Do earthquakes happen on the moon, please explain? I vividly remember this question in an English course examination in my school days. Many of us felt angry because this is a physics question that has been asked during the English exam. Some just answered ‘yes or no’ thinking there will be some grace marks awarded. A few of the students even got their parents to the class next day to argue with the teacher that this is an out of syllabus question and the pupil should be awarded full marks to it.

In the next class, our teacher explained that this is a perfect English exam question and the answer to it is – “Earthquakes happen only on earth, if it happens on moon then it is called a moonquake”. Our teacher definitely disrupted the rote style of learning and forced us to comprehend more deeply what we read than just a shallow pass through of memorised phrases.

To get to the mastery, you start by the rote but you have to get into deeper comprehension soon or else we will be stuck forever in a knowledge loop without any applicable wisdom. Software development is no different, especially when agility is the key. What is the problem that is troubling the software industry now? It is the glut of rote learners who are misguided to stay as rote learners with an assurance by a small set of high profile consultants and trainers.

In this model, the luminaries will be happy to certify the rote learners as experts because it benefits them a lot in terms on money and fame. They ill equip a lot of people from developers to executives with runbooks and cheatsheets full of jargons and metrics which then becomes the industry standard. Recent article from a major consulting company on measuring developer productivity is one such example. It has angered all the real practitioners and experts forcing them to vent out their opinions including me. Some experts have taken such articles and have given a thorough explanation of why it is wrong.

What should be the model then? Rote learning is always a starting point. At the starting point people are dealing with the knowledge in an abstract way, which means they won’t understand the underlying value and intent of a practice or a process. When they do something by rote, they work along and observe the practitioners, understand the value and intent behind what they are doing and are able to internalise. Over time, because of deep understanding gained on the ground and interactions with other people, people develop expertise. The common theme here is collective growth as a community not as an individual who excels in arbitrary metrics.

If an organisation wants to improve their productivity, driving through metrics will always result in behaviours that encourages excellence in metrics. It will easily runaway into a toxic culture where no one wants to help another unless it helps their metrics. Organisations should bring in a culture of resilience where communities exists with an intention of upping the notches of everyone, Some people when they gain very good expertise they go even further and extend the community reach outside of their workplace. Information should flow more freely across layers and people should feel safe to try new things and fail safely to improve productivity and resiliency. These things are hard to keep in effect and hard to measure, so not many leaders try it.

Now coming back to the title of this blog. I am beginning to see that the so called experts or luminaries are not calling out that ‘it should be called moonquakes’ instead they are selling tools and frameworks to observe and detect earthquakes on moon and reinforcing wrong understandings to a great deal.

A fighter is plane flying at 1000 kmph fires a few bullets which leaves the barrel at 2000 kmph, the plane then accelerates to 2000 kmph and catches the bullets back. Is this statement true of false?

If the above question is asked to a primary school student who sees only the facts in the statement, the answer is likely a NO but a student who has understood physics and all the dynamics that are in play will certainly bring in air resistance, terminal velocity and acceleration into the equation and say it is possible. The reason why the plane can catch the bullets is – a bullet will be subjected to air resistance and lose its forward velocity and fall towards the ground, while doing, so without any external force it will reach terminal velocity because of more denser air towards the ground; the plane can dive faster and break terminal velocity because of its thrust and will eventually over take the bullets given this happens at a sufficient altitude. It happened in real so it is not a theoretical possibility, check this link.

Many of the professional management and executives in software development I observed belong to the static understanding of agility in software development. They are well versed and equipped with a lot of terms and concepts which in their mind are non negotiable and have to be followed to the dot. Examples – estimations using fibonacci numbers only, planning only for repeatable velocity iteration over iteration, enforcing only either solo or pair programming, locking main branches and allow only PR through gated reviews. The list goes on without addressing the nimbleness needed.

The management’s mindset is opposite to agile philosophy, they treat the development team like resources to be utilised and deal them as numbers. The best software teams are the ones that understand business and measurable results to back some decision, which will help them to continuously modify the software to facilitate business goals. If a business idea takes about a month or more to implement and test it out, more or less there isn’t any agility, it is just that people are convinced that they are doing the so called AGILE PROCESS. If we have the right ways of working (not process) then barring major platform work most of the idea to production cycles should be within weeks or days timescale, not in months.

What gives agility? Before we get to that, we should answer what kills agility. My take is, two things that kill agility are (1) professional management and (2) certifications/processes/frameworks claiming to be AGILE. Professional management is conventionally equipped to deal with industrial and manufacturing, they always start from there which is an efficiency oriented management style and adapt to an effectiveness oriented management style.

When people are dealing with software development as if it is a static system, things appear in black & white and easy to understand and predict. It is hardly the reality, software development is a complex dynamic system which inherently has to be driven by people on the ground than the management. The focus for management is to be a facilitator to ensure that the staffing is good, right tools in place, work environment is not toxic and remove the impedance for business knowledge & communication. As much as possible avoid the cargo cult way (a.k.a. some major agile framework) of managing software projects, understand the dynamics at play and plan based on that.

Some things that has worked so well for me

• Very good individual contributors who excelled at their work were better leaders as managers than professional managers from reputed schools
• Long running small and stable team compositions achieve big results compared to frequently churned or large teams. They may start slow but accelerate to a good velocity which sometimes is hard for the business to keep up
• Certifications and frameworks in the AGILE business do a lot more damage than help
• There is no substitute for XP practices, continuous delivery and engineering rigour
• Developers who understand business contribute multi-fold than business people who understand tech

LLMs are trained on public data, the limitation to grow is not the hardware or parameters but it is the quality of content that is available to further train the model. With these new tools, newer content will be generated at a pace that can’t be consumed easily by humans and we may need to use these tools again to summarise and create action items for us to follow. A new risk that has emerged is that we will be stuck in a content whirlpool where these tools create more content based on the content they had already created.

This is similar to what many algorithmic feeds are doing to us already. You get thrown similar content that you have watched, songs that you have heard and articles that you have read. Those algorithms can be worked around by going private mode at least for that window so that the serendipity factor increases. The LLMs and similar tools in a way work from a same public data set, making them behave like coupled systems. The more self generated data they feed on the public domain the more synced they are. This is very similar to what is explained in Kuramoto model. When the underlying foundation is the same and more content generated is being added to the same foundation from different models, then models may begin to converge on what they can generate.

There are paths possible that this convergence does not happen. One is to start getting into personal and proprietary data which is out of the bounds of the models now. Who gets access to what will start getting to matter a lot, our personal data is a goldmine and will be monetised well. The other is to advance the technologies to start reason well with first principles and heuristics, this will require less magnitude of data and may be years away. So the first option to getting into private data is more of a possibility, entities with fiduciary responsibilities of data will be tempted to go for legal ways of monetising which can be loop holes that may not be in the best interest of a layperson.

Before more reasoning capabilities are built, it is better to live in the content whirlpool than feed the private lives to insane computing power of these models.