Exponential curve. Straight up.

Welcome to AI Collision 💥,

In today’s collision between AI and our world:

From smart glasses to…
Google as a chip maker
AI fear and dystopia

If that’s enough to get the smart glasses fitted, read on…

AI Collision 💥

Siri, Alexa… Astra?

I bet you’ve never heard of that third one?

Neither had I until recently. But “Astra” may end up being that most commonly used name in the world.

Astra is Google. And if Google has their way with our future, Google may even become Astra.

AI assistants are coming. Arguably they’re here. But not at least on a useful scale yet. We’ve seen chat bots, AI agents, but we have not yet really seen a mass release, day-to-day, go with you anywhere AI assistant that helps guide you through the world, through your life.

But we’re close.

And Astra may be the answer to it all.

We know that Amazon has been working hard on levelling up Alexa’s AI assistant capabilities. We’ve covered that a few times here at AI Collision.

You can see what direction Alexa is heading here:

While Amazon and Alexa may become the first to-market with their AI assistant, Google is going to be hot on their heels with Astra.

In a recent interview with US 60 Minutes, Demis Hassabis, lead at Google’s DeepMind and AI pioneer believes that we may well have artificial general intelligence (AGI) in five to ten years, but in the next 30? He expects a system that understands everything around you at a deep level that is abundantly useful and ubiquitous.

And he believes the speed of change with AI is exponential. A curve “straight up”. One term in particular he refers to is this idea of “radical abundance”. It’s the elimination of scarcity, thanks to the advances of AI technology.

It’s a novel idea at least. Achievable? We’ll see. If AI can accelerate drug development and understanding of disease, if it can help us to harness free, abundant energy, and rocket productivity higher, then maybe radical abundance is a possibility.

And I’d suggest he’s also confident that whatever the outcome, it will be Google’s Astra at the heart of it.

The Project Astra they refer to you can see for yourself here – Project Astra – Google DeepMind

What I found interesting here is that Google also sees the long-term consumer device that we’ll access AI through is smart glasses.

This is something that Amazon has already been doing with (you guessed it) Alexa (which I had tested out a couple years ago at CES in Las Vegas, see below) and also Meta is doing with their Orion smart glasses and Llama AI.

I wonder if iPhone and Android dominate the smartphone market now, is it Astra and Alexa and Orion in the future, or two of them…or is there just one winner?

While this is the novel consumer technology, I’m not sure it tells the real story.

Google is pressing ahead hard with their upgrades to their AI ambitions. Be it through Project Astra or through Gemini. But I think the real story here is the development of their Ironwood processor.

You see, Google clearly sees AI as their future, as does Amazon and Meta. But in order to achieve that future, they need power, a lot of power. That means huge data centres which we know they’re building. It means huge amounts of energy which we know they’re tapping into.

But it means huge amounts of hardware, in the form of AI chips. And while we know they predominantly look to Nvidia for these, Google is also deep into the development of their own AI chips, or TPU’s (tensor processing units) as they call them.

And quietly, somewhat under the radar two weeks ago they released Ironwood, their most powerful TPU to date.

The leap in power from their previous generation TPUs is significant.

As Google explains,

Ironwood represents a unique breakthrough in the age of inference with increased computation power, memory capacity, ICI networking advancements and reliability. These breakthroughs, coupled with a nearly 2x improvement in power efficiency, mean that our most demanding customers can take on training and serving workloads with the highest performance and lowest latency, all while meeting the exponential rise in computing demand

I still don’t see Google making big enough leaps towards Nvidia, or a competitor in the chips space like Cerebras short term. But it is clear this is something they see as being critical to the future of Gemini, AlphaFold, DeepMind and Astra.

So, if you’re weighing up chip companies to invest in, sure, think of Nvidia, think of AMD, Broadcom…but also add Google the chip maker into the mix too. It may end up being a much bigger and lucrative part of their business than we might first assume.

Boomers & Busters 💰

AI and AI-related stocks moving and shaking up the markets this week. (All performance data below over the rolling week).

man in black suit jacket and black pants figurine

Boom 📈

Brainchip (ASX:BRN) up 25%
iRobot (NASDAQ:IRBT) up 8%
AeroVironment (NASDAQ:AVAV) up 7%

Bust 📉

Allegro Microsystems (NASDAQ:ALGM) down 15%
Nvidia (NASDAQ:NVDA) down 12%
AMD (NASDAQ:AMD) down 10%

From the hive mind 🧠

Financial services is the perfect breeding ground for advanced AI. It’s nothing but data and information where high-speed accurate answers command big money. But how does that change the way in which financial services are created, delivered, and does it erode the nuance of the markets and ultimately eliminate the need to invest at all? Maybe we all just end up with Astra that invests for us?
I was very literally just listening to the radio as I prepared today’s edition, and heard an ad for the new Samsung S25 smartphone, packed with Google’s Gemini, thinking to myself, why is Google packed into a Samsung phone? Then I saw this article and it all made sense.
Ahh, it’s been a while since we’ve gone back to out and out AI dystopian fear. So here’s a big chunk of it.

Artificial Polltelligence 🗳️

With the Easter break past us, we’ll get back to polls on Wednesday and Friday with a review of the results on Mondays.

Weirdest AI image of the day

Big Bird deported to a federal correction complex on Mars

ChatGPT’s random quote of the day

“Debugging is twice as hard as writing the code in the first place. So if you’re as clever as you can be when you write it, how will you ever debug it?”
— Brian Kernighan

Thanks for reading, and don’t forget to leave comments and questions below,