AI is quickly integrating into education, with major releases from Anthropic (Claude) and Gemini (Google). However, OpenAI's contribution is pretty underwhelming. I call it the lack of the Actionable Outcomes Bias Principle.
AI is quickly integrating into education, with major releases from Anthropic (Claude) and Gemini (Google). However, OpenAI's contribution is pretty underwhelming.
I call it the lack of the Actionable Outcomes Bias Principle.
Nearly 20 years ago, I witnessed firsthand how a single smart tech decision could transform a 'boring' business overnight.
The company I was working for was in one of the most boring (yet necessary) sectors in the world. Transport and Logistics š¤®
Boring, uninteresting, and 100% necessary.
Now, before all you modern business people jump in on the conversation and blast me with all sorts of facts about the importance of supply chains to global business, keep in mind the context in which Iām referring.
Today's modern world of globally integrated supply chains and just-in-time manufacturing recognizes the intrinsic value of effective transport and logistics. Business schools all over the world teach their students about how important logistics is to global business leadership.
Back then, it was one of those boring, slow-to-innovate sectors that everyone knew was necessary, but preferred to ignore.
Yet one seemingly small change led to an overnight 20% increase in productivity.
In this article, I'll share some thoughts around why this happened, and then, how it applies to my theory that OpenAI is losing its edge.
Sign Up for AI insights, analysis, and how to guides for all things AI.
No spam. Unsubscribe anytime.
At the time, I was a very junior office worker who had just given up my forklift driving career to move into an administrative role.
I had no idea that in the future I would get an electrical engineering degree, become a cybersecurity expert, and move into the world of technology.
What I did know is that the system we used to track all of our logistics sucked. It was slow, annoying, tedious, and honestly, looked like poop.
Of course, all the management types at the time loved it because they could build fancy dashboards and track all sorts of interesting metrics.
So basically, it was locked in.
However, thatās not the worst of it. Not only was this system slow, tedious, and ugly, it also had this annoying feature (I use this word politely) that meant every time you pushed a āsaveā button, there would be an up-to-one-minute delay while it updated the server.
One minute where you could do nothing except watch an annoying rotating hourglass animation, hoping and praying that it would remember to save all your work and that you didnāt have to do it again.
Heaven help you if something urgent happens while waiting for it to save!
Then one day we turned up, and that one-minute delay had evaporated. That annoying save button now took 3-5 seconds to do its thing.
It felt like magic.
As I later learned, the change was something that today we consider pretty simple.
The business had paid a seemingly exorbitant amount to lay a direct fibre line from our warehousing facility to the central server, boosting our data throughput by a factor of at least 1,000.
We still had to track and enter endless amounts of metadata.
We still had to use this slow, clunky, ugly interface.
We still had to turn up each day and work through the seemingly endless amount of data, spreadsheets, and reports.
We still had to push the save button.
However, that one change increased our productivity by 20% overnight and locked in a powerful competitive advantage for the company.
Over the years, Iāve had many opportunities to reflect on this remarkable experience.
There are many businesses Iāve worked with who would happily pay millions of dollars to achieve a 20% productivity increase, especially when it requires no retraining of employees.
Itās not like the company I worked with was the only transport and logistics company at the time.
They werenāt the only company that saw the opportunity presented by globalization and wanted to get in on the action.
Yet somehow, they became one of the first companies to see the opportunities presented by technology infrastructure and then actually do something about it.
They were literally years ahead of their competitors, some of whom had far more resources and a far greater capital expenditure budget.
How?
Why?
The principle of Actionable Outcomes Bias - a focus on real, immediate outcomes - is what separates leaders from laggards in tech today.
And thatās exactly where OpenAI is starting to slip.
What Iāve come to realize over the years is that all companies take actions to move forward. Itās inevitable.
However, a few of these companies bias their actions towards actionable outcomes. I call it: The Actionable Outcomes Bias.
Companies that nurture and cultivate this powerful bias have some distinct characteristics.
They focus their thoughts, efforts, and press releases on actionable, tangible outcomes.
They say and share things like, āYou can do action x with our product.ā
When they announce partnerships, itās always in the context of how this partnership improves the tangible outcomes for their business.
They demonstrate how the partnership improves their productivity or somehow improves the value of their product to their customer base.
Internally, their teams operate in a psychologically safe environment. Individuals are expected to provide and receive honest, robust feedback.
Ideas are evaluated, accepted (or rejected) based on the merit of their business impact, not squashed because it doesnāt suit a leader's internal political agenda.
Above all, at all costs, these companies avoid the merest hint of tokenistic virtue signalling that lacks any kind of substance.
Why? Because they know that the moment they take their eyes off a relentless focus on business improvement, they make space for the competition.
The companies that do this well? They change the world.
Those who don't? They disappear.
Just like OpenAI is starting to do.
I couldnāt help bearing this opinion in mind as I researched this week's news in preparation for this newsletter.
Last week, I briefly mentioned Google Geminiās foray into AI for Education.
I didnāt have time in that newsletter to expand on what I learned; however, I walked away from my research seriously impressed. It looks like a genuine attempt to unleash the enormous power of generative AI in the education space. Working through their website, theyāve clearly thought about how this product could improve outcomes for students, parents, educators, and the systems that support each.
This week, I also came across Claude's (Anthropic) foray into AI for Education.
I also walked away from my research seriously impressed, and probably even more excited. Claudeās approach is to partner with high-quality information providers (such as Wiley) to ensure that people researching knowledge have access to high-quality, glue-on-pizza-free answers. Combined with the exceptional power of Generative AI, itās a powerful take indeed.
So far, so good.
We donāt yet know which of these approaches will work the best, and maybe it will be a mashup of both.
Maybe both can coexist peacefully.
What is clear, though, is that both companies are implementing the Actionable Outcomes Bias.
The press releases, demonstrations, and generable hubbub around both companies demonstrate a clear value proposition for the perceived target market.
They emphasize how this will help, why itās important, and most importantly, what their product can ACTUALLY DO.
Both have a clearly opinionated and articulated path forward for their products.
Then you have OpenAI.
Now, I love OpenAI.
I think what they did with ChatGPT is amazing. However, hereās their contribution to AI-Education: āWorking with 400,000 teachers to shape the future of AI in schools.ā It sounds impressive until you read the article.
There you learn that over the next five years they are working to establish a flagship facility and launch education hubs, blah blah blah.
Five years. Weāre not even two years into the Age of AI, and everything has changed. Five years from now is 2.5 times longer.
And also, have all those smart and intelligent people not even read āthe smartest kids in the world, and how they got that wayā by Amanda Ripley?
OpenAI has an incredible first mover advantage, and thatās what they wanted to announce?
Crazy.
Can you imagine all the random board meetings, executive meetings, conferences, and brainstorming sessions theyāre about to go through?
To achieve absolutely nothing?
To me, this sounds like a company that has made a conscious decision to move away from the Actionable Outcomes Bias, and towards a future filled with tokenistic, non-value-adding announcements that meanā¦nothing.
Hopefully, they will course correct soon.
The best prompts arenāt the ones that get types into ChatGPT.
These reflective questions are crucial for you as a CEO or business leader.
So, reflect on what youāve read here, starting with these questions:
Calendly
Because AI is so complicated with so many terms, I created a quick reference glossary you can research here.