How to Think About A.I. in 2025
Perhaps we should start the new year with a fresh perspective on this topic…
Happy New Year!
As I was sitting down to write this piece I remembered the saying, “meet the new boss, same as the old boss.” I’d heard it many years ago and had always assumed it was an old Soviet saying. I’d heard a resigned Soviet captain say it once in a movie and it always stuck with me as the quintessential statement of situational surrender. This quote is also often credited to George Orwell. I went and looked it up and was surprised to learn it was actually coined by the British musician Pete Townsend of The Who when he wrote it as a lyric in the song, “Won’t Get Fooled Again” (1971) . It reflects a sense of resignation when one comes to realize the change at hand isn’t necessarily the change one wanted or envisioned. The words themselves are a sign off at the very end.
“I'll tip my hat to the new Constitution
Take a bow for the new revolution
Smile and grin at the change all around
Pick up my guitar and play
Just like yesterday
Then I'll get on my knees and pray
We don't get fooled again
Don't get fooled again, no, no
Yeah
Meet the new boss
Same as the old boss”
It may seem like a stretch, but to me, these lyrics and their sign off at the end remind me of the current conversation around A.I.. I’ve been watching the steady stream of news articles and predictions and they perfectly reflect the idea above. The current conversation around A.I. continues to be a shifting, yet utterly-predictable mix of hyperbole and techno-babble with a small side portion of helpful information. While the headlines may shift, they don’t really change. As a result you, the participant, don’t really benefit from reading them. In my mind that’s the biggest problem.
For example, MIT Technology Review ended the year with its A.I. Hype Index. It’s actually pretty clever but the contents are really just the same old conversation, dressed up from a different vantage point - hype vs. reality, doom vs. utopia (see image below). MTR is one of the most respected technology innovation publications so the fact it has embraced the hyperbole around A.I. as a topic is interesting. But is their approach really change? Is this really any different than what everyone’s been writing about for the last few years? Meet the new boss, same as the old boss.
If you step back and shift your perspective ever so slightly, you see a sharper picture comes into focus. We, as readers and interested parties, have largely been sidelined from the real value of the story based in large part to a pernicious feedback loop driven by the technology’s immature nature, it’s infinite complexity and the age old story line of greed and avarice (on the part of both the writers and the hypesters). And the headlines leading into the new year didn’t seem any different. Meet the new boss, same as the old boss.
I was so struck by this realization I sat down and looked at the various headlines throughout all of 2024 and they almost uniformly fell into one of three buckets:
Bucket #1: Technology Developments and Milestones
Example: “OpenAI Releases ChatGPT o1, ‘World’s Smartest Language Model”
Did you know? There were 19 LLM product announcements last year (as best I could count). This averages about one every three weeks. This included Stable LM 2, Grok 1, Llama 3.1, ChatGPT 4o and others. Each month, like clockwork, readers were treated to a flurry of articles talking about how said model improved some feature or how it beat a benchmark. But in reality the average user experience for applications like Claude, ChatGPT and others didn’t appreciably change. For ChatGPT users things actually went backward with many of the supposed enhancements being hidden away in its “Pro” subscriber level which runs around $200 a month. It’s a given that these companies inevitably release updates. What’s not given is how the news will appreciably improve your knowledge of the category or your ability to more effectively discuss the topic. I love the new Perplexity pages function but that’s one I found on my own through clicking around.
Bucket #2: A.I. and Business Competition
Example: “AI Stocks: Tech Giants, Cloud Titans Face 'Show Me' Moment. Nvidia Mojo Gone?”
This headline is largely emblematic of the broader industry reporting talking about topics ranging from Google’s inevitable demise (and now it’s inevitable resurgence after its quantum computing announcement) all the way to how A.I. is evolving faster than any technology before it and how it is on the cusp of disrupting your industry as well.
We get it, the A.I. business landscape is highly competitive. And it stands to reason it will remain so this year. What’s lacking in so many of these articles is the absence of any real competitive insight. Everyone seems to understand the category is hot. On the other hand no-one seems to know whether it will stay hot and if so, why it will stay hot or if not, what will influence said downturn. This is 2024 all over again… Will NVIDIA climb? Will it cool?
Meet the new boss, same as the old boss…
Bucket #3: A.I. and Our Dystopian Future
Example: “Shazam!' star Zachary Levi compares AI to biblical disaster”
It really pained me to use a Fox News reference but I just couldn’t resist. This headline fits nicely with so many other apocalyptic pronouncements that make their way into news feeds on a daily basis. Case in point, no sooner did I see that earlier headline announcing ChatGPT o1 than these headlines popped up”
“OpenAI’s o1 model sure tries to deceive humans a lot”
“OpenAI's new ChatGPT o1 model will try to escape if it thinks it'll be shut down — then lies about it”
“The new follow-up to ChatGPT is scarily good at deception”
Be Your Own A.I. Boss in 2025
Meet the new boss… you. I sincerely mean this. As I sat back and thought about what was so frustrating about all of this I kept rounding back to what a waste of time it all is given the importance of the topic. I can’t tell you how many highly-intelligent people I’ve spoken to about A.I. feel largely relegated to spectator status due almost entirely to the dysfunctional category conversation I outlined above. As I mentioned earlier, the current conversation around A.I. largely eliminates any real benefit to interested parties or curious intenders. So, in the vein of “be the change you want to see…” here are four categories you can focus on yourself to avoid the predictable cycle that’s already upon us. You can reframe your participation through these lenses and avoid falling into one of the predictable buckets.
#1: What part of A.I. am I actually interested in?
What’s always struck me is how people talk about A.I. as if it’s this giant, monolithic thing. A.I. is an incredibly complex topic. A.I. is a catch-all term that includes LLMs but also Machine Learning, Computer Vision, Robotics and other applications. It includes tech, data and innovation. But it also includes ethics, design, training, commerce, vertical market considerations and that’s just what I can think of off the top of my head. Finding a specific area of interest will allow you to not only source more relevant and useful information, it will also allow you to enjoy the topic.
Be your own A.I. boss in 25: Find a part of A.I. you’re passionate about and go deep on it. You may even carve out a role as a subject matter expert.
#2: What problem do I need to solve?
It’s important to remember that A.I. (and it’s various components) is a means to an end. It’s an incredibly interesting technology which can lead many to suffer from “shiny object syndrome.” But at the end of the day A.I. is about teaching machines to think and act like humans. So once the shine wears off many people are left wondering what they would actually do with any of the tools or assistants they’ve been exposed to.
Before you try and decide which application you like the best, ask yourself, what could A.I. do to make my life easier? A.I. will serve one of two functional roles for you. It will either make you better at something that you do or it will automate a task or process for you. All of the data and applications and systems really come down to one of those two benefits. If you are a writer it will either make you a better writer or write something for you (though I’d argue if you want to be a writer having an LLM write it for you is pretty lazy).
This challenge extends from the individual to the personal level. You are just as able to ask how you can enhance your own personal productivity as you are to build a use case for your team, division or company. The key is avoiding the predictable buckets and starting your inquiry with the question of, “how can I solve this problem.” You will be surprised what you find along the way to answering that very question.
Be your own A.I. boss in 25: Brainstorm 2-3 areas in either your personal or work lives you would kill to make easier, faster, simpler. Is there something you’d just give away to an assistant if only you could? These are all great use cases for A.I. and you will learn a lot in the process.
#3 What can I learn?
While this may seem elementary, you’d be surprised at how few people are actually taking advantage of the various courses and certifications freely available online. When I’ve asked why that is the almost universal response has to do with the perceived technical nature of the topic. A.I.’s personal brand is both awe inspiring and intimidating for many. But nothing could be further from the truth. Sure, many advanced A.I. topics like Deep Learning can get complicated quickly. But much of today’s commercial A.I. environment requires no coding or math skills at all. In fact if you listen to some of the Dystopian Bucket #3 authors coders and mathematicians may now be an endangered species.
Last year I received a certification and a multi-part specialization on Coursera. One was from Vanderbilt focused on prompt engineering and the other was a multi-part certification on A.I. and Business from the Wharton School. I have rudimentary technical and math skills. Trust me, if I can do it, so can you. And if paid courses aren’t your jam there are plenty of free courses. In fact there are YouTube videos that explain some of the courses so you don’t actually have to take them.
IBM offers a free, self-paced course.
DeepLearning.ai offers a free for everyone course with a beginner’s level track.
You can either take the Coursera Google Prompt Engineering Course or watch Tina Huang summarize it for you in 20 minutes.
Be your own A.I. boss in 25: Take a course. Or two, or there. There are many free courses out there if you’re worried about time or money.
#4 What deficit can I address?
This one is a little more advanced though, still, it doens’t require a PhD or equivalent tech wizardy skills from Stanford. Many organizations are racing to embrace A.I. and as they do, they are coming to realize they face significant challenges in one of three areas:
Resources - One example would be data. LLMs and Machine Learning applications require substantial amounts of data to be effective and that is typically left out of the sales pitch when a salesperson is pitching their new tool. Data is just one example of a resource challenge. If you dig a little you will likely find many more A.I. related resource challenges.
Talent - A.I. applications often require skill set updates. Per IBM.com, “According to a 2024 Randstad survey (link resides outside ibm.com)4, respondents said that companies adopting AI have been lagging in training or upskilling employees on how to use AI in their jobs. There are also gender and age divides in how well AI training adequately prepares workers.” You could be the hero that helps solve this issue.
Processes - Given A.I. is still a work in progress for almost everyone few teams or organizations have well-defined processes for workflow or even basic governance. The United Nations has been very vocal on this topic as have many others. A.I. needs strong guidance at almost every level. As you learn about A.I. and the various applications, you have the opportunity to step back and assess what process deficits exist at different levels and either lead or contribute to the solution.
Be your own A.I. boss in 25: Don’t assume everything is “A Ok” when it comes to A.I. in your work world. There are likely numerous deficits you could shed light on and help address. This will lead to growth along the way.
Meet the new Boss in ‘25: You
Meet the new boss in 2025 - yourself. By focusing on your specific interests, identifying problems to solve, pursuing learning opportunities, and addressing deficits in your organization, you can cut through the noise and hype surrounding A.I. This approach empowers you to become an active participant rather than a passive spectator in the A.I. revolution. Remember, the most valuable insights often come from your own experiences and efforts, not from sensationalized headlines or recycled narratives.