Niche Twins - AI today, not tomorrow.

Twitter is where I consume nearly all of my Artificial Intelligence (AI) related news.

But lately I've been getting serious echo chamber vibes, to the point where AI has become almost synonymous in my mind with 'chatbot'.

And for good reason.

The release of ChatGPT a little over a year ago now is what brought AI to the masses and in all likelihood it will forever be remembered as a watershed moment in the AI revolution - the tipping point from which we accelerated and never looked back.

OpenAI's release of their text-to-video model Sora yesterday, is further evidence of this.

Using just a text based prompt, Sora can generate entirely unique videos from scratch up to a minute long and with stunning visual quality.

It's INSANE how quickly OpenAI is shipping products.

We're undoubtedly accelerating. And at a rate that has even some of the most accomplished developers stunned.

While all of this is may be true, it can often feel like it's really just a bunch of nerds on Twitter who are the only end users playing around with these things.

In some sense, that's to be expected.

After all, nearly all world changing technologies start out looking and feeling like toys.

But I wanted to get off Twitter and into the real world.

What are the real world examples where AI is being deployed today?

Not tomorrow. Not five years from now. TODAY.

This is why your keyword research tool’s Starter Plan costs $129.95 a month. That and 1,004 employees who all have salaries that need paying… KeySearch doesn’t have any offices. Our team is lean. Just $17/month. ​ –> KEYSEARCH FREE TRIAL <–

Unlocking 2000 year old secrets

Have you ever heard of the Herculaneum papyri?

I hadn't either, until I came across this incredible article in Bloomberg Businessweek.

Get this - it turns out that the same eruption of Mount Vesuvius in 79 AD that led to the lost city of Pompeii, also resulted in the preservation of an Italian countryside villa that is believed to have belonged to Julius Caesar's father-in-law.

In that villa, there is a massive library of scrolls (referred to as the 'Herculaneum papyri') that may represent the largest collection of ancient texts ever discovered.

Only a small portion of the villa has been excavated to date, uncovering about 800 scrolls.

Insane right? These literally hold the potential to rewrite our collective human history.

There's just one problem. We can't read them.

What's equally incredible is these scrolls have been known about for hundreds of years, but because they've been covered in volcanic ash, they are so brittle they can't be opened.

And believe me we've tried.

As far back as the late 1700s, a priest by the name of Antonio Piaggio created a contraption that slowly onrolled the scrolls at a rate of about 1 inch per day. It worked on some, but damaged most.

Since then we've tried just about everything, including creating 3D scans of the scrolls and virtually unrolling them. But even that's yielded limited results with an unfeasible price tag.

Insert former CEO of Github Nat Friedman who was convinced that AI algorithms may hold the answer.

In classic Silicon Valley fashion he put a $1 million bounty in 2023 for anyone who could develop an AI software that could read at least 4 passages of a scroll.

Sure enough, a few 20 something-year-olds kids managed to train a model that was able to interpret what they refer to a 'crakle data' - more or less just dried up ink that's lifted up on the page over time.

To you and I, meaningless. But to a trained AI model? Valuable clues.

Their model was ultimately able to interpret that crackle and convert it into Greek letters, resulting in 5% of one of the scrolls being deciphered.

Small progress, but for a seemingly unsolvable problem, significant.

Friedman estimates that this will get up to 80% over the next couple contests - at which point there's not telling what we'll discover.

Do you want fries with that?

Earlier this week I had a business trip in Seattle and I noticed as I pulled up to my hotel, there was a MOD Pizza directly across the street.

A nice buffalo chicken pizza for Superbowl Sunday? Yes, please.

When I called to place my order a few hours later, it took me a few seconds to realize that I was talking to a conversational AI.

Look, I fully appreciate that there's nothing sexy about automated order-taking. It seems like it's been around forever.

And as it turns out, the conversational AI platform MOD partnered with, SYNQ Voice, has been around since 2014 - for over a decade!

But to be clear, it's not simply the fact that MOD has an automated order-taking process that interested me.

What absolutely blew me away was how fu*king good it was.

From start to finish it was probably a 60 second conversation. The voice sounded incredibly organic. There was hardly any latency. The AI wasted zero words in our interaction and made zero mistakes.

It was so flawless, that the second I completed my order, I hung up and started Googling the company and reading all about them.

SYNQ3 is already partnered with some big names like Chipotle, Applebee's, Five Guys and more.

But it doesn't stop there.

Carl's Jr., McDonald's, Rally's and Popeyes have all rolled out their own versions of voice AI solutions at select restaurants.

It's an example where the technology went from bad (we've all had that uniquely unhinged experience of yelling "I want to speak to a human!!!!" at a robot on the other end of the line), to still bad but good enough, to good, to now VERY good.

The bar has been cleared and the technology has caught up.

It's crystal clear to me that every single ordering service in the world will be handled by conversational AI.

Big Brother

And finally, I had to leave you with an Orwellian example to dwell on.

It was recently reported that Walmart, Delta, Chevron and Starbucks are all monitoring their employee messages uses AI software from "Aware".

The seven-year-old startup enables businesses to better understand company sentiment by analyzing employee communications.

Just recently hired a new COO? Interested in how the company really feels about it? Aware will let you know in real time.

In essence, it creates a company social graph.

While the data is anonymous and they say their intent is not to flag individual employees, their platform has the capability to do exactly that.

And sure they're clear such a targeted flagging is reserved for extreme cases, like threats or other high-risk scenarios.

In general, I'm personally incredibly optimistic about where all this is heading.

But at the same time, it doesn't take much of an imagination to see how this could all wrong.

In some ways Seeger captured it perfectly - “Technology will save us if it doesn’t wipe us out first.”

Have a Friday! ✌️

-Keith

Niche Twins newsletter sent weekly on Fridays at 8:30 AM ET

​​​

Want to get your company/brand/service in front of 8,700+ bloggers and internet entrepreneurs?
​--> SPONSOR OUR NEWSLETTER <--