Rahim Hirji on Nothing Ventured: what happens to judgement when nobody gets the reps

Aarish ShahFor a lot of immigrant founders there is this invisible ladder. Survival, stability, then ambition. Does that lens create better entrepreneurs, or does it limit how far people are willing to think?

I talk about survival skills. Survival skills is the inherent first step in a super skills ladder. So I've got survival skills, street skills, specialist skills, which I think we call vocational skills here in the West, then soft skills, which EtonX was built around, and then super skills on top.

I believe that to be the ladder that anyone goes through. Immigrants go through it as they're thinking about their generational growth. My grandparents and great-grandparents went through that and so did yours. Many people who've lived in this country all their lives have had to go through that. Some will be the first people to go to university in their family.

As an immigrant I think there is this thinking around survival skills and survival thinking which becomes like lifeblood to people who are in the entrepreneurial space. I don't know what the link is between the survival skills that we might have had because we couldn't get a meal, and where we might be in a startup where we might not be able to pay the staff, but I do think that they are very important skills, and that's why they're the first thing on my ladder.

Survival skills is really about, are you going to take that big risk or not? Which in both of our cases was one of our predecessors getting on a dhow, on a boat, and shipping himself from Gujarat in northwest India down to the east coast of Africa. I talk about that story in the book, that the risk that generation took was that they probably had a meal if they were living in India, but they saw a better life for themselves and they took that risk, even though there was probably a one in three chance of them making it across to another country.

Aarish ShahWhen my grandfather and his brothers first moved out to Kenya they moved because of famine. Even a meal was literally table stakes.

The progression you get from that is interesting. You go from survival to street skills, and that's anything from communication to negotiation to running a shop, which is what my grandfather did in Dar es Salaam in Tanzania. Being able to work with the locals, learning the language, Swahili was the language in Tanzania. And then you progress up from that. My parents moved to the UK and they decided that they needed to learn and progress with their career. Accountants, doctors. That's the thing I call specialist skills.

You and I both know that you can learn a job, you can learn the skills within a job, but it's those soft skills, and soft skills is the worst possible name for this, because they're the hardest things to learn. You need to be able to be a good leader, you need to be able to communicate well, you need to be able to lead a team, you need to work with a team, you need to have the critical thinking to be able to make judgement. All of those things become really important. And then on top of that I've built these meta skills.

I don't think it's necessarily just an immigrant story that you need to go through this. I think it's for everyone. Everyone's going on about it being a very tool-driven society now, because everyone's going to Claude or Claude Code or ChatGPT, but really we're forsaking some of these human skills, or super skills, that become even more important when everyone has access to every single LLM at their fingertips. Multiple superpowers. What becomes different then, when information is no longer the barrier?

Aarish ShahIn professions like accountancy, legal or consulting, the mad rush is to get rid of all of the juniors and the analysts. That creates two problems. Where do those people then work, and it is the people at the beginning of their careers who are paying into the system to fund those at the end of theirs. But the more dangerous problem is that you will have far fewer people coming up through the ranks learning the human skills of judgement and synthesising information.

If you go back to when we started our careers, my first day I was probably photocopying, doing data entry, job shadowing. You learn by being in the office, you learn by osmosis, you see what people are doing, you learn about the interaction, you see what's important, what's not important, you see what the priority is. It was an apprenticeship of sorts. You start off and you don't really know what you're doing, at least I didn't know what I was doing. You turn up there and very quickly through that osmosis you learn what's going on and what is important. And that's the beginning of judgement.

Aarish ShahThe other major shift is that with COVID there was a big swing to remote and then it swung back, but not fully. That apprenticeship tends to happen in person.

Absolutely. The thing that I'm seeing is a lot of these very smart people who make it into the workplace, maybe one or two years in, who want to learn an expertise. You're able to do that very quickly now. And this is a concept I call synthetic seniority. Someone with not very much experience is able to emulate the fact that they've got ten years experience and able to sit at the table with someone else who's got the same level. And that creates this problem, not because they're able to have that conversation. The problem is when things go wrong and they don't have the scar tissue of what went wrong last time.

The more I speak to organisations, that's what they say to me, and I say, this is the name for it, and they go, yeah, that's exactly right. The question is what do you do about it. Invariably there are those individuals with a level of synthetic seniority, and if I was that age in that position I'd be doing exactly the same as them. And then there are those who are worried about those with synthetic seniority. Pair them up together. The power of those teams, someone who's got this augmented mindset to be able to think in that way, with individuals who've got that level of wisdom, those two together, or a group of those people, unstoppable. That is AI plus human, judgement plus judgement, all of that together becomes this amazing thing that could be a rework of work.

I'd say to organisations and leaders and teams to really think about how you concoct those new teams of the future, what are those pods that you put together of individuals with different skill sets.

Aarish ShahSome advisers tell parents to limit children's exposure to AI, others urge early hands-on use. What do you tell them to prioritise, and what is the failure mode you fear?

Everyone's got an opinion on this. My view, and it has been up to now, is that I have a hard line on no AI until fourteen. I have not allowed my youngest to touch any AI up until now. That said, I'll be sitting down with her this summer and trying to go through it in a very clear way so that it doesn't appear to be a cheat code, because that's the issue that you might have with kids.

If you take a longitudinal view of this, the brain is still growing until you're twenty-five. So if you start to short circuit this and you lose that cognitive ability to do certain things, which in the UK has dropped dramatically because of lots of other reasons including social media. If your child is susceptible to social media and using WhatsApp and Snapchat and gets addicted to it, stay away from things like AI at least for a while, because that can be even more addictive. Those are infinite rabbit holes.

I draw this line at fourteen for that reason, but also because I didn't want my daughter to use it until I was a hundred per cent sure about what I was going to do. When you come out of university, or before you go into the workplace, you need a body of work. A portfolio. Here's an AI project that I built from scratch. Here's a company that I did. Here's a real life project. Here's some human stuff that I did. Here's some stuff I did by myself. Here's some stuff I did with a team. AI fits within that.

My eldest, who's just finished her first year at university, is pro AI, but she will solve the problem first and then she will use AI for citations and additional elements. Long story short, if you sell your soul to all of these tools too early, as a parent, if you say to your kid, go and build, you need to give them the right guard rails, because the cliff is so high that you could fall and you could jeopardise their learning for the rest of their academic career.

Aarish ShahIt would also be dangerous for a parent to completely shelter their children from AI, because it's going to be such a massive part of their lives. The best outcome is that they are comfortable using it but recognise that it is a tool. It is not a replacement of you.

That's absolutely right. I would take a step back and say that whatever kids go through over that period between fourteen and twenty-one at least, there is a level of friction. They need to be solving problems all the time. And they don't think of the tool, whether that tool is AI or not, as a quick way of doing something. Friction is a paramount trait of learning, and not just in formal education, in life. Friction and failure and figuring things out. That painful process that you need to go through on becoming an individual is important. Whether you're at fourteen, whether you're at twenty-one and going into the workplace, whether you're at fifty, you still need to be learning.

If you're fifty or fifty-five or sixty and you say, oh, AI is not for me because it's a bit complicated, you need to change the way you think about it, because this is the future. It's just that you need to use it in the right and the most appropriate way. If you look at someone mid-career, they should be thinking, I want to leverage all of my experience but put myself centric and say, how am I going to solve this problem using these tools? Which tools should I be using? It's not always the default tool that you think about. What is the problem that you're solving? And how does it augment and accentuate what I have got? We're in this amazing era now where we can accentuate and do things in a way that we could never do before. They were impossible three or five years ago.

Do some stuff in the AI space. Figure things out. Learn what you need to do. But also don't forsake who you are as an individual. Don't forsake the relationships that you have. Don't forsake all of these human skills and these meta human skills that you need to have, because those are the things that are also going to be important.

Aarish ShahI'm really bullish that in real life activities, events and human to human interaction become more and more paramount in a world of AI. If everyone has access to all of this information and can do the same thing by prompting, what becomes super important is how I interact with you, and what experience I bring that the AI doesn't have context of.

AI is this amazing autocorrect for the future based on the data that it has been trained on, but it hasn't got lived experience, and that's what humans have. It's not so long ago that we went through COVID, and humans were shut away for so long. We made good with Zoom, but most people were craving human connection. My daughter was thirteen at the time. That was the worst period for her, to not have that interaction with her classmates and her friends.

We'll get to a point where companies will realise that they've got rid of too many individuals. They have lost the lived experience. They have lost the wisdom. They have these great tools, amazing workflows, and it'll be this magic machine that's doing stuff, and there won't be any value or soul or truth in that. And that might be a great world for some investors, but I don't think I want to be part of that.

Aarish ShahProprietary data is also what is sitting in your head as an individual, and that is massively valuable. If everything a consulting firm produces is generated by AI, what's the purpose of the firm? There needs to be an edge and that edge can only come from a human being. And I always come back to trust.

We call it the three T's. Truth, trust and taste. You think about those three areas in this new world. You may get some truth, you may get some trust and you may get some taste, but then it's, where can you get more of that? Can you get more of that from humans? There is this fine line where people dip their toe into this new world and they think everything's going to be okay. There will be groups who will get burnt by overinvesting or overspending or outsourcing too much of what they do, and then they'll revert back to some of the processes.

Especially when you're making large transactions, it's very much a relationship. If you're buying a service from someone, they might be buying some of your workflows, but they're actually buying from you because they trust you. And sometimes your taste, although that's debatable sometimes. My taste is I wear black t-shirts.

Aarish ShahHow should large organisations think about synthetic seniority with their existing staff?

If you play it forward, if you start saying, I'm going to trust the person who's able to use the tool better than someone who's got the wisdom, you'll make a judgement call that the person who's got the wisdom is slower because they can't use the tool. And you might say, I'm going to get rid of that person and promote the person who's very good at using the tool but can't deal with the edge case, or can't deal with something that goes wrong.

Aarish ShahThe perfect example is, we can build a financial model using Claude. You can get something approximating a model, but it falls over because it doesn't understand what the assumptions are, the uncertainty, the risk, the tolerance. I can output a model in three hours without my judgement, but it's not going to be as good as a model that's taken me three days to build with my judgement.

That's where you're paying for the actual output that you get, or you're actually paying for the output with someone standing behind it. Because it's one thing having an answer. What's the backup plan? The model may be an asset that is output, but it is not the output of the work. The work is the stuff around it. It's the thinking and the judgement that creates the model.

Aarish ShahA finance leader told me they used Claude to interrogate their ERP to pull out a reconciliation, and the auditors turned around and said we can't accept this, because we cannot audit how Claude reconciled this. It's the proof of work. Like maths A levels, you would get marks for the work even if the answer was wrong, because it shows how you're thinking.

There's a concept that I try and get clients to think through. Is this human at the start? Everyone talks about human in the loop, and obviously I do believe in that, but I also believe in human at the start. Where you're talking about identifying everything, actually, is that the right decision? Should you be using AI at that point? That's part of my augmented mindset skill. Human plus AI is very powerful. The question is when should you use AI, when shouldn't you use AI, and should you be solving that problem first.

There's a wild west out there. There are organisations that are turning to reverse pyramids and keeping the people at the top and having fewer people coming into the funnel, and I'm not sure where they're going to be in five or ten years' time when you get the natural attrition and who are they going to be, the leaders. You have those that are ripping people out from the middle. There are those who haven't done anything yet and are kind of rabbits in headlights. Some industries are too regulated to change certain things, and some are too protective over the creative process.

The only thing we do know is that those who are struggling at the moment are struggling to get a job. Those are the sixteen to twenty-four year olds in the UK. Which is very worrying.

Aarish ShahWhat did you get wrong about edtech?

Hindsight's a wonderful thing. Tech inherently is very quick. You want to do everything very quickly. Education, if you look at it, is over many years. There's a risk element associated with education. You don't want to get it wrong for these kids.

If you take the two years to GCSE, there are so many different elements in that. There are state schools, private schools, different exam boards, teachers who teach differently and in a different order. So when you look at an edtech product trying to get into schools, it's very difficult to think about usage and proof of efficacy, because you've got a very small window and then you can't do it again. You can't do it again with the same kids. You've got to pick a different cohort, and sometimes you're waiting a year. The proof point for a product which is seasonal in education is very difficult.

The other thing is the assumptions that you make. You think that the channels are going to be easy. They never are. In each country there's a different government, different regulations. In higher education sometimes the decisions take years. As an investor you've got to bite the bullet and say, this is going to take this amount of years. Investors invariably are impatient, because they want to see product market fit across many segments, and you can't prove them in a single sprint.

Aarish ShahThere's a story about a beach bar.

There's this girl called Sugar. She owned a beach bar in Thailand where we went to visit last year. Compared to all of the other beach bars on that beach, it was doing the best. I went to speak to her and said, this is the most popular place on the beach, what's your secret? And she said, oh, I use ChatGPT. And I said, how do you use ChatGPT? She said, I use a mixture of ChatGPT and different models to look at the pricing across all of the beach, to understand what other people are doing, where we haven't got products. For example, I needed to have halal to bring in a certain group. And she goes, I do that once a week to augment what I'm doing.

She was this girl who looked as if she was fifteen or sixteen, I think she was twenty-five. And she was dominating because she had this mindset around thinking about how am I going to do this using this technology, but still maintaining, you wouldn't know that she was using tech, because everything else was very human and very manual. It was the whole culture of why you might go to Thailand and go to a beach. It was about how do I improve the experience for the humans.

Aarish ShahWhat is one thing you have a contrarian conviction about?

Everyone's saying, what should I use AI for? I'm saying, what shouldn't you use AI for? Don't assume that it's a panacea for everything. You can over-optimise. Don't assume that the path is AI only, AI first. Assume that there is a path, that you've got to make the decision, that you have the agency to make the decision, and you can take that step.

Aarish ShahWhat is something that moves you emotionally?

I feel this sense of time is really important, and the thing that we just don't know is how much time we have. We might see someone for the last time. You might have a WhatsApp with someone and they never reply. You hear about someone who's passed away after a period of time and you just didn't know for a year.

But the thing that you can't get away from is, we've got kids, and we've seen them grow up. You had this relationship with this baby first, and then with this toddler, and then a tween and then a teenager and then invariably an adult. And all the data is, once they've reached that phase, you're not going to spend that much more time with them. But the thing that people don't tell you is that you have said goodbye to each version of that kid.

I see these pictures of my daughters when they were three, and I remember these really happy times, and I'd love to go back to that time and hold my daughters again and lift them when I used to lift them and have those laughs and giggles, but that person's gone. You meet friends along the way and usually you're at the same phase, and you may have met them as a teenager but you grow together. But as an adult, seeing your own child go from version one to version two to version three to version adult, you can never go back to version three. Which gets me to say, value your time with your family, with your friends, with your children, because you just don't know when it's going to be.

Aarish ShahRahim, this has been an awesome conversation. Just been phenomenal. Really appreciate you having spent the time.