Be a consultant, not an academic; why building companies when you can copy good ones?; Weyland-Yutani can't do marketing; build business proprioception.
They build rockets, but businesses need bicycles.
A colleague and I automate business processes for local companies, sometimes using LLMs.
Talking to others in the same field, we've noticed a certain "academic" tendency.
Two guys working for a large IT management company told us about their project: an LLM-powered chat connected
to various tools (Asana, Microsoft Teams...).
It took them 6 months.
SIX MONTHS.
Because they built everything from scratch, the chat interface, the MCP connectors, the LLM API
integrations...
EVERYTHING FROM SCRATCH.
We did a similar project in 2 weeks using the technology the company already had (Copilot, Power
Automate, ...).
But of course: employees already struggle with basic tools, if the goal is to optimize processes you
can't do it by COMPLICATING things.
To us it's obvious to first understand business needs, the team's technical skills, stakeholder
interests...
But apparently it's not obvious to everyone (to anyone?).
I call this tendency "academic" not simply because it's common among people from universities, but
because it's the approach that universities teach.
The academics I worked with were top experts in one tiny niche of knowledge, and completely incompetent
at everything else (even in applied mathematics, the field I worked in: I met the worst programmers...
despite coding being 90% of the job).
But it's not just incompetence: it's impermeability.
They refuse to contaminate their expertise with a surface-level understanding of anything else.
In this case, the "anything else" would be "business": understanding how the companies actually operate,
identifying their needs, and making decisions based on that.
And in academia, anything involving companies and money is considered the root of all evil.
I must say that on the business side things aren't much better: managers are completely clueless about AI.
But there's a fundamental difference: corporate incompetence on AI is a necessity, while the
academic has turned impermeability to the outside world into an identity.
A company manager simply doesn't have the time to keep up with a topic that changes literally every day.
But wait: I'm not angry at the academics.
Quite the opposite!
All the better for us: it's so easy to be better consultants!
Having good ideas is hard. STEALING good ideas is getting easier.
A friend and I were planning a management tool for doctors that would speed up bureaucracy using LLMs.
We gave up because any feature could be replicated in a day of vibecoding by market-leading competitors
(who already have cash flow and a technical team).
My prediction is that this will only accelerate: fewer startup acquisitions, more startups failing because
market leaders replicate their winning features.
We're already seeing it with AI-frontier labs destroying startups by releasing native connectors.
So yes: this is the era of the idea guy.
But sadly: this is the era of the idea guy.
It's just a 3-questions solution.
At 15, playing Aliens vs. Predator, I was struck by the Weyland-Yutani motto, the
corporation that experiments on xenomorphs that end up escaping and wiping out every human on the base.
The slogan, broadcast by overly composed recorded voices, is: "Building better worlds."
I've always found it amusing that such a corporation would choose such a motto.
Obviously I'm not critiquing the marketing choices of a fictional company, but in real life I find that
some companies have made the exact same mistake.
Here's a list of examples.
For each one, I want you to ask yourself:
Only AFTER you've asked yourself these questions, click the link:
I'm guessing that for all three slogans, no image came to mind, they left you with a vague impression, and
NO brand came to mind.
In other words: they have the opposite effect of what a slogan should do.
(In my head, every time I read a slogan like these, I hear it in the Weyland-Yutani voice from the
videogame.)
And as you've seen, these are NOT small companies!
These are massive corporations that produced a generic, meaningless slogan, disconnected from the
product.
These slogans could easily belong to a motel, a car rental agency, or to LEGO, Adobe, an art school...
Lucky for us the fix for a bad slogan is simple: Harry Dry's 3 questions.
If all answers are "no" you've written rubbish, if they're all "yes" you have a valid slogan.
Let's take "Do what you can't":
Meanwhile, slogans that followed the rules have become iconic:
So yeah: don't be a Weyland-Yutani.
If you have to look at your legs to walk: VERY BAD.
I currently work as a consultant for local businesses: I implement automations and LLMs in business
processes.
The typical situation I face is the department head who wants to "use AI because we can surely do it
faster."
Then I analyze the situation and the problem is almost always one of:
One of the cases that struck me most in "The Man Who Mistook His Wife for a Hat" is Christina's
("The Disembodied Lady"): a woman who lost her proprioception due to a neurological condition.
To understand what proprioception is, try this: close your eyes and touch your nose.
How did you do it? You didn't use sight or any other sense: proprioception allowed you to do it.
Christina had lost this ability: she no longer unconsciously knew where her legs were, so she was forced
to look at them in order to walk.
Many companies lack proprioception: that unconscious (organizationally speaking) awareness of what is
actually happening across departments.
I don't mean it in the literal sense, nor in a "non conscious" way.
I mean a sense of the non-obvious, non-explicit dynamics of how the company works.
Not "we need to automate the DB extraction," but: how technical is the person doing it manually? What
kind of solutions can I propose to them? How many people does the solution impact?
In fact, when we go in and analyze the problem up close, we end up discovering it together with
the boss...
The symptom of that is that the person who contacts us is never the person who has the problem.
They don't know the details and become an unnecessary intermediary who complicates things: they must
attend every meeting and must approve the quote, but they slow down the analysis and the decisions.
It's nothing new: in management theory it's called the Iceberg of Ignorance, where the managers are aware of only
a small percentage of the frontline problems.
But now it's worse: they don't just not know the problem, they think they know how to solve everything
(with AI!).
The Iceberg of Ignorance is universally attributed to Yoshida's 1989 paper "Quality Improvement and TQC Management at Calsonic in Japan and Overseas," but no official academic reference actually exists. Basically a "friend of a friend told me..."
I don't know what the solution is, but I know the cause.
In Italy (and in Europe), the "boss" is almost never the founder nor someone whose income depends on the
company's results: they're just another employee.
They're not paid to have business proprioception: they're paid to sign off on salaries, approve budget
allocations, and attend meetings with other top executives.
Working with the self-employed, the dynamics are completely different: they know what they need, they
can explain the tools they use and why, they can articulate what they can and can't do.
Everything is easier, and the solution comes faster and works better.
Classic principal-agent problem, AI-FOMO-flavored.