Our AI Future — Jobs vs. Tasks
Automation, the Great Resignation, and Bullshit Jobs are all related.
I know, a delay! Don't worry, getting my sea legs on the new gig. More to come on that at some point! For those wondering, I'm really liking the team at CARE and there is a LOT of good work going on and to be on all things data and technology!
But also, thinking I might move my publish date to Tuesdays so that I don't ultimately succumb to the Sunday Scaries.
On to the good stuff...
When looking at AI news, I've started finding it useful to distinguish between jobs and tasks.
Wait — don't leave!
I basically assume that if you write "AI" in any piece of content on the Internet today, you're either going to have everyone flock to it or everyone's eyes glaze over.
Or ideally — both?
There's been a lot of e-ink spilled about the trials and tribulations of this era of birthing our AI. Are we on the right track for Artificial General Intelligence (Skynet)? Are we in another "AI Winter"? Is GPT-3 even good, or just mostly pseudo-linguistic garbage? So many questions!
Even the term 'AI' being bandied about so carelessly is a little obnoxious. Are we really ushering in the next era of intelligent evolution with recommender engines based on Enron emails and pictures identified as cats? It does seem that from a hype perspective, we've definitely crested onto 'AI' really just insinuating everything that...we just don't have a great way for doing it yet. Magic! Pixie Dust! AI! And we do kind of know who to blame — over-zealous slide decks from tech bros:
Difference between machine learning and AI: If it is written in Python, it's probably machine learning. If it is written in PowerPoint, it's probably AI.
AI is a buzzword, but it's still real.
But let's not throw the byte out with the bit-water here, either.
There's enough real firepower behind advanced modeling, recommender systems, classification, categorization, and more that the front edge of unsupervised learning, reinforced learning, and real-world applications do start to feel a little like magic. Here are a few:
Egads! And actually, these are all from JANUARY 2021! (Note: one place to track what AI can do, DeepIndex. Great for building those pitch decks for the board members.)
It's all over the map, though!
Yeah, that's the challenging bit about tracking what's going on in the AI world — it's everywhere right now. And it's only going to get worse (better? worse? better/worse! yeah — better/worse.)
To steal and wildly misuse an old adage, when it comes to AI — never was so much developed by so many to...so many. There are just so many folks carving out areas where algorithms can play a role in transforming their business.
One way of carving it up — jobs vs. tasks
This is not going to be that comprehensive AI post.
But the intent of this newsletter is to tease out what's happening across the digital / tech landscape — what lenses make sense to peer at this area, whether it's about social impact, or sustainability, or even just rote aesthetics. We aren't above stabbing in the dark a little, to try and find some ways of understanding the parts of the elephant that we are currently closest to.
That's all to say; I'm seeing one way of looking at AI is through the lens of what is going to affect jobs, versus what is going to affect job tasks.
Jobs that people have, and tasks that people perform — these are not the same beast.
McKinsey did a study in 2019 comparing human skills to computational skills, making the point that demand for technological, social and emotional, and higher cognitive skills will grow, while needs for many manual labor skills and even simple cognitive ones will decline.
There's a potential future here where we have an opportunity to design and develop systems that draw on the best of the abilities of a human workforce, partnered and supplemented with AI assistance. Cooperative systems will be more efficient, more effective, more transparent, less biased, and more resilient.
And really, this isn't a controversial or groundbreaking statement. We've known for nearly 40 years that teams made up of humans and machines form stronger alliances. It's almost cliché at this point to say that any...good enterprise is coupled with good data. 😎
The Great Resignation meets Bullshit Jobs
I also picked up the book Bullshit Jobs last year — right as we were heading into the era of the Great Resignation (which, maybe this doesn't actually exist?). This is a fantastic read from a thinker who we sadly lost just as he was hitting his stride. Maybe I'm biased because of my Washington D.C. residence and history of working in/around magnets for bullshit jobs (but seriously it's a good book). But a lot of areas that fall under Graeber's rubrik for a bullshit job (box tickers, taskmasters, duct tapers) are ripe for automation and AI-sponsored pattern recognizers. The days are only numbered for many of these roles.
The timing and possibilities around automation are interesting as we are having a conversation (read: negotiating for better offers) about the world of work, what we find most important after 2+ years living through a pandemic, and how we relate to this little organic sphere hurtling through space getting hotter and hotter.
Reasons for optimism
Jobs are changing. We are changing. COVID made us change even faster. It's not all bad. We are seeing is the algorithmification of job tasks, which can be good! It's good until it becomes the distilled recipe for a job, which is bad.
Computers are really good at some tasks. And humans are really good at some tasks. These tasks, are not necessarily overlapping, and in fact, they often complement one another.
In 1930, John Maynard Keynes predicted that, by century's end, technology would have advanced sufficiently we would have achieved a 15-hour workweek.
It seems long overdue — but better late than never.
And you know, when my little AI friend can take over writing this fledging newsletter (tell your friends!), I won't even let you know. You might see a hint of it somewhere. Dedicated readers will notice a subtle change in style, but like Michael Caine's Alfred at the end of the Dark Knight trilogy, you'll see me in Italy, and hopefully give him a little head nod of recognition — we finally did it.
What else I'm reading
The Great Siberian Thaw | The New Yorker — Sigh. We're all screwed.
We Need to Let Go of the Bell Curve — Comparing "normal" distributions to Pareto distributions. I'm still wrapping my head around this one.
Scott Galloway on the False Promise of Web3 | Marker — Really smart overview of all the hype surrounding web3.
These Turtles Fly South for Winter | Hakai Magazine — Cover photo alone did it for me.
Thanks for reading, Gabriel
GoodTinker is a weekly email from Gabriel Krieshok about technology, design, and social impact. If you’ve enjoyed this edition, please consider forwarding it to a friend. If you’re reading it for the first time, consider subscribing (it’s free!).