AI & Machine Learning

The AI Human Turbocharger: A Human & AI Model of Acceleration

I find this metaphor an interesting one regarding the application of “AI” technology to our businesses. A Turbocharger intakes hot air from the exhaust of an already running engine, cools, compresses, and pushes it back into the motor. This creates higher compression ratios and by extension creates greater horsepower and torque. In other words, a faster accelerating and moving vehicle.

A Turbocharger, by itself, is completely pointless. You essentially have an air compressor without any air to compress.

The vehicle is your business, the humans the motors and AI the turbocharger. The driver is your C-Suite, providing vision, obstacle avoidance, direction and when to accelerate or brake.

It's a similar situation when considering AI within your business. An AI does not replace a human. It does not reason or create. It does parse input and data to arrange it into meaningful sets. It does find patterns in information extremely quickly, creating summaries that human minds might miss. Yet, an AI operating in a vacuum is essentially pointless.

In order to operate, an AI needs a human mind to pass it information to process. That information is often creative, directive, corrective, and preorganized. In a word: reasoned.

Leadership that is looking to reduce their staff because they think they can go as fast or faster with a few AI implementations are misunderstanding the nature of the math. By reducing their head count, they are effectively shrinking the size of their motor, hoping the turbocharger will make up for the loss. While that may be true, they are making a drastic, fundamental change to their organization that will increase wear and tear of the engine that remains.

The turbo(AI) only works with input generated from the motor(your employees).

Your business only operates because of the reasoning, creative abilities of your staff. Imagine: rather than replacing staff you focused on how AI could eliminate the busy work they do day to day, what value you might generate if you focused on automation of drudgery and boilerplate work. Let humans do human work: vision, direction, creativity, insight, intuition, and asking questions!

This is a reflection on discussion and content from my #MITProfessionalEducation #CTO #CPO #Innovation #Leadership coursework. We have spent quite a bit of time discussing the application of #AI. While understanding the theory of our discussions, I wanted to translate the theory to a practical, easy to grasp model in business that speaks directly to the shifts happening across the markets. You cannot replace reasoning humans with non-reasoning algorithms, no matter how cleverly they arrange information. You CAN effectively accelerate your operations by combining AI abilities with human creativity and intuition!

Another day and another few thousand AI-written posts and AI-scripted reels.

The Dead Internet is quickly becoming the AI-Swamp-of-Sameness across websites, social media, and video platforms. Everything reads the same, sounds the same, because, well, it’s all generated from the same underlying statistical generative chatbots. No, wait, not chatbots, AI.

I wish we could rewind a little bit and whisper into the ear of a particular marketer that calling these systems AI would engender a sweeping disassociation of smart people from their own voices and critical thinking.

How much content is actually written by humans now? What percentage would you wager is considered, conceived, and written by human hands?

Now, don’t get me wrong, I have used AI in various ways in any given week. I’ve tinkered with producing music on Suno, have built mocks and templates, written some software to parse and format heavy equipment specifications, and built images for Product Vision slide decks.

BUT there are a few conclusions that I’ve come to from this experimentation and usage… and it’s that we are giving up the very thing that makes humanity a miracle.

Our brains are capable of producing lyrics and music, software that create interactive 3D worlds or generate natural language, we write novels and comics, plays and movies, engineer airplanes, cars, send people to the moon, and the computer I’m writing this with.

What are we doing with these AI’s?

We are losing our voice, I think.

Every day I scan LinkedIn and see post after post with the same voice.

Occasionally I scan the music produced in Suno and it all sounds generically the same.

I scroll reels and see one AI-generated video after another, all with that….weird lighting, coloration, and movement that isn’t right.

Let's take a breath and think about what we are doing. These sycophantic AI’s are bringing out a weird side in humanity and I suspect specifically in leadership. These things are not reasoning at all.

They are tools.

And should be used as such! There is potential in the measured, thoughtful application of a tool. But any craftsman will tell you that an untrained enthusiast with a band saw and some wood is likely to injure themself.

Is that what we’re seeing now? Companies that have driven full-speed into the AI swamp discovering they’ve just cut off their fingers and out-sourced their minds to statistical models?

How about:

• If you only have time to write a post prompt, then don’t post.
• If you can’t write the algorithm yourself, then don't vibe-code and commit it.
• If you can't discuss a decision, then don't do it.

Don’t lose your voice and your critical thinking. Don’t out-source yourself! Do the work to exercise your mind: write those emails. Build that software. Don’t lose your abilities!

Accelerate your mind with AI, but don’t replace what makes you - us - unique in the world.

#inmyownwords #ai #useyourvoice

Where do we apply AI?

Last week I posted about using your voice and making sure to write in your words, not an AI’s, and to remember that the “AI” are really statistical models, not reasoning or thinking things.

They’re also stupidly expensive. In CPU, memory, and cooling, they are not cheaper than humans. Also consider the down-stream impact companies and homes are facing due to the inflation of CPU and Memory expenses.

❗Everyone is paying for the rush to build data centers.

So if I’m saying we don’t use AI for simple things like writing emails, outlines (really, consider the expense of asking an AI to draft an email rather than typing one out in your own words), or complex vibe-coding, what am I saying that we use these tools for?

Generally speaking, I think the formula is a function of:

Value = TimeSaved / (Size Of Prompt * Complication of Work * RiskOfHallucination)

Obviously, there are far more details and nuance in these variables than what’s coverable in a single post, but the gist is: “how much time am I saving and how complicated is this prompt to process?”

🤨 Also: “am I being lazy or do I really need to spend this money writing this email, post, etc?”

More If-this, then-that, do-this, but-don’t-do-that types of logic require more memory and CPU to process. Stacked complications in logic increase the risk of faulty outputs - an element I think people are often overlooking in their work!

Recall: these are statistical models. The more specific your needs, the fewer related data points exist in their models and you will see an increasing amount of chaff or unrelated additions. This is also why there are programming, image, music, and video models. They are trained on data sets specific to that topic - but what you don’t know is how much of that training model is actually related to the type of work you are asking it to do.

High quality and low (relative!) CPU work is generally found in:

• Stacked simple instructions for processing a dataset looking for patterns, outliers, and trends.
• Basic, self-contained programmatic tasks that are common boilerplate types of work. Think: “I need a function that takes dataset Y and converts it to Z” or “I need an endpoint in API X that receives a payload in format F and converts into a format/datatype D. Name the endpoint N” etc etc.
• “Here’s my post, please check it for grammar and tense inconsistencies.”
• “Provide a series of sources for and against: X.”
• “I need a basic HTML and CSS page that displays X,Y,Z data”

Again: Think TURBO CHARGER! Not “replace my engine”! There should always be a human partner for this work checking the results. 1:1 ratio. Not 1 human to 20 AIs each spawning agents.

Related:

Ford Rehires Engineers fired for AI after quality concerns 6/14/26: https://www.bbc.com/news/articles/cgrkd41n2v9o

NVidia on AI vs. Human Expense 6/14/26: https://fortune.com/article/why-is-the-cost-of-ai-higher-than-human-workers-nvidia-executive/

#ai #inmyownwords #turbochargerai