What's Next - GenAI

I just did a post on a Gen AI omission problem, the near future plus optimism. This morning on that post someone asked me what I thought the biggest development in Gen AI would be this year, and after a while with my favorite tool (the whiteboard), here's my answer:

Three things:

1. The New Enterprise Platform

Someone will launch a broad enterprise platform that includes:

• GenAI,

• realtime Edge,

• KM with smart harvest/ingest/exhaust

• clever human interaction to afford continual improvement,

• retro-alimenting, evolutionary intelligence, and extends to a firms entire ecosystem

• next gen GenAi infused RPA/BPM

• snackable smarts architected into longitudinal learning (what Exec Ed and most Ed should be), including XR

Maybe, like the Web Services early days, this will first show up as half a dozen elements that need to be cobbled together.

It will be the next ERP-like platform except truly useful and not a Gordian Knot to implement and support.

New competitors will emerge architected from scratch for this reality and will have a damaging impact on incumbents. A small version of this process is what Workday Adaptive Planning did to PeopleSoft ERP Solutions @PeopleTech .

(I am happy to help construct this...need 14 months and about $8B in funding for top team/acquisitions/build. Getting slices of the relevant market market cap is about a $300B valuation in a couple of years ).

2. The Demand for Intelligent

We will see a broad demand for 'intelligent' across everything. This will be in products, services, experiences, supply chain, commerce, and for organizations themselves. This will be a brand imperative driven by economics. The simple design question will be 'is it intelligent'? That will require understanding that it is a two-part question. The first part is binary. It is or it isn't. The second part is understanding that 'intelligent' is a relative term. It is really not a thing by itself it is a comparator...in this case, to your competition. Like Neanderthals and early Homo Sapiens, for example. Note that the latter flourished while the former became extinct.

Think about it. You have two products, services, or experiences. One is Intelligent, and the other isn't. One is proactive, preventative, pattern matched, peer connected, process performant precise, predictive, and personalized... the other isn't. Which do you choose ? Think of it as a hospital you have to pick for yourself. Layer on top of that the Intelligent one is also cheaper.

That will have huge ripple effects on strategy and competition as some successful companies realize they are the best knife fighters at what has become a gunfight.

In business terms, the gun versus the knife is simply more agile, has a better cost structure, higher quality everything and better future vision. Yes, it's not the best metaphor, but it's visceral.

Of course, none of this works with bad, uninformed, past-anchored strategy. Great tech with bad strategy is just a faster, better trainwreck.

3. Super Hyper Scale Era (distant future - maybe late 2025)

The ability to process massive-scale data at speeds and price points almost unimaginable today that make economic sense is the hallmark of the next era.

This will include massive Edge capability moving more AI to the realtime world, you know, the world we live in.

GenAI will be a utility and quickly relegated to how we feel about water from taps or electricity from plugs. Both of those are true marvels, but we really don't get excited about that. Ocient is an early example of hyperspeed and hyperscale data.Vantiq is the same for real-time Edge intelligence, including brilliant GenAI integration. Dreamscape Learn is the beginning of what's next in Education. Positively leveraging neuroplasticity will be incorporated into the 'health', performance and education fabrics. NESTRE Health and Performance is an early example. CrowdSmart.ai already has AI plus GenAi, enabling cumulative IQ and continuous improvement. This has started and is coming. The real trick of LLMs is the scale/speed/cost structure. That specifically, by the way, will be 200x more efficient in three years.

This will change science, education, and healthcare, and by change, I mean change what they are, not just fancy them up around the edges. Many sectors, supply chains, branding, and commerce will change.

This will begin to kill what we know as advertising and advertising-based social media. The latter being one of the less 'intelligent' things you have. Platforms that harvest your commercial value for themselves with neuro-trickery content manipulation that eat a chunk of your most valuable asset (time) and have various negative human/societal impacts won't be hard to supplant when intelligence at scale lands. Think of a platform that does the positive connection things of social media today but that pays you for your commercial value, optimizes your commercial existence, doesn't have horrible societal side effects, saves you time in myriad ways and positively impacts your brain rather than uses it to trick you into dopamine loops for their purposes. (I need a scale partner and $300M to build that)

Like my last post above is a picture of a 'tech guy lost in a hedge maze'. I used Google to get this, which I think was 100x+ less carbon footprint than having Dall-E2 generate it. Yep, 100x...

Timing for building new solutions is tricky. I have participated in several absolutely brilliant technologies backed by top VCs (e.g. KPCB) or led by world-class people/brains (e.g., Prof Jim Cash HBS) that were just too early. Inforay, Metamatrix, and a KPCB-backed excess inventory play when I was at Moto are dramatic examples of being too early. You can't sell the best electric bike before bikes take off.

To end this, here are two quotes:

“Prediction is very difficult, especially if it's about the future!” Niels Bohr

"The future is already here; it's just not evenly distributed." William Gibson