, where “viewers” or “readers” (we call them “Interactors”) can become part of the content.

End-user comments, contributions, interaction and participation — can all be incorporated into the interactive content itself.

That’s what we’re building.

I’m sitting here, fighting for attention and mindshare so that I can keep the momentum moving forward with Instigate.ai.

I’m busy crafting linguistical techniques so that our Bots can be the most attentive, whimsical, intelligent, articulated Bots on the block.

Our users need to be TOTALLY captivated — so we need a system which adapts itself to whatever a user wishes to do (while keeping it all butt simple),,,,,,,

,,,,,,,,while underneath the covers we’ve got a system in place that can listen, learn, turn feelings and intentions into meta-data, index and label everything being input into it, create Pop-up suggestions, ask questions, walk down a list of actions, offer suggestions, increase and expose vocabulary, guide the user through pathways — while learning and growing over time.

And that has nothing to do with the data itself. Where does it come from? What is it based upon? How did you get it? Is your data the same as mine — can’t we share?

Do you own your data or will you be renting this data — which you’ll become addicted to it (and as any heroin addict can tell you — that’s not necessarily a good thing.)

Having the right data is exactly what differentiates successful, positive Bot experiences from horrible, boring, worthless Bot experiences.

So teaching your Bot to be smarter — should be the onus of the end-user. And your Bot’s responses, wittiness, range of expression and media “comebacks” — should increase — as you invest time and energy into training your Bot.

What I’m trying to say is that “Users should BENEFIT DIRECTLY as their Bot gets smarter and smarter.”

By providing direct benefit and constant evidence that one’s Bot is getting smarter — CAUSE OF ME — we believe a magical relationship will be established between the human Puppeteer and their Puppet.

In fact we’re betting on it. That’s the magic sauce of Instigate.

Why Instigate? To get Magic.

Comparing one’s growth and progress through the system to others’ progress in growing their Bot will be an “gamificiation” motif which should also (hopefully) keep “them” coming back for more.

Having one’s own Personal Bot and watching its progress — is sort of like a modern day version of a Tamagotchi. Only your Bot can talk to you, playback your fav snarky video “comebacks” and gets smarter as more and more of your friends — join your posse.

I’m imagining Animorphic Ariel taking her journey through the Instigate Bot neighborhood, growing and training her Personal Bot — via video selfies, contributed meta-data and compelling content.

Ariel invites her friends to contribute to her Bot and creates and uploads a whole bunch of Ariel’s Selfies — both photo and video — which get used as the mechanism for training the Bot. Her Bot “articulates” itself, expresses (via videos) and even can lay down a soundtrack (as long as you provide it to her.)

Yes — we can make semi-autonomous beings — but they’re only as good as the videos and music you provide it. That’s where the state-of-the-art is — right now.

Yes — you can add in generative language, dynamically controlled conversations and even establish guidelines and pathways to guarantee “compelling experiences” — but the secret to GREAT Bots is their content.

And where should that content (data) come from?

Well from the PEOPLE! That’s where!

That’s what Dave Winer said when he invented blogging.

That’s what Steve, Jawed and Chad said — when they launched YouTube.

That’s what Matt Mullenweg has been saying for 15 years!

The People are the ones who will provide the content — dudes and dudesses.

Ariel earns points for everything she does and as she draws others into the environment — she TOTALLY gets MORE points when THEY get points. Its a virtual pyramid scheme!

Meanwhile — the system’s structures and ontologies guide me through the arduous task of finding enough data to even START our machine learning process. Each task is its own universe — with topics, mythology, people and context — as underlying data structures and key functionality.

For any learning process to be successful, the feature engineering of the neural networks need to understand (in mathematical, algorithmic form) just what is it that we’re learning for you.

What are the sample queries, searches, learned concepts, parallel and serial pathways and tasks, background processing, foreground actions — ALL OF IT?

Be specific please — cause there’s nothing worse then spurious, unlabeled or nefarious data.

Everything that has to be learned, must be an algorithm. And for that algorithm to work, it has to be based upon (typically) VERY large data set.

So how do you start?

BTW there’s a new company idea in here for somebody — a DaaS (data as a service) play that can sell me learning data for $.000001 a transaction, so I don’t have this HUGE wall to climb — just to get started.

We need data sets like: “Young Love”, “Silicon Techie snark”, “Leftist attitude pontificating”, “Valley grrl snarl”, Sports (NFL and College Football for me), Music (Reggae, Bhangra, Classical, Jazz), Movies (RomCom, Sci-Fi), TV shows (Time travel) or Food domain database. You offer up a menu of domain databases like that — and you’re in business! Then we ALL can START doing some serious deep learning.

But without this data, I’m fucked. So what do I do? Hire entire classrooms of young data scientists?

As AWS was to the cloud, so must some newDaaS startup be to Deep Learning. Without data, your AI is shite (to quote the British.) This is the AI startup I outlined — in New kind of data play — please.

And lets be clear — all this data I’m talking about — is “fictional data.” Entertainment, Fun. Hah hah — you remember that kinds stuff — right?

Before the Evil empire and all our worst nightmares came true?

Instiage.ai will start off as strictly an entertainment product, based on ictional data. But as more than one friend and potential investor has pointed out — there’s no reason we could start plugging in factual data — right?

“One step at a time — young Padawan. Go slowly, perfect, refine, then expand.”

Placeholder holder logo

Try as they can, they can’t take my attention away from my journey.

I find myself, right now, using this writing as a respite — a break from the intense documenting, graphical prototyping, pivoting, agile adjusting, rotating, UX/UI creating, fractal turning, LEANing, gazing into a projected use case of my daughter building her Personal Bot — for herself, her friends and (as she most aptly asked me) “why are we coming back Dad? Why do I care?”

Instigate.ai is for my daughters. And my sons — as well. And my wife — too!

And everybody I know, have ever known and will know.

Canter family all in one place. Its a rare day indeed, as two of my sons live in NYC — with their women. Third son is in 2nd law school at Berkeley, so we get to see him fairly often but my wife is so busy playing bassoon (and being a certified PMP) that she barely even gets to see — me.

For young people who aren’t hung up on what’s current, as they know they’ll ALWAYS be something new coming along.

Young people who assume that they’ll be able to utilize AI as a tool, just like a knife or screwdriver.

And they’ve long wondered “why is commenting for lame?”

“There’s GOT to be a better way of interacting with an author or creator!”

Instigate.ai is live folks — so we begin.

--

--