Notes from Tech Buddies II

The second Tech Buddies Space was supposed to be about careers in AI. It was, eventually. But like the first one, the conversation refused to stay in one place. We started with a political movement in Ghana, moved into AI and jobs, went backwards into the history of technology, talked about software careers and degrees, and somehow ended up talking about luck.

Thanks to Gemini, Chinedu, Bubu, Edem Kumodzi and everyone who joined, spoke, challenged an idea or simply listened. I am writing this from memory, not as a transcript, and I am probably missing some of the things people said, so I would rather keep this as a record of what I took away than pretend it is a perfect account of the conversation.

Can a movement survive the people around it?

Gemini and Chinedu kicked this part off with something concrete: the government’s new arrangement for getting workers to work. In the mornings, the roads coming into Accra are choked with traffic, while the lanes leading out of Accra are much freer. So instead of sending workers through the choked inbound lanes, the government is using the outbound lanes to bring them into the city on the Aayalolo buses. Gemini pointed out that arrangements like this have worked in other countries as temporary solutions, and this one is temporary too. It is not meant to be permanent. It is a way to ease the traffic and get workers to their workplaces on time.

The concern was how it was introduced. Many drivers did not seem to know about it, and the communication amounted to a one-day campaign by the minister, and that was it. That led us to Ghana Must Work and the wider question of why movements like this emerge in the first place.

One concern raised in the conversation was the psychology of participation. Visible personalities can make it easier for people to notice a movement and join it. But that creates another question: can the movement survive the personalities?

Gemini said this is something they are trying to deal with. Their answer is to try their best not to centre the movement around one person or a few people, so that when those people move on or go their separate ways, the movement is not left standing on them. Chinedu’s view was that, psychologically, it may already be late for that. He still hoped the best for them, and so did all of us.

We have seen previous youth movements, including #FixTheCountry, and that history made me think about what happens when the people who initially create momentum move on.

But my concern went further than that. I wondered whether political parties can learn to use the frustration behind these movements for their own advantage. A political party in opposition can recognise that young people are angry about real problems. It can support that frustration, speak the language of the movement and position itself alongside the people demanding change.

Then power changes hands. What happens next?

My fear is that some of the people carrying the movement can gradually be pulled into the political system through appointments, incentives, access or other opportunities. The movement loses some of its original energy, not necessarily because the problems disappeared, but because the people who were pushing it are no longer pushing from the same place. I do not know that this is always what happens. It was a concern I was raising.

That made me think about something else: people may prefer to be led than driven. A movement needs people to feel that they own the reason they are there. If the entire thing becomes attached to a personality, a political party or a particular moment, what happens when that person or moment disappears? The real test of a movement may be whether the idea survives the people who first made it visible.

Then we got to AI

Eventually we reached the reason I had planned the Space in the first place: careers in AI. The conversation was messy in the best way because nobody really knows where this is going.

One participant pushed back on the idea that AI is already replacing jobs at the scale people sometimes claim. Another view was that we have seen technological cycles like this before, and that the current AI boom may eventually cool down before we understand what the lasting software actually looks like. Chinedu also brought up the possibility of the current bubble bursting and things settling afterwards, and that idea made sense to me.

Every technology boom has a period where everything suddenly becomes about the new technology. Companies rename things, investors chase the trend, products add the new buzzword and everybody starts trying to predict the future. Then reality arrives. That does not mean the technology disappears. It can mean the opposite. Once the hype settles, the useful parts become easier to see.

The interesting question for me is not simply whether AI will be huge. It probably will be. The harder question is what software looks like after the excitement becomes normal. Maybe the next generation of software will be less about screens full of buttons and more about systems that can actually do things for us.

Not every layoff is an AI story

The COVID period came up too. People were excited about remote work. Companies hired aggressively. Some companies overhired. Now we see companies laying people off and AI is often part of the explanation.

Chinedu made a point I liked: some of this may be economic rather than technological. That does not mean AI is not changing work. It means we should be careful about assuming that every person laid off during an AI boom was replaced by an AI system.

A company can hire too many people during a period of unusually high demand, realise later that the numbers do not work, and reduce the workforce. It can then happen to be doing that at the same time AI is becoming more capable.

Those are two different stories.

AI can still make some work cheaper or unnecessary while the company is also correcting decisions it made years earlier. I think that distinction matters because otherwise we start treating every headline about layoffs as proof of a future that nobody has actually demonstrated yet.

So is software engineering still worth learning?

This was probably the question underneath most of the AI conversation: if AI can write code, should somebody still study computer science?

Bubu’s answer was one of the things I liked most: the fundamentals still matter. Software engineering is not going anywhere just because the tools are changing. What changes is the set of skills that make somebody useful. If AI gets better at writing code, knowing how to type code quickly may become less valuable. Understanding systems, debugging, architecture, trade-offs, data, security, product thinking and how software actually behaves in the real world does not suddenly become useless.

Edem Kumodzi made another point that pushed the conversation further: repetitive work is the part most naturally exposed to automation. That applies far beyond software. If most of your job is a predictable sequence of actions, a machine has a clearer path to doing it.

For a software engineer, the answer may eventually be to build the agent that performs the repetitive work instead of spending your entire career performing it manually. That changes the job. It does not necessarily eliminate the need for the person who understands what should be built, why it should be built, whether it worked and what happens when it fails.

The degree question is more complicated than “degree or no degree”

We also got into whether studying computer science is still worth it. I do not think the conversation produced a simple answer, which is probably a good thing. You can become technically capable without a traditional computer science degree. But that does not mean the degree has become worthless.

At some point we started talking about opportunities where a credential itself can open a door. Edem Kumodzi gave an example from experience where a degree mattered when accessing an international opportunity. The point was not that the degree made someone technically capable overnight. It was that some systems still use credentials as a filter.

That changed the question for me. Instead of asking “Do I need a degree to learn software?”, maybe we should ask “Which doors does a degree open, and which doors can I reach without one?” Those are very different questions. A degree can be useful even when it is not the only route to competence.

Luck is everywhere

The last part of the conversation went somewhere I did not expect. We started talking about the Salifu and Master game and the debate around a later, more refined Nigerian version. I do not think the most interesting part of that discussion was even the argument over who copied whom.

It made me think about how ideas travel. Someone can have the idea first. Someone else can execute it better. Someone can have better distribution, enter the market at the right time, meet the right person, or simply be in the right place when a technology becomes possible.

Then we somehow ended up talking about luck.

Edem Kumodzi brought up the kinds of advantages that can exist before someone even starts competing. Bill Gates came up as an example. His ability mattered, but so did his environment: the family he was born into, the school he attended, and unusually early access to computers. That does not reduce the achievement. It makes the story more interesting. Success is rarely just talent plus hard work. There are usually conditions around those things that are easy to miss when we tell the story backwards.

That led me to a question I keep thinking about: if luck matters, can you increase your exposure to it? Maybe you cannot manufacture luck. But you can put yourself in more rooms. You can learn more things, build more things and meet more people. You can keep trying things long enough for an unexpected opportunity to find you. Maybe that is one version of creating your own luck.

What I actually took away

The Space was supposed to help answer whether careers in AI make sense. Instead, I left with a bigger set of questions. Can a movement survive the people who start it? How much of the AI job disruption is actually technological, and how much is ordinary economics happening during an AI boom? If repetitive work becomes automated, what should humans become better at? Is software engineering changing or disappearing? What are degrees still useful for? And how much of success comes from things you had no control over before you even started?

I still do not have clean answers. But I think that is becoming the point of Tech Buddies for me. I do not want every conversation to end with somebody announcing the correct answer.

I want to leave with a better question than the one I walked in with.