The future of machine learning
Sooner than later machine learning is just going to be an afterthought. Automated efforts to utilize and optimize machine learning will allow deployment as necessary and only to the advantage of the use case in question. We are also seeing so much of the machine learning efforts being packaged and built into products being delivered by day to day vendors. Yeah, I’m arguing that the future of machine learning is that it becomes omnipresent and no more or less exciting than virtualizing the hardware layer away from your day to day operations.
For every edition of “The Lindahl Letter” I pick an image for the social preview icon that people get when they visit the online edition. Every one of the emails as far as I can tell always comes with one of my favorite images of Denver at the top of the email. Naturally, I went out to my repository of Google Photos and searched for something with snow in it to make a joke about the impending artificial intelligence winter that must be coming.[1] The best quick read about that is probably from Towards Data Science and is linked below.[2] That site includes a wonderful graphic about general slow downs in the world of artificial intelligence that occurred in 1973 and 1988 respectively. You can probably tell from the first paragraph that I do not believe another round of AI winter is coming based on how ubiquitous machine learning has become. At this point, I think it is more likely that machine learning will become so omnipresent that it's an afterthought vs. it ever falling out of favor.
It is entirely possible that machine learning will morph into something more unexpected or change names, but the same principle and spirit of using artificial intelligence to generate models that work inherently with data will continue without interruption. The reason I’m so confident about that is that the technology underpinning has fundamentally changed. Automation and the need to do things more efficiently has changed the paradigm of how things are expected to occur and with what pace and cost. Between those two influences people who are creating technology are going to use the best methods available to them to succeed. That means layering in machine learning or whatever amazing phrasing for it comes next to make the next leap toward a future where the intersection of technology and modernity occurs. That occurrence will happen without fanfare or any celebration; it will be an afterthought and assumed to be mundane at that point.
Links and thoughts:
Alaina from PCWorld did a great tear down this week “Azza Pyramid Mini 806 Teardown & Walkthrough
This week I watched a good bit of the WAN show with Linus and Luke “Talk me Down from The Wall... - WAN Show July 23, 2021”
This video from NASA was pretty decent “The Shuttle's Last Flight | An End. A New Beginning”
Check out the AI Show this week from Microsoft Developer “AI Show Live - Episode 23 - Prebuilt Docker Images for Inference in Azure Machine Learning”
Ok. I did watch the “Ghostbusters: Afterlife Official Trailer 2” this week and have to admit it would have probably been better as a 10 part series vs. a Thanksgiving weekend movie release.
Top 5 Tweets of the week:
Footnotes:
[1] If you were not aware of the history of AI winter references, then you might be surprised that it even has a Wikipedia page https://en.wikipedia.org/wiki/AI_winter
[2] https://towardsdatascience.com/history-of-the-first-ai-winter-6f8c2186f80b
What’s next for The Lindahl Letter?
Week 28: Machine learning certifications?
Week 29: Machine learning feature selection
Week 30: Integrations and your ML layer
Week 31: Edge ML integrations
Week 32: Federating your ML models
I’ll try to keep the what’s next list forward looking with at least five weeks of posts in planning or review. If you enjoyed reading this content, then please take a moment and share it with a friend.