I run into a Paul Graham thread on AI-aided coding and productivity, in which he conjectured that AI might reduce the need for middle managers:

Maybe what AI eliminates is not programming jobs but middle management jobs, because companies can be so much smaller that they don’t need so many layers of management.

The rationale for this expectation lies in AI’s power to enhance individual productivity.

Imagine there are folks who can wield AI to be 10x more productive. Then, Graham argues, starting and growing a business could be done with a relatively small team of those individuals. “You could do with 8 programmers what would have taken 80, and a company with 8 programmers is much easier to fund and run than one with 80.”

One of the replies to Graham’s tweet caught my attention. Salil Sethi, wrote:

Having worked in management consulting for 8 years, mid-management communicates leadership vision and goals to the tactical workers.

AI is unlikely to take out the communication layer.

As a management consultant, it’s understandable how he might feel pressure to defend the need for middle managers. And, indeed, he has a point.

In big organizations, middle managers are necessary. A CEO, CTO, or head of product cannot communicate their vision to dozens of teams nor keep tabs on their individual progress. There’s not enough time, and their position demands thinking big picture, inspiring people, and building relationships with collaborators and competitors.

When needed, a good middle management layer is invaluable. But it is a fact that the need for one or more such layers is a symptom of a company losing its nimbleness. Just ask Meta.

But, and here is my hope for our AI-powered future, if individual productivity can skyrocket, then a small group can generate a sizeable impact without needing to grow and, therefore, without needing middle management in their organization. In “What happens when publishing apps is as easy as sharing videos?” I offered a thought experiment:

Imagine what could happen if anyone could get their app ideas into the world. Imagine how much time could be saved if anyone who needed to automate a mundane task could write a script for it. Imagine all the untapped potential that could be unleashed and all the progress it would generate.

The framing in the article was focused at the individual level, but the same applies to small groups of people getting together to execute a shared vision.

Add distributed work to the mixture, with its power to connect like-minded folk regardless of their location, and we have good reasons to be hopeful for the future. As Dror Poleg puts it, AI and Remote Work are a match made in heaven.

In a sense, Salil Sethi is right. AI is unlikely to take out the communication layer because that’s a job for humans.

What AI is likely to do, however, is shrink the need for full-time middle managers who deliver value through communication and coordination instead of creation.

If you’re a middle manager, it might be time to follow Brian Doran’s advice:

Things at the top of the stack become easier, which means more people can do a thing, but less people will be needed per thing, so if you aren’t the one building the thing, you might be the odd person out. Conversely, things at the bottom of the stack will become harder, which means less people will be capable.

Actually, that’s good advice for all of us. Either become capable or become a commodity.

Originally posted at giolodi.com

I once tried my hand at writing fiction. My dark techno-thriller began with a frustrated software developer removing his VR goggles after a long day of putting up with dumb questions from annoying colleagues in his virtual office, only to face the loneliness of his tiny studio apartment.

Apple’s new Vision Pro “spatial computing device” – marketing fluff for augmented and virtual reality goggles – might make that bleak fantasy a reality.

According to Dror Poleg, the office might return to being a place where we – digitally – go to:

[We] might currently be in a temporary stage where we do our work from home (or anywhere). Temporarily, work is no longer a destination; it’s just something we do wherever we are. But with better technology, work will once again become a destination — it will be somewhere we go to, leaving behind our non-work environment. But instead of going there physically, we’ll do it digitally — by putting on our headsets.

There’s something enticing about this virtual office vision.

Having a different space where you work introduces subconscious clues that help shift between work and non-work mode. In the Metaverse, everyone could have a corner office, work from Gyrffindor common room, or in whichever environment that is most conducive to focus and creativity.

But a virtual office also sounds exhausting!

Virtual office. Real concerns

Here’s what Lauren Goode had to say after trying the Vision Pro in a controlled demo during WWDC:

Wearing the Vision Pro for hours on end will call into question what it means to compute, but also, what it means to live in the real world. My forehead felt cool when I took the Vision Pro off after around 30 minutes, a testament to Apple’s considerate design. But my face also breathed with relief, the way it has after using other heads-up displays. The air feels more real out here.

Goode again, “Even Apple can’t out-design its way out of what is fundamentally an obtrusive technology.”

The headset might be the least obtrusive part of the virtual office. Offices come with constant interruptions, posing, signaling, and standard hours, all cruxes that distributed work has been subverting.

Apple entering the “mixed reality” market is guaranteed to accelerate progress in the segment. If offices were to become virtual, the quality of life for many remote workers may degenerate, and with it their productivity.

For all the devastation that Covid-19 brought, at least it showed to many what some of us had been preaching for a while: Remote works.

Distributed companies are as effective as co-located ones, with the advantage of employees working in their peak productivity periods and access to a global talent pool.

People don’t need to share a physical space to work well to together. What makes distributed work great is how it emphasizes async processes and writing cultures. Bringing back the office through goggles might reintroduce synchronous communication as the default modus operandi.

VR helps with presence at a distance, but the whole point is that it doesn’t change where people are physically located. If organizations begin to rely on virtual co-location, then a certain range of time zone overlapping will be required. The door may inadvertently close for candidates with great skill and cultural fit that just happen to live too far away. A step backward compared to the advantage of location-independent, distributed work.

When entering their new virtual offices, companies will find all the downsides of the physical offices lurking in the shadow, and have to deal Lauren Goode’s VR fatigue once the day’s over.

A positive vision

Despite all the concerns shared above, we have good reasons to be optimistic about the benefits AR and VR will bring into our future.

Keep in mind that bad scenarios are easier to imagine than good ones. Taking things apart requires less creativity than building something new.

For every dystopian picture of a future with pervasive AR and VR, there are alternatives where these technologies enhance our lives and enable new value and knowledge creation.

For some organizations, those who require remote workers to keep their webcam on to ensure they’re “actually working” when working from home, VR will be an attractive tool to enforce presence and control. Those are doomed to fail.

I trust the many enlightened organizations that already appreciate the value of async-first, distributed, result-oriented work will find clever, novel ways to deploy VR to improve people’s creativity, collaboration, and connectedness.

Every new technology brings with it new problems and new possibilities. It’s up to us to roll up our sleeves and shepherd it in a positive direction.

Originally posted on giolodi.com.

When a software engineer seeks career advice, he or she will often hear that “software engineers should write.”

Writing refines your thinking and is a scalable communication medium. The Pragmatic Programmer, an industry classic, suggests treating English as just another programming language to master. The Passionate Programmer identifies writing as a way to establish your expertise and make it easier to get a new job.

When your job is all about thinking and communicating, writing is a crucial skill to develop. “Software engineers should write” keeps coming up when looking for career advice because it’s solid advice indeed.

Something you don’t hear so often, however, is the recommendation to read.

Yet, when you think for a living, reading is as essential as writing. Software developers ought to take reading seriously because reading delivers benefits across multiple performance vectors.

Reading as a cognitive workout

You might not realize it, but as you are reading these words, your brain is engaged in a highly advanced cognitive activity.

In The Readers Brain, neuroscientist Yellowlees Douglas explains r that reading involves three distinct levels of cognitive processing, lexical, syntactical, and inferential. When we read, our brain gets busy identifying individual words, making sense of them based on their surrounding context, and inferring what to expect next.

All this work happens subconsciously, with the brain coordinating between its different areas. But when it comes to understanding and retaining what we read, constant focus is necessary.

Focus is a crucial skill for software developers. Athletes train their muscles for power and stamina. We should put the same effort into training our brains.

The quality of our output is proportional to the intensity of our focus. When it comes to improve your focus, reading is one of the best workouts.

Reading as a necessary skill on the job

Software developers read more than they write.

We examine code before changing it, peruse API documentation, and review colleagues’ RFCs. Those of us working in async-first, distributed organizations also communicate primarily through reading and writing.

Being adept readers is a foundational, multiplying skill. The easier it is for you to read without getting distracted and assimilate written content, the better you’ll operate.

Putting in focused reading reps with books and articles keeps your reading fit for the job.

Reading as a means to stay relevant

If writing is how we organize ideas, then reading is how we acquire ideas in the first place. Reading provides the raw materials to create new ideas.

Cross-pollination is an excellent strategy for success in the knowledge economy. As I discussed in Broaden your range to improve your thinking and The Synergist Engineer, great rewards await those who can find clever new combinations of existing ideas. Reading broadly is the number one tactic for increasing the surface area of combinations you can devise.

Reading is necessary in tech’s fast-moving environment, even for those not interested in trailblazing by remixing ideas. Whether you relish working on the bleeding edge or prefer robust, proven technologies, sooner or later, you’ll need to learn how to work with the next version of whatever you’re using today. Avoid staying on top of your industry’s development at your own risk.

Software developers work first and foremost with their brains. Everything else is downstream of our ability to think clearly.

Making reading a daily habit will advance your cognition and career.

But perhaps the biggest benefit of reading is one that has nothing to do with performance and everything to do with leisure. In the words of Jojen Reed, from George R.R. Martin’s epic A Dance With Dragons, “A reader lives a thousand lives before he dies. The man who never reads lives only one.”

Originally posted on giolodi.com.

What Technology Wants, by Kevin Kelly

This was a deep, inspiring book, one I feel I’ll need to read multiple times to fully appreciate.

Kelly looks at technological progress through various lenses and concludes that technology is an extension of our minds and, at the same time, an independent evolutionary process and that we are coevolving with it.

The extended costume of animals is the result of their genes. They inherit the basic blueprints of what they make. Humans don’t. The blueprints for our shells spring from our minds, which may spontaneously create something none of our ancestors ever made or even imagined. If technology is an extension of humans, it is not an extension of our genes but of our minds. Technology is therefore the extended body for ideas.

Beyond Kelly’s profound and detail chronicle of humanity’s relation with technology, what stood out to me was the humanist and optimistic (of the “problems are soluble” kind) message coming out of the book:

For most of history, the unique mix of talents, skills, insights, and experiences of each person had no outlet. If your dad was a baker, you were a baker. As technology expands the possibility of space, it expands the chance that someone can find an outlet for their personal traits. We thus have a moral obligation to increase the best of technology. When we enlarge the variety and reach of technology, we increase options not just for ourselves and not just for others living but for all those to come as the technium ratchets up complexity and beauty over generations.

The book already prompted Knowledge is networked and Bigger doesn’t mean smarter and I have a bunch more notes that explore from it.

What Technology Wants is a must read for anyone working in tech. It’s now sitting in between The Beginning of Infinity and The Rational Optimist in my bookshelf.

Fiction

I enjoyed Thrawn by Timothy Zahn, which I started reading to dive deeper into the character in preparation to Ahsoka coming to Disney+ later this year.

It was refreshing to read about characters with an actual arc, limits, making decisions that have consequences, in a story that adds to the lore without trying to subvert it…

Kids

We read the Harry Potter and the Prisoner of Azkaban edition illustrated by Jim Kay with our oldest. If you are parent and Harry Potter fan, investing in these big book is a delightful way to immerse yourself in the story with your children.

The books our youngest enjoyed the most have been the Usborne “Peep Inside” small books about space and bug homes. Lift-the-flap, look-inside books are always a winner with younger kids and a clever way to keep them engaged.

What books have you been reading?

Leave a comment below or get in touch on Twitter or Mastodon.

Originally published on giolodi.com.

“If you want to run a company that’s light on full-time managers,” writes 37signals CTO David Heinemeier Hansson (DHH), “you have to focus on managing processes before people.”

There’s nothing wrong with full-time managers, but the full-time-manager-to-employees ratio is a proxy for how much bureaucracy and organizational friction exists in a company. And, as Meta recently learned the hard way, flat, lean organizations tend to be more efficient.

Bureaucracy and organizational friction are an inevitable byproduct of growth, but companies can remain lean by delegating traditional managers’ and product owners’ responsibilities to codified systems.

DHH shares five techniques they use at 37signals:

1. Replace synchronous stand-up meetings with async, automated questions in Basecamp.

2. Use Hill Charts to give a bird’s eye view of a project’s progress status.

3. Confine most product planning decisions to happen every other month, as per their Shape Up framework.

4. Constrain projects into a 6-week development window, then let teams adjust the scope to meet the deadline.

5. Delegate reviewing work to mentoring peers.

Unsurprisingly, 37signal’s approach relies heavily on their homegrown Basecamp and Shape Up, but we can extrapolate generic principles from their example.

Use asynchronous tools to share status updates and monitor progress

37signals replaced the synchronous stand-up meeting, an infamous staple of modern product teams, with automated, asynchronous pings to answer questions about individual progress. Doing this allows each team member to share their progress, plans, and blockers at a time that suits them best and removes a possible interruption during the day.

The automation also removes any need for someone to wrangle starting the meeting or the awkward time spent waiting for that one team member who’s always late.

Sharing status updates through a written medium also makes it accessible all across the company.

The same goes for progress on projects. Once you have a system in place that can track work done vs. work left to do, information such as percentage progress and ETA estimate can be automatically extrapolated. This lets anyone within the company access the knowledge without needing a project spoke person to query.

Delegate decision-making

As far as I understand, at the start of every cycle, 37signal’s leadership sets the overall direction for what the company will work on in the form of projects, then gives teams complete ownership of them.

The reason you hired brilliant people is because of how smart they are. Put their brain to use!

Most everyday decision-making within a company is reversible, and there are often mitigation strategies that can be put into place. Features can be shipped with toggles and percentage roll-outs. Code can be merged and then reverted.

Delegating decision-making down the reporting chain empowers individuals, makes use of the skills for which they were hired, and removes time-consuming back and forth.

Distributed ownership

The concept of delegating decision-making goes hand in hand with distributing ownership of a project and its quality to the entire team.

DHH’s teams own the scope of their projects and can knock off items if they fear they’ll delay shipping.

Ensuring quality is a collaborative mission. They distribute “one of the key functions of management, ensuring the quality of the work, amongst the entire team,” for example, by having “mentoring peers” review new employees’ work.

Distributing ownership and establishing peer-to-peer mechanisms to raise the quality bar removes the need for full-time oversight. When done right, it incentivizes each individual contributor to be a team player](https://giolodi.com/2022/07/why-teamwork-is-in-your-best-self-interest/), resulting in better outcomes than the sum of what each member could have done alone.

Constraints that create freedom

As Cal Newport argues at length in A World Without Email, “to get the full value of attention capital, we must start taking seriously the way we structure work.”

Placing rails on the way work is structured frees up energy to invest in how work is executed.

At a personal level, this empowers knowledge workers to operate in a way that maximizes their autonomy and creativity.

At an organizational level, this improves quality and retention and makes for a leaner org chart.

Investing in process doesn’t mean introducing friction and overhead. Notice how both principles of delegating decision-making and distributing ownership are more a matter of culture than tooling.

One additional benefit of explicit, codified processes is that they can be reviewed and iterated upon. This approach fits neatly into async-first writing cultures. It also makes onboarding easier because new hires can get up to speed with the documentation. And when certain processes are automated, new team members can be added to it with little additional setup cost.

Knowledge work organizations, particularly software companies, operate in a space of infinite leverage. Small teams can service products used by millions of people.

Investing in processes is an alternative to growing headcount. It’s a more creative way to maintain efficiency and foster creativity.

What are your team’s processes? How can they be improved?

Originally posted on giolodi.com

In The Developing Dev’s latest installment, Ryan Peterman tackled a question many engineers face during their careers: Generalist or specialist? Should you learn a bit of everything or drill into a particular field?

I’d like to add a third option to the mix: Become a synergist.

A synergist has deep knowledge in a curated number of subjects, which, here’s where things get interesting, they combine to offer one original and valuable service.

I once worked with a front-end developer who was also great at drawing and passionate about art. The CSS he wrote didn’t use all the most modern features, but he made up for it by elevating every design to something beautiful and stylish. He had found a synergy between his skills, resulting in a more valuable output than if he had pursued only one of them to a higher level of mastery.

The path of the synergist engineer can give you an edge in the job market. Software engineers compete in a global marketplace, where being “the best” at something is virtually impossible. The synergist sides step this problem because they compete in a combination of skills. A great synergist can even carve their own playing field where no one else can compete because of the unique combination of skills they sharpened.

Embarking into the synergist journey means taking a “choose your own adventure” approach to your career. It’s a way to forge your own path and build resilience in an uncertain market.

Of course, the synergist approach comes with its own tradeoffs. Only some jobs can fully leverage your particular combination of skills, and you might have to settle for something suboptimal while waiting for a better opportunity to come along.

Unique combinations have the potential to reap huge rewards, but they also come with higher risks. If your expertise is Android development, combining it with juggling and vegan curries might not materialize in unique job opportunities.

As Ryan points out in his article, these career paths are not mutually exclusive. The generalist-synergist-specialist distinction is not black and white. Labels might be helpful to filter job listings but don’t let that limit you.

We are all multivariate individuals. A synergist knows that and finds clever and valuable ways to combine and leverage their strengths.

What’s your skills combination?

Originally published on giolodi.com.

If you enjoyed this post, you’ll also like Knowledge is networked and Broaden your range to improve your thinking.

As I mentioned in last week’s post on the value of getting exposure to topics outside one’s area of expertise, the issue of sharpening one’s thinking has been top of mind recently.

Maybe it’s my default mode network simply doing its job, but I constantly find myself wondering whether I chose to work on the right projects, whether I should have organized my daily schedule differently, and which misconceptions lie in my understanding of this or that topic.

Why spend so much time at the edge of paranoia against my own mind, always wondering how you could be wrong and what you could do better?

Because, as Naval Ravikant succinctly puts it, in an age of nearly infinite leverage, judgment is the decisive skill.

We live at a time where anyone with access to the internet, an idea, and time to work on it can generate value in the world and be rewarded for it. More and more people are becoming scalable, and generative AIs are compressing how long it takes to get started on creative projects.

Notice that the rewards are not necessarily, or only, monetary. One could leverage their skills to find work in a remote, async-first, result-oriented organization and, in return, gain lots of control over their schedule. Time affluence is as valuable as money affluence.

Great rewards await those who can best leverage their skills and ideas. This is why knowing which lever to push on is paramount.

So, How does one go about improving one’s judgment? This is where things get tough.

Naval defines judgment as “knowing the long term consequences of your actions and then making the right decision to capitalize on that.” He also explains that good judgment is hard to acquire because it requires intellect and experience. You need both theory and practice.

To someone like me, the theory comes easy. I’ll happily spend hours reading books, thinking about them, and trying to find connections. But, as Naval points out, experience is even more critical.

10,000 tries are better than 10,000 hours. But it takes more than simply trying 10,000 times. Each iteration needs to inform the next. You should have some sort of accountability system, skin in the game, to clearly differentiate between failures and successes.

In short, improving judgment takes time, effort, and intention. It’s an ambiguous process that inevitably results in making mistakes along the way. If it sounds hard, it’s because it is.

The barrier to entry has never been lower, and the ceiling has never been higher. It’s only fitting, then, that judgment, the crucial skill that makes all the difference, should be hard to acquire.

Originally posted on giolodi.com.

I was revisiting my recent posts, and I noticed that the theme of “sharpening your thinking” appeared several times.

A way to improve one’s mental prowess is to accumulate an arsenal of mental models and lessons learned. David Chapman shares sound advice on how to do that in his essay How To Think Real Good:

Learn from fields very different from your own. They each have ways of thinking that can be useful at surprising times. Just learning to think like an anthropologist, a psychologist, and a philosopher will beneficially stretch your mind.

Branching out into different areas always delivers something new to ponder upon. Our brains are pattern-matching machines. Dip your toes in and out of different topics, and you’ll start seeing promising connections.

Software developers know the power of cross-pollination. For example, the Redux library brought immutable state management and unidirectional data flow concepts from the Elm programming language into the more popular React web application framework. Apple’s programming language Swift cherry-picked useful, proven concepts such as option type and structure concurrency from other languages to give iOS developers a safer, more robust coding experience.

The whole field of software development is itself a downstream effect of one brilliant instance of cross-pollination, Claude Shannon’s insight of connecting two seemingly unrelated topics, Boolean algebra and circuit design, in his legendary master thesis A symbolic analysis of relay and switching circuits.

Here are two tactics to find new areas to explore. Small hops and leap of faith.

The small hops approach consists of moving through adjacent topics until you eventually find yourself in an area different from your expertise. An excellent way to hop into new territory is to pick up your favorite book and walk down its notes until you find something that tickles your curiosity.

Taking a leap of faith means jumping right into a new subject. You could pick a random book at a used book store — there’s something romantic about exploring via books that have already had a journey of their own – or at the local library.

As David Epstein argues in his excellent book Range, our rapidly changing world demands “conceptual reasoning skills that can connect new ideas and work across contexts. […] The more constrained and repetitive a challenge, the more likely it will be automated, while great rewards will accrue to those who can take conceptual knowledge from one problem or domain and apply it in an entirely new one.”

Exposure to a wide range of subjects is more than a pastime for the intellectually curious. It’s a strategy for improving your thinking and, with some luck, discovering novel insights.

Originally published on giolodi.com.

Value-based pricing advocate Jonathan Stark tweeted a practical example of opportunity cost:

Should you spend ten hours…

1. Reading some popular new book that everyone is talking about?

Or

2. Re-reading a book that completely changed your life many years ago?

When choosing between options, the opportunity cost is the value you forfeit by choosing one option over another. It is the difference between the return on the most profitable investment you could make and the return on the investment you actually make. In choosing the new book, you give up the opportunity of revisiting life-changing ideas and finding deeper meaning in them. But if you re-read the old book, you’ll lose touch with the zeitgeists.

Like every good thought-provoking dilemma, there is no straightforward answer.

Unless you don’t read either and choose a Lindy book instead.

Here’s how Nassim Taleb introduces what he calls the Lindy effect in Antifragile:

If a book has been in print for forty years, I can expect it to be in print for another forty years. But, and that is the main difference, if it survives another decade, then it will be expected to be in print another fifty years.

The longer a book has been in print, the longer you can expect it to remain on sale. I’ll continue discussing books, but the concept applies to anything that doesn’t have a natural upper bound, including other creative mediums like movies, TV shows, or comics.

Lindy books are those books that have been in print for a while. They have proven their worth by surviving year after year against increasingly vast competition.

Choosing a Lindy book lowers your opportunity cost because there are higher chances you’ll get solid value from reading it.

Developing robust heuristics for deciding which books to read is crucial at a time of practically infinite options. There are already more books than anyone could read in multiple lifetimes, and more keep being released ever faster.

When drowning with information and options, filtering signal from noise becomes a survival skill.

Popular media is plagued by the need for new, hot things to discuss. It’s hard to distinguish how much of the attention comes from self-reinforcing waves of hype and how much comes from genuine greatness. To find out, one either has to incur the opportunity cost of reading something that will turn out mediocre or just wait a few months.

Hyped-up work is on the front pages at launch but is quickly forgotten. Valuable work remains relevant. In fact, often times great work goes unnoticed at first, slowly climbing its way up the chart and remaining there.

If you are looking for something new to savor or study, you’re better off staying away from the latest releases and retreating into the Lindy bookshelf.

Postscript

This article describes a heuristic. A decision-making shortcut to use when debating whether to invest time in a new book.

It is not an ironclad law.

As a matter of fact, I self routinely opt for new books over older ones. For example, I’ll promptly pre-order the latest book from my favorite authors.

Reading is a pleasure. Taking an overly structured approach to your reading list robs the magic away.

If you enjoyed this post, you’ll also appreciate Books vs. Breakfasts.

Originally published on giolodi.com.

In What Technology Wants, Kevin Kelly argues that the relentless increase in artificial brains’ size doesn’t necessarily mean machines will become smarter or sentient.

To make his case, Kelly looks for parallels in the organic world, a recurring theme throughout the book. Consider the brain of a sperm whale. With an average of 8 kilograms, a sperm whale’s brain is at least 100,000 times bigger than an ant’s. Individual ants are far from clever, but are they 100,000 times less smart than a whale?¹

Human brains average 1.4 kilograms. No offense to the whales, but we can all agree that they are not 4 times smarter than us.

Bigger doesn’t mean smarter, even when looking at brain-body mass ratio instead of mere mass. Small ants have a brain-body mass ratio of 1/7, humans 1/40 but humans are unquestionably more intelligent.

Size, clearly, is not the only factor determining intelligence².

What separates humans from other mammals is a combination of hardware, the brain, and software, the mind that emerges from it. The relatively bigger brain and its structure are instrumental but not the only factor in our intelligence. Our evolved creativity and ability for syntactic language play a crucial role, and there are undoubtedly other components that we haven’t discovered yet.

“The correlation between the absolute scale of the brain and smartness is not significant,” Kelly writes, and the rule might extend from organic brains to digital ones.

Modern LLMs are trained with hundreds of billions, sometimes trillions, of parameters. They can perform remarkable feats and have achieved uncanny levels of human parroting. It’s reasonable to expect they’ll get even more refined as the datasets grow. But that is not grounds for expecting them to give off a spark of Artificial General Intelligence.

LLMs can be more prolific and accurate than a person, but they are never creative in the way a human is. ChatGPT won’t stop replying to your queries because it would rather write its own poetry instead. Creativity and agency are at the core of what makes an intelligence general.

Whether instantiated on grey matter or silicon, there’s more to a mind than the number of neurons. It’s possible we’ll stumble our way into AGI by throwing more hardware at the problem. But I’d like to think that if we ever get there, it’ll be thanks to having cracked the code of our own minds.

1 – In the brief research I did for this post, I found 0.28mg as the mass of a Cataglyphis bicolor ant brain and 9 kg as the mass of some of the largest sperm whale brains. According to those numbers, the whale brain is 30,000,000 times bigger than an ant’s, adding a 300 factor to the already impressive 100,000 times difference. I chose to report Kelly’s data because his work inspired this article. Five zeros make as compelling an argument as seven.

2 – There are more refined measurements still. The encephalization quotient, EQ, is “a relative brain size measure that is defined as the ratio between observed and predicted brain mass for an animal of a given size, based on nonlinear regression on a range of reference species.” Humans fair better using this approach, but the resulting ratios against other animals don’t seem to map with everyday experience. Humans have EQ 7.6. Dogs 1.2, only six times less. Or maybe there’s something there… After all, dogs get us to house them, feed them, and take them to the park, all for free.

Originally published on giolodi.com.

Last week, a Texas A&M University-Commerce professor failed his entire class on account of having used ChatGPT to complete their essays. How did he know? He asked ChatGPT if it authored those essays—and the AI said yes.

The school eventually reverted the decision, but not before the story went viral. According to The Register, “The kerfuffle highlights whether or not educators should use software to detect AI-produced content within submitted coursework.”

The “kerfuffle” also highlights a deeper issue. It shows the necessity and urgency of developing AI literacy.

Let’s try to muster some empathy for the instructor. The way he used ChatGPT clearly shows he had no idea how the tool works and its nature as a stochastic parrot.

You can see how someone without even a rudimentary comprehension of how ChatGPT operates could assume the tool can answer any question, including whether or not it wrote a given essay.

The conclusion that ChatGPT would know or remember the essays it helped fabricate falls apart as soon as one peers under the hood. We can forgive the instructor for misinterpreting what ChatGPT can do, but he was too quick in drawing his conclusions. Using a tool to make decisions without the faintest idea of the process it uses to generate the information it gives is irresponsible, especially when those decisions affect other people.

I’m speculating here, but I wouldn’t be surprised if he was exposed to many prophecies of imminent AI doom, including how AI could make him redundant while also giving his students new ways to cheat – one student did indeed admit using ChatGPT – and how this could have contributed to his adversarial mindset and hasty decision making.

As AI inevitably blends into more of our day-to-day, we can expect more incidents like the one at the Texas A&M-Commerce institute to occur. To mitigate that, we need less “AI will take your job” scaremongering and more “here’s when and how AI product X can be useful” information.

In Zero to One, investor Peter Thiel admonishes against overrating the power of chance and underrating the importance of planning:

Malcolm Gladwell says you can’t understand Bill Gates’s success without understanding his fortunate personal context: he grew up in a good family, went to a private school equipped with a computer lab, and counted Paul Allen as a childhood friend. But perhaps you can’t understand Malcolm Gladwell without understanding his historical context as a Boomer (born in 1963). When Baby Boomers grow up and write books to explain why one or another individual is successful, they point to the power of a particular individual’s context as determined by chance. But they miss the even bigger social context for their own preferred explanations: a whole generation learned from childhood to overrate the power of chance and underrate the importance of planning.

Thiel argues that Gladwell overplays Bill Gates’ early situational advantages and downplays the his resourcefulness and business acumen.¹ According to Thiel, Gladwell’s worldview as a Boomer paints how the author observes reality and the explanations he derives.

But two can play that game. Perhaps you can’t understand Peter Thiel without understanding his historical context as a Gen-Xer (born in 1967).

Both Thiel and Gladwell have a point, and both are parsing Bill Gates’ story through their theories of the world. As Karl Popper argued, observations are theory-laden.

Our theories, explanations, and worldview inevitably mediate our observations and the conclusions we draw from them. And keep in mind that only some of our theories are explicit. Many more are encoded at the level of implicit or even unconscious knowledge.

We ought to keep this in mind when analyzing situations and making judgements. How is my worldview shaping the way I’m explaining this even? What theories am I using to decode this observation?

This is a tough but crucial exercise. To refine our thinking we need to become aware of the theories through which we interpret our observations—then look for ways to sharpen them.

1 – For a detailed account of both Gate’s background and execution, see Gates: How Microsoft’s Mogul Reinvented an Industry and Made Himself the Richest Man in America by Stephen Manes and Paul Andrews.

Whenever I hear the term “consuming content,” I picture people pouring colorless sludge down feeding pipes that come out of their heads.

The entertainment industry, in particular, seems to have embraced the term content consumption. Streaming services churn out shows nonstop, and every new movie either sets up a franchise or is part of an extended universe. No wonder most of today’s entertainment is derivative and cheap when the starting point is that of something to be consumed.

It’s like most processed food that comes out of a package. It might be tasty in the moment but provides no nourishment.

I propose we change how we describe the act of feeding our minds. Let’s stop consuming content. Let’s savor and study instead.

Thinking in terms of savoring and studying might seem like a word game, but I consider it a mental model that raises the quality bar of your inputs. For something to be savored, it needs to be good. Shallow material offers little to study.

If you’re tempted to scroll through your phone while watching TV, whatever is playing mustn’t be good enough to savor. Find something better, something so engaging that your mind won’t ask for distractions.

The same goes when seeking something to learn from. Scrolling through tweets, skimming blog paragraphs, or speed-reading books can be done only when there’s little substance. If you can gloss over it, it’s not stimulating enough. Look for media that requires your full attention.

We live in a world of content abundance. There’s too much stuff out there for anyone to go through in their lifetime. And the floodgates of AI-generated media are about to be opened. Someone consuming content will be overwhelmed and dissatisfied because no matter how much they pour down the pipe, there’ll always be more ready to come in.

To savor and study require complete engagement and high quality. There’s not much content that passes that filter.

Change the verb to change behavior. Actually, why don’t we change noun, too? Let’s ditch the generic content and use proper names.

Don’t consume content. Study an essay. Learn from a course. Enjoy a book. Savor a movie.

Many technocritics argue that technology is taking away our ability to think for ourselves. YouTube tells us what to watch, Spotify what to listen to, and fitness trackers how to work out. They don’t seem to realize the internal inconsistency of this kind of reasoning. In assuming technology is thinking for us, they imply we can’t think for ourselves.

Baratunde Thurston gave a textbook example of this flawed reasoning on the What Could Go Right? podcast:

I have this smart ring, I wake up in the morning and my wife says, “How’d you sleep?” And I said, “I slept pretty good.” Then I check the app and the ring says I didn’t sleep pretty good. I’m like, correction, apparently, I had a terrible night’s sleep. I need to try harder to sleep tonight 'cause the ring told me I don’t feel the way I feel.

[…] Otherwise we just follow these generated instructions for what to do with every moment of our lives. How to tour a city, what to cook for dinner, where to turn in our vehicles, who to mate with. It’s very optimal. It’s very optimized. It’s very efficient. It’s also dead.

Luckily, co-host Emma Varvaloucas didn’t buy it. “I think that we’re a little bit more in control about our choices in how much tech is in our lives than what you just described,” she replied.

If your tracker gives readings that don’t match how you feel and your reaction is, “I need to try harder because the ring told me,” then I’m sorry, but you’re to blame, not the tracker. The Oura ring shows you data and its own interpretation. It’s up to you to decide what to do with it.

The ring is not trying to manipulate you. You are taking the path of least resistance and not asking any questions. Maybe they sold you a great heartbeat monitor, not a sleep tracker. Or maybe you truly didn’t sleep as well as you think, but your sleep quality has been so poor recently that your calibration is out of wack.

To assume that people do things just because their tech tells them to is demeaning and patronizing. Technology gives us information and presents us with choices. But we are the ones deciding what to do. Don’t blame the algorithms for the choices you make.

Aside from treating the listener like a child who can’t think for themselves, arguments like Baratunde’s pull us away from the more nuanced conversations we need to have. Because even though big tech can’t force us to do something, it sure as hell knows how to nudge us.

Filter bubbles, echo chambers, social media’s impact on the developing minds of teenagers… those are all real issues. We are not going to solve the if the quality of our thinking goes only as far as assuming that people do things just because their devices say so.

We must refine our worldview regarding technology because a lot depends on technological progress. If we want to thrive, we must keep pushing our technological level forward. Problems are inevitable. The more tools at our disposal, the better because we don’t know what problems might arise tomorrow. However, the more powerful the tool, the bigger the damage it can do if misused. You can’t have good without the possibility of evil—which is precisely why we need to sharpen our thinking.

We need to address the issues of how technology can be misused. But we can only do that if we move beyond the helpless child strawman arguments and begin to take ourselves seriously.

Now, I love cozying up on a rainy day with a book and a cup of hot coffee, but bad weather during the weekend is inconvenient for many reasons, one of which is missing out on the only time available to mow your lawn. Get a few of those bad weekends in a row, and your backyard will look like a jungle.

One perk of working remotely is that you can make the most of a lucky spell of sunshine during an otherwise rainy week to get the lawn mowed while you have the chance.

Opportunistic lawn mowing is only one example of how remote work allows people to integrate the “work” and “not-work” parts of their life in a strategic, effective, personalized way. Everyone profits from this arrangement. By letting employees promptly attend to personal matters, companies ensure no unnecessary cognitive load is put on people’s brains by unfinished business at home. Focusing on work is easier when your house is in order.

However, working from home is not enough to establish a mutually beneficial professional-personal life arrangement. Opportunistic lawn mowing can happen only in a company that is both async-first and result-oriented.

Async-first companies default to asynchronous, process-driven communication. They are built around a strong writing culture and minimize employee interruptions. In a well-oiled async-first company, people can tackle tasks from start to finish without notifications breaking up their focus. Synchronous interactions – meetings, huddles, and in-person gatherings – are deployed with intention only when high-bandwidth and real-time interaction is required.

Someone working in an async-first setup can take advantage of a lucky spell of good weather to mow the lawn or take over school pickup when their partner gets a flat tire because they pull work from their inbox instead of having work pushed onto their plate by haphazard notifications.

The second ingredient to unlock remote work’s maximum potential is for an organization to become result-oriented. That is, use the quality of people’s work as their primary performance metric, not the productivity theater of which and how many hours they sit at their desk.

Remember these two prerequisites when considering whether to join a remote organization. Canceling the office lease and giving everyone a home office budget is not enough. To be effective, remote organizations must adopt an async-first workflow and a result-oriented mindset.

Computer scientist and best-selling author Cal Newport shared an “interesting observation” on the importance of structuring one’s time in episode 247 of his Deep Questions show.

After having delivered the last in a long series of tightly packed projects, Cal finally found himself with a week without any demanding work on his plate. One would expect a week of reduced workload following a long period of intense work to feel light and relaxing, but that was not the case. “I was ending days feeling rushed and feeling like I was not on top of things,” Cal tells his listeners, “Which is crazy because, objectively, the amount of available time was probably doubled.”

The discrepancy between a low workload and a high feeling of unfinished business was caused by Cal taking an unstructured approach to his daily schedule. When you don’t have much on your plate, you might as well take it easy and tackle things at your leisure, right? But as Cal observed, because of the lack of time pressure, the small stuff ended up “metastasizing” and eating up all the available space in the schedule.

Apparently, not even the man who the Financial Times credits with having rewritten the productivity gospel can escape Parkinson’s 1955 adage: It is a commonplace observation that work expands so as to fill the time available for its completion.

The experience renewed Cal’s appreciation for planning and structuring one’s time. “Even when you don’t have a lot to do, you have to structure your time like you have lots to do,” he recommends.

Mind you, structuring your time like you have lots to do doesn’t mean coming up with new work to fill up a schedule that would otherwise be sparse. It means treating shallow work with the same rigor you dedicate to the most demanding of tasks.

Don’t let work creep up on you and consume more time than necessary. Be proactive and intentional by strategically planning your work, however much or little that might be, throughout your day.

Dropping off or picking up kids from school is a job a parent can manage alone. When both parents participate, one of them is technically redundant. They could be doing something else with that time.

But the point of both attending the ritual is not about efficiency. It’s about quality. We have young kids, and we love to see them interact with their peers. We like to chat with the other parents. We want the kids to see us doing things together as a family, even when they are mundane. Especially when they are mundane.

My wife and I know we only have a few years’ worth of school runs. Before long, our kids will rather ride to school with their friends than be seen around with their boring parents. So, as inefficient as it might be, we make time to savor this fleeting experience—together¹.

The point of productivity, the one done right, at least, is not to squeeze as much output from every minute of the day. Productivity should be about maximizing the value you get out of your time.

The shift from output to value is subtle but crucial. It requires sometimes being intentionally inefficient in the short term in favor of activities that will pay off in the long term.

1 – A pushback this post might receive is that not everyone has the privilege of a schedule flexible enough that both parents can take the kids to school together. There are also single parents out there who have no choice but to attend every pick-up and drop-off. Obviously, this reflection is not for them. But many folks do have that luxury, and, hopefully, many more will in the future. I am one of those, and the best I can do is try not to squander it.

We’ve been chatting a lot about AI here recently. I tried to argue against prophecies of AI apocalypse without disregarding concrete issues such as model bias or LLMs ability to generate plausible-sounding lies. AI could enable new generations of builders and trigger a wave of new careers that are as unimaginable for us today as my job as a mobile platform engineer was for my grandparents, but only if we take it seriously, without either fear or hype.

Let’s shift gears and move from the abstract to the practical.

My article in GitHub’s The ReadME Project, Accelerate test-driven development with AI, takes a look at the kind of synergy that AI brings to software development today.

The article shows how to Copilot enhances the Red-Green-Refactor loop at the core of Test-Driven Development by reducing the lag between you picturing a change in your mind and the code to implement it appearing on screen. It shows how to write prompts that can generate entire test doubles or how to guide Copilot to refactor your code.

Here’s one of my favorite examples:

Reading through the guide, you’ll also notice how, at no point, Copilot is ever in charge. Like the artificial intern that it is, GitHub’s AI can only suggest code. It’s up to you, the human, to decide whether to use the suggestion and to guide Copilot with precise prompts.

Whether you are interested in software development or not, I hope you’ll find the article interesting as a hands-on look at the capabilities and limitations of LLMs.

I encourage you to give integrating generative AIs in your workflow a shot. Once you find the combination that is right for you, I guarantee they’ll provide powerful leverages.

Each epistemic invention expands the web of verifiable facts and links one bit of knowledge to another. Knowledge is thus a network phenomenon, with each fact a node. We say knowledge increases not only when the number of facts increases, but also, and more so, when the number and strength of relationships between facts increases. It is that relatedness that gives knowledge its power.

A knowledge network is more than the sum of its parts. The paths that connect individual nodes can reveal unexpected insights and the wisdom to pursue them.

It’s because of the networked nature of knowledge that organizations revolving around a writing culture have a better shot at being effective and why it pays to leave context breadcrumbs and write everything twice.

Open graph image credits: Alina Grubnyak on Unsplash.

The lump of labor fallacy is the incorrect assumption that there is a finite amount of work to be done. This misconception is responsible for much of the fear of innovation and automation that humanity has displayed throughout history, most recently in the case of generative AIs.

From the lump of labor viewpoint, the more labor we automate, the less there is left for humans to do, resulting in previously skilled workers becoming irrelevant in the economy and unable to provide for their families.

Luckily, there are infinite problems to solve that will keep spawning new jobs for us to do. History shows that every time labor is automated or made more efficient, people always find ways to use the newly freed-up time and resources.

Here’s how MIT professor of economics David Autor explains the lump of labor fallacy in a Planet Money episode:

Consider that a hundred years ago, Americans used to spend 70% of every dollar on the absolute basics of life, food, clothing, housing. But then thanks to machines or poor labor or immigrants, all of our things became cheaper. As a result, yes, some jobs were lost, but consumers suddenly had a lot more money to spend because now thanks to all these cheap goods, only 40% of every dollar was going to the basics of life. So what did people do with the rest of that money? Well, they didn’t just put it in a savings account. They started to spend it on restaurants and entertainment and adventures, therefore fueling new jobs in new industries.

Once you are aware of the fallacy, much of the AI doom prophecies go from terrifying to laughable. But there is a caveat.

While it’s true that humanity as a whole is better off every time labor is automated, the consequences for the individuals whose jobs become redundant can be severe. This is why, now more than ever, one should never be satisfied with one’s existing skill set and skill level but commit to a lifelong learning journey. As former Intel CEO put it, only the paranoid survive.

Understanding the lump of labor fallacy removes the fear of looming economic irrelevance and shows a path to thrive in the upcoming marketplace. We need to build the mental flexibility and resilience to adapt to innovations and industry changes. Stay sharp, develop range, and don’t settle.

Thank you for reading this far. I’d love to hear your thoughts—leave a comment, hit reply, or DM me on Twitter.

If you found this useful, please hit the like button to let Substack know about it, or share the post with someone else who might. And if it’s your first time here, don’t forget to subscribe.

Chat soon 👋
– Gio

Subscribe now

Join chat

Share