After Machine Learning Is Everywhere
ML is both pervasive and nascent. It’s been around for 60 years, yet enterprises still struggle to grasp what it’s limitations are.
Your product has AI, can it clean my teeth, raise our stock price, and make me a pizza too?
A flurry of startups are coming out with annotation tools, annotation services, training tools, models, frameworks etc. It feels like Javascript did 8 years ago with a new framework every day. Teams that used NodeJS benefitted tremendously from this tech as it allowed a smaller team to do more.
Now that’s happening with machine learning too.
The lead time on purchasing and adopting new tech in a large organization is somewhere between 8-18 months, though some fortune 500s are already deployed, and some have yet to start this clock.
So the question is what happens when machine learning is everywhere, and in everything? It will be accelerated either by chips, GPU, or novel model architectures. It will be cheap and pervasive.
So let’s assume we’re there now. Every org has machine learning capabilities to enhance their company. There are no more “AI startups” because if you’re not using it to gain a competitive edge, you’ll be left out.
Possible effects:
- The talent shortage is over. A few kingpens create the majority of novel architectures, disseminate them through academic papers that come with code to reproduce the steps. The code is repurposed and productized by enterprising companies young and old.
- Data science becomes overpaid data entry. Clean data, unique data sets are easy to gather, clean and deploy across an array of devices. Cloud infrastructure enables click to deploy machine learning models that can be federated on locally private data.
- Data, clean usable data. Resell this? Hoard this? Use it then resell it? If it’s cheap and pervasive is it the new oil?
- Incrementally better tracking, targeting, and advertising.
- Retail efficiency
- Employee theft nearly eliminated. Point a camera at the cashier, correlate with POS system
- SAAS being applied to companies who were never thought of as SAAS (how can a real estate company be a SAAS company, how about a commercial real estate portfolio?).
- Will open source data catch up with GAFA? Likely no, but they will begin to be widdled away at. The chink in the armor will show and it will look possible that someone could unseat a Google.
- Both the point above and the death of the Innovators Dilemma. Large companies who know the playbook of the Innovators Dilemma realize it does not need to apply to them anymore.
Liked the article? Sign up for our newsletter. This article was brought to you by SugarKubes. Sugarkubes is a container marketplace. Want to start running AI at the edge? Need some sweet machine learning models that work out of the box? Tired of paying per api call? Check us out at https://sugarkubes.io.
