TLDR: “Anything you can do, I can do better” - Computers
None of my predictions have come true, but they seem a lot more possible today than they once were.
About eight years ago, in a post on LinkedIn, I made a prediction that tools like Tableau would turn from data lake aggregators to decision making control centers. That AI would develop insights into the business and make recommendations over what would need to change to achieve various operating priorities.
Then about five or six years ago, in another post on LinkedIn, I made a prediction that advertising would become a self-improving by automatically split-testing different forms of copy and images that were automatically generated, and measuring their performance similar to the closed loop system that was driving so much growth for Amazon’s advertising business—Advertising Cost of Sales (ACoS).
Then four years ago, I made a presentation for my sales team about what was changing in sales technology, and I outlined a strategy I was calling “Autopilot or Pumping Gas?” The premise was essentially that the most time-intensive sales activities were driving the least amount of revenue. On a scale of time invested vs. revenue generated, things like research, CRM management, and outreach were all automatable tasks compared to the in-person negotiation that actually moved a deal from fiction to reality.
None of these predictions are materially true yet, but the reality is that they are all more possible today with recent breakthroughs in NLP and LLM’s like few shot learning and programmatic labeling.
Soon we will have executive co-pilots that understand our data lakes—these are essentially neural networks, internal-facing versions of ChatGPT if you will.
Soon we’re going to have self-improving advertising—the copy & imagery tools are already available, and we’re just waiting on the attribution stack and/or new distribution models.
And on sales automation, well that one is just getting started, but I’ve proposed that CRM’s will become listening technologies many times, and when that happens sales & marketing alignment will reach its potential, buyer and seller marketplaces will form, salespeople will go from gas pumpers to board room negotiators, and the datasets will lead to novel vector databases that can be leveraged across departments, federated learning models for technology providers, and clearer paths to sales efficiency for investors.
Taken together, everything is moving to autopilot.
Maybe that’s obvious and that’s what ‘software is eating the world’ really means.
If you think about a coder’s job and the development of new languages, engineers have been writing code to make their own jobs easier for decades. They’re automating the repetitive tasks to spend more time on the creative and challenging ones.
And I think that is happening because we have created a scoreboard for the human race. Wealth.
How do you become a world superpower? Money.
How do you feel like you are winning the game of life? Having more than others.
How do you create more capital? Borrowing, and efficiency gains.
If the scoreboard is personal wealth, corporate wealth, country wealth, then the path is simple. Do more with less.
And we’re getting exponentially better at doing more with less, so the question is, will you be happier when you have less repetitive tasks and more creative work to do or will you be out of a job?
See you Monday.