Technology

Power is no substitute for Skill

So many interesting things are happening that it can be difficult to know where to start. When so much is capable of grabbing our attention, focusing becomes a vital discipline.

So it is with New Artisans. At the heart of it are ideas on what craft might look like in an age dominated by technology, and what it could mean for those looking to master something that lasts, rather than run on someone else’s mill wheel, trying to survive.

AI has become an ouroboros, constantly eating its own tail, as new claims swallow old claims, with almost fantastical tales of capability in pursuit of investment dollars. Everywhere, there are those claiming expertise in something we barely understand, and it reminds me of the definition of an expert as someone who knows 5% more than we do.

Nearly every online tool we use now has AI for parsing information, writing notes, and researching. It’s becoming difficult to find places online where we can avoid it; it is becoming a default. Little wonder we feel unsettled, having spent all that time learning to do things that a plugin will now do for us (sometimes, whether we like it or not) or, more alarmingly, those who employ us.

The effect is insidious. Large organisations begin to look like hot air balloons, kept airborne by stories they tell about the benefits of AI rather than propane tanks. Looking down while realising the fragility of what keeps them airborne makes them a bit precarious, and as someone who has been in a forced landing in a balloon, I can testify that coming back down to earth suddenly is less than elegant and dignified.

It is sobering to read that increasingly credible questions are arising about the nature and effectiveness of AI reasoning models. I read of Apple Engineers, Chinese Universities, and others finding gaps, seeing these reasoning models more as pattern matches, as “stochastic parrots” rather than truly curious.

So I did the obvious, and asked ChatGPT to go all whistleblower.

Recent research from Apple, leading Chinese institutions, and others has exposed critical limitations in the reasoning abilities of large language models (LLMs). While these models excel at generating fluent language, studies show they rely heavily on pattern matching rather than true logical reasoning. They often fail simple tasks involving time, spatial understanding, and irrelevant details, revealing a lack of genuine comprehension. Described by some as “stochastic parrots,” LLMs tend to produce plausible-sounding answers without real understanding, raising important questions about their reliability in roles requiring consistent, deep reasoning.

We will see. None of this diminishes the potential LLMs have to contribute, but it does raise the question of the wisdom of a headlong rush to defer to AI.

In my youth, amongst other things, I was a keen amateur rally driver, and one of the lessons I remember was being told there was no point in getting a more powerful car until I could drive the one I had to the maximum of its capabilities, because “Power is no substitute for skill”

That resonates when very clever people are building ever more powerful but unproven platforms, giving us enormously powerful tools that we have not yet learned to use effectively.

Running low on Propane in a hot air balloon is a salutary experience. You have limited control of altitude, and no control of direction - the wind takes care of that. You can only choose the best place to land before you run out altogether. That’s when you find out how good the pilot is.

Many businesses feel like that. If we listen carefully, I suspect we can hear the digital propane sputtering. Everyone is looking at the pilot, and when it comes to AI, there are no skilled pilots. We are all still learning how to fly it.

This means that the real power of LLMs lies less in retrofitting old organisations created before they were, than in helping new forms of organisation to get airborne—ones that are simpler, lighter, and smaller.

It is much easier to design with AI in mind at the beginning than to go through the pain of designing out of existing businesses elements (people and relationships) that technology means we no longer need.

Most of those in big organisations are not qualified to fly an AI-enabled business.

They cannot integrate AI into their business plans. AI's capabilities and reaction time make business plans obsolete; it moves faster than plans can keep up with.

Which brings me back to Artisans, and specifically, their creativity.

I suspect creativity is the new skill shortage. Less in the traditional sense of “creatives”, but more in the sense of those who can find new ways to do things that move us forward, not just deeper into where we are.

Imagination and Creativity are needed when we do not know what to do and time feels upon us.

This TED talk shines a tiny light into how new businesses might take form:

We need those who can start where technology stops, when we don’t know what to do.

Maybe we can start there on our regular network weekly call tomorrow

If you are not a member of the Outside the Walls Group and would like to join us, drop me a line, and I’ll send you a link. Our next conversation is tomorrow (Wednesday 14th May) evening 5:00 pm UK time.

There is no charge - we want your voice, not your money